Cornea and External Eye Disease - Beximco Pharma · Anterior lamellar keratoplasty is presented by...

ESSENTIALS IN OPHTHALMOLOGY: Cornea and External Eye DiseaseT. Reinhard · D.F.P. Larkin (Eds.)

ESSENTIALS IN OPHTHALMOLOGY

G. K. Krieglstein · R. N. WeinrebSeries Editors

Glaucoma

Cataract and Refractive Surgery

Uveitis and Immunological Disorders

Vitreo-retinal Surgery

Medical Retina

Oculoplastics and Orbit

Pediatric Ophthalmology,Neuro-Ophthalmology, Genetics

Cornea and External Eye Disease

Editors T. ReinhardD.F.P. Larkin

With 138 Figures, Mostly in Color,and 20 Tables

Cornea and ExternalEye Disease

123

Series Editors

Guenter K. Krieglstein, MDProfessor and ChairmanDepartment of OphthalmologyUniversity of CologneJoseph-Stelzmann-Strasse 950931 CologneGermany

Robert N. Weinreb, MDProfessor and DirectorHamilton Glaucoma CenterDepartment of Ophthalmology – 0946University of California at San Diego9500 Gilman DriveLa Jolla, CA 92093-0946USA

ISBN-10 3-540-22600-1Springer Berlin Heidelberg New York

ISBN-13 978-3-540-22600-0Springer Berlin Heidelberg New York

Volume Editors

Thomas Reinhard, MDProfessor of OphthalmologyUniversity Eye Hospital Killianstrasse 5 79106 Freiburg Germany

D.F.P. Larkin, MD, MRCPF, FRCSCornea & External Diseases Service Moorfields Eye Hospital London, EC1V 2PD, United Kingdom

ISSN 1612-3212

Library of Congress Control Number: 2005933478

This work is subject to copyright. All rights arereserved, whether the whole or part of the material isconcerned, specifically the rights of translation,reprinting, reuse of illustrations, recitation, broad-casting, reproduction on microfilm or in any otherway, and storage in data banks. Duplication of thispublication or parts thereof is permitted only underthe provisions of the German Copyright Law of Sep-tember 9, 1965, in its current version, and permissionfor use must always be obtained from Springer-Ver-lag. Violations are liable for prosecution under theGerman Copyright Law.

Springer is a part of Springer Science +Business Media

springeronline.com

© Springer-Verlag Berlin Heidelberg 2006Printed in Germany

Cover picture “Cataract and Refractive Surgery” fromKampik A, Grehn F (eds) Augenärztliche Therapie.Georg Thieme Verlag Stuttgart, with permission.

The use of general descriptive names, registerednames, trademarks, etc. in this publication does notimply, even in the absence of a specific statement, thatsuch names are exempt from the relevant protectivelaws and regulations and therefore free for general use.

Product liability: The publishers cannot guaranteethe accuracy of any information about dosage andapplication contained in this book. In every individ-ual case the user must check such information byconsulting the relevant literature.

Editor: Marion Philipp, Springer-Verlag Heidelberg,GermanyDesk editor: Martina Himberger,Springer-Verlag Heidelberg, GermanyProduction: Elke Beul-Göhringer, Springer-VerlagHeidelberg, GermanyCover design: Erich Kirchner, Heidelberg, GermanyTypesetting and reproduction of the figures:AM-productions GmbH, Wiesloch, Germany

Printed on acid-free paper24/3151/beu-göh 5 4 3 2 1 0

Essentials in Ophthalmology is a new review se-ries covering all of ophthalmology categorizedin eight subspecialties. It will be published quar-terly; thus each subspecialty will be reviewedbiannually.

Given the multiplicity of medical publi-cations already available, why is a new seriesneeded? Consider that the half-life of medicalknowledge is estimated to be around 5 years.Moreover, it can be as long as 8 years betweenthe description of a medical innovation in apeer-reviewed scientific journal and publica-tion in a medical textbook.A series that narrowsthis time span between journal and textbookwould provide a more rapid and efficient trans-fer of medical knowledge into clinical practice,and enhance care of our patients.

For the series, each subspecialty volume com-prises 10–20 chapters selected by two distin-guished editors and written by internationallyrenowned specialists. The selection of thesecontributions is based more on recent and note-

worthy advances in the subspecialty than onsystematic completeness. Each article is struc-tured in a standardized format and length, withcitations for additional reading and an appro-priate number of illustrations to enhance im-portant points. Since every subspecialty volumeis issued in a recurring sequence during the 2-year cycle, the reader has the opportunity tofocus on the progress in a particular subspecial-ty or to be updated on the whole field. The clin-ical relevance of all material presented will bewell established, so application to clinical prac-tice can be made with confidence.

This new series will earn space on the book-shelves of those ophthalmologists who seek tomaintain the timeliness and relevance of theirclinical practice.

G. K. KrieglsteinR. N. WeinrebSeries Editors

Foreword

This volume of the series Essentials in Ophthal-mology aims to present recent developmentsregarding the cornea, with a discussion of diag-nostic measures and particular emphasis beingplaced on treatment.

The therapeutic repertoire for surface disor-ders has increased considerably within the pastdecade. The chapter by Geerling and Hartwigreviews the application of autologous serum eyedrops for this indication. Dua and coworkers intheir chapters first help us to understand thelimitations of amniotic membrane transplanta-tion and then give us an overview regarding thevarious possibilities of limbal stem cell trans-plantation. Güell and coworkers provide anillustration of the potential of limbal stem celltransplantation following ex-vivo expansion.

Anterior lamellar keratoplasty is presentedby Melles as a promising technique for patientswith low grade keratoconus or opaque corneaswith healthy endothelium. Endothelial immunereactions may be avoided using this procedure.In penetrating keratoplasty, immune reactionsand astigmatism in patients still represent themajor postoperative problems. Slegers andcoworkers illustrate immunopathological phe-nomena, clinical features and risk factors ofgraft rejection. Antiangiogenic procedures arediscussed by Cursiefen and Kruse which mightcontribute to minimizing the immunologicalproblem in the future. Böhringer and coworkers

outline modern matching techniques (major –triplet, minor matching) as a prophylactic im-munological measure for patients with normal-risk as well as those with high-risk keratoplasty.Modern strategies of systemic immunosuppres-sion following high-risk penetrating kerato-plasty are presented by Reis and coworkers.An overview is given by Seitz of how best totrephine the cornea in penetrating keratoplastyin order to minimize postoperative astigma-tism. Watson and Daya provide an expert opin-ion on a serious postoperative complicationfollowing LASIK, i.e. infection.

Adenoviral corneal opacities are still a thera-peutic challenge. Hillenkamp and coworkersprovide an update on the different treatmentregimes. Guthoff and coworkers present con-focal microscopy of the cornea as a valuable tool for the in-vivo description of corneal struc-tures on a cellular basis. Finally, a overview ofocular allergic disease is given by Manzouri andcoworkers.All the topics covered by the book have a directclinical relevance, and we hope they will make asignificant contribution to the development ofoptimal diagnostic and therapeutic proceduresfor patients with disease of the cornea.

T. ReinhardD.F.P. Larkin

Preface

Chapter 1Autologous Serum Eyedrops for Ocular Surface DisordersGerd Geerling, Dirk Hartwig

1.1 Introduction . . . . . . . . . . . . . . . . . . . 21.1.1 The Rationale for Using Serum

in Ocular Surface Disorders . . . . . . 21.1.2 Legal Aspects . . . . . . . . . . . . . . . . . . . 31.2 Production and Application . . . . . . 41.2.1 Important Parameters

of the Production Process . . . . . . . . 41.2.2 Current Standard Operating

Procedures Used at the University of Lübeck . . . . . . . . . . . . . . . . . . . . . . 6

1.2.3 Quality Control . . . . . . . . . . . . . . . . . 81.3 Clinical Results . . . . . . . . . . . . . . . . . 81.3.1 Persistent Epithelial Defects . . . . . . 81.3.2 Dry Eye . . . . . . . . . . . . . . . . . . . . . . . . 101.3.3 Other Indications . . . . . . . . . . . . . . . 131.4 Complications . . . . . . . . . . . . . . . . . . 151.5 Alternative Blood Products

for the Treatment of Ocular Surface Disease . . . . . . . . 16

1.5.1 Umbilical Chord Serum . . . . . . . . . . 161.5.2 Albumin . . . . . . . . . . . . . . . . . . . . . . . 171.5.3 Plasma and Platelets . . . . . . . . . . . . . 17References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

Chapter 2Controversies and Limitations of Amniotic Membrane in Ophthalmic SurgeryHarminder S. Dua, V. Senthil Maharajan,Andy Hopkinson

2.1 Introduction . . . . . . . . . . . . . . . . . . . 212.2 Proposed Mechanisms of Action

of the Amniotic Membrane . . . . . . . 222.2.1 Amnion Structure . . . . . . . . . . . . . . . 222.2.2 Amnion Composition . . . . . . . . . . . 22

2.3 Intra and Inter Donor Variations of the Membrane . . . . . . . . . . . . . . . . 24

2.4 Processing and Preservation of the Membrane . . . . . . . . . . . . . . . . 25

2.5 Clinical Studies and Outcomes(Definitions of Success and Grading of Disease Severity) . . . . . . . . . . . . . 27

2.6 Efficacy of Membrane in Relation to Other Established Techniques and Options . . . . . . . . . . . . . . . . . . . . 30

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

Chapter 3Transplantation of Limbal Stem CellsHarminder S. Dua

3.1 Introduction . . . . . . . . . . . . . . . . . . . 363.2 Stem Cells . . . . . . . . . . . . . . . . . . . . . . 363.2.1 Definition . . . . . . . . . . . . . . . . . . . . . . 363.2.2 Characteristics of Stem Cells . . . . . 363.2.3 The Stem Cell ‘Niche’ . . . . . . . . . . . . 373.3 Limbal Stem Cells . . . . . . . . . . . . . . . 373.3.1 The Clinical Evidence . . . . . . . . . . . . 373.3.2 The Scientific Evidence . . . . . . . . . . 393.4 Limbal Stem Cell Deficiency . . . . . . 413.4.1 Causes of Limbal Stem Cell

Deficiency . . . . . . . . . . . . . . . . . . . . . . 413.4.2 Effects of Limbal

Stem Cell Deficiency . . . . . . . . . . . . . 413.4.3 Diagnosis of Stem Cell Deficiency . 443.5 Limbal Transplant Surgery . . . . . . . 453.5.1 Principles . . . . . . . . . . . . . . . . . . . . . . 453.5.2 Preoperative Considerations . . . . . . 453.6 Surgical Techniques . . . . . . . . . . . . . 463.6.1 Sequential Sector Conjunctival

Epitheliectomy (SSCE) . . . . . . . . . . . 463.6.2 Auto-limbal Transplantation . . . . . 483.6.3 Allo-limbal Transplantation . . . . . . 493.6.4 Adjunctive Surgery . . . . . . . . . . . . . . 513.6.5 Postoperative Treatment . . . . . . . . . 52References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

Contents

Chapter 4Limbal Stem Cell CultureJosé L. Güell, Marta Torrabadella,Marta Calatayud, Oscar Gris,Felicidad Manero, Javier Gaytan

4.1 Introduction . . . . . . . . . . . . . . . . . . . 574.2 Epithelial Phenotype . . . . . . . . . . . . 584.3 Preparation of Human Amniotic

Membrane . . . . . . . . . . . . . . . . . . . . . 584.4 Culture of Explanted Tissue . . . . . . 594.5 Tissue Procurement . . . . . . . . . . . . . 594.6 Preliminary Clinical Experience . . . 604.6.1 Principles for Taking the Biopsy . . . 604.6.2 Advantages of Limbal

Stem Cell Culture . . . . . . . . . . . . . . . 614.7 Case Report . . . . . . . . . . . . . . . . . . . . 614.8 Future Standard Staging Approach

for Ocular Surface Reconstruction . 64References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64

Chapter 5Deep Anterior Lamellar KeratoplastyGerrit R.J. Melles

5.1 Introduction . . . . . . . . . . . . . . . . . . . 655.2 Main Drawbacks

of Conventional DALK . . . . . . . . . . . 655.3 Different Concepts . . . . . . . . . . . . . . 655.4 Important Preoperative

Considerations . . . . . . . . . . . . . . . . . 665.5 Psychological Preparation

of the Patient . . . . . . . . . . . . . . . . . . . 675.6 Choice of DALK

Surgical Technique . . . . . . . . . . . . . . 675.7 Clinical Results . . . . . . . . . . . . . . . . . 68References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70

Chapter 6Corneal Transplant RejectionT.P.A.M. Slegers, M.K. Daly, D.F.P. Larkin

6.1 Introduction . . . . . . . . . . . . . . . . . . . 736.2 Incidence . . . . . . . . . . . . . . . . . . . . . . 746.3 Factors Predisposing

to Corneal Graft Rejection . . . . . . . 746.4 Clinical Features . . . . . . . . . . . . . . . . 756.5 Histopathology . . . . . . . . . . . . . . . . . 75

6.6 Immunopathological Mechanisms 766.6.1 Immune Privilege

and Its Breakdown . . . . . . . . . . . . . . 766.6.2 Afferent Arm

of the Allogeneic Response . . . . . . . 766.6.3 Efferent Arm

of the Allogeneic Response . . . . . . . 776.7 Treatment of Rejection . . . . . . . . . . . 776.8 Prevention of Rejection . . . . . . . . . . 776.8.1 Immunosuppression . . . . . . . . . . . . . 776.8.2 HLA Matching . . . . . . . . . . . . . . . . . . 786.9 Future Prospects . . . . . . . . . . . . . . . . 78References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Chapter 7New Aspects of Angiogenesis in the CorneaClaus Cursiefen, Friedrich E. Kruse

7.1 Introduction . . . . . . . . . . . . . . . . . . . 837.2 “Angiogenic Privilege of the Cornea”

or “How Does the Normal CorneaMaintain Its Avascularity?” . . . . . . . 84

7.3 Corneal (Hem)angiogenesis . . . . . . 857.3.1 General Mechanisms of Corneal

(Hem)angiogenesis . . . . . . . . . . . . . . 857.3.2 Common Causes

of Corneal (Hem)angiogenesis . . . . 867.3.3 Clinical Consequences

of Corneal Hemangiogenesis . . . . . 867.3.4 Corneal Hemangiogenesis

After Keratoplasty . . . . . . . . . . . . . . . 877.3.5 Corneal Angiogenesis

Due to Contact Lens Wear . . . . . . . . 907.3.6 Angiogenesis as a Cause

of Disease Progression,not a Sequel (Herpetic Keratitis) . . 91

7.3.7 Surgery in Vascularized Corneas . . 917.4 Corneal Lymphangiogenesis . . . . . . 917.4.1 Mechanisms of Corneal

Lymphangiogenesis . . . . . . . . . . . . . 917.4.2 Importance of Lymphangiogenesis

for Induction of Alloimmunity After Keratoplasty . . . . . . . . . . . . . . . 94

7.4.3 Non-immunological Effects of Corneal Lymphangiogenesis . . . 94

7.5 Antiangiogenic Therapy at the Cornea . . . . . . . . . . . . . . . . . . . 95

7.5.1 Established and Novel Antiangiogenic Therapies . . . . . . . . 95

X Contents

7.5.2 Novel Antihemangiogenic and Antilymphangiogenic Therapies to Improve Graft Survival AfterKeratoplasty . . . . . . . . . . . . . . . . . . . . 98

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98

Chapter 8Histocompatibility Matching in Penetrating KeratoplastyDaniel Böhringer, Rainer Sundmacher,Thomas Reinhard

8.1 Introduction . . . . . . . . . . . . . . . . . . . 1018.1.1 Immune Reactions Constantly

Threaten Graft Survival . . . . . . . . . . 1018.1.2 Major Transplantation Antigens

(HLA) . . . . . . . . . . . . . . . . . . . . . . . . . 1028.1.3 Minor Transplantation Antigens . . 1058.2 Time on the Waiting List Associated

with Histocompatibility Matching . 1068.2.1 Waiting Time Variance Has Been

a Barrier to HistocompatibilityMatching . . . . . . . . . . . . . . . . . . . . . . . 106

8.2.2 Algorithm for Predicting the Time on the Waiting List . . . . . . . . . . . . . . 106

8.3 Recommended Clinical Practice . . . 107References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

Chapter 9Current Systemic ImmunosuppressiveStrategies in Penetrating KeratoplastyAlexander Reis, Thomas Reinhard

9.1 Introduction . . . . . . . . . . . . . . . . . . . 1099.2 Immunology . . . . . . . . . . . . . . . . . . . 1099.2.1 Acute Rejection . . . . . . . . . . . . . . . . . 1109.2.2 Major Histocompatibility Complex 1109.2.3 Chronic Rejection . . . . . . . . . . . . . . . 1109.3 Normal-Risk Versus High-Risk

Transplantation . . . . . . . . . . . . . . . . . 1119.3.1 Normal-Risk Transplantation . . . . . 1119.3.2 High-Risk Transplantation . . . . . . . 1119.3.3 Rationale for Systemic

Immunosuppression . . . . . . . . . . . . 1119.3.4 Why Is Immunomodulation

with Topical Steroids Not Sufficient To Prevent Immunologic GraftRejection in High-Risk Patients? . . 111

9.4 Immunosuppressive Agents . . . . . . 1129.4.1 History . . . . . . . . . . . . . . . . . . . . . . . . 1129.4.2 Corticosteroids . . . . . . . . . . . . . . . . . 1149.4.3 Cyclosporine A (CSA, Sandimmun,

Sandimmun Optoral,Sandimmun Neoral) . . . . . . . . . . . . . 114

9.4.4 Tacrolimus (FK506, Prograf) . . . . . . 1149.4.5 Mycophenolate Mofetil

(MMF, CellCept, Myfortic) . . . . . . . . 1159.4.6 Rapamycin (Sirolimus, Rapamune) 1159.4.7 RAD (Everolimus, Certican) . . . . . . 1169.4.8 FTY 720 . . . . . . . . . . . . . . . . . . . . . . . 1169.4.9 Biologic Agents . . . . . . . . . . . . . . . . . 1179.5 Guidelines for Practitioners . . . . . . 1179.5.1 Preoperative Evaluation . . . . . . . . . . 1179.5.2 How To Use Cyclosporine

in High-Risk Corneal Transplantation . . . . . . . . . . . . . . . . . 118

9.5.3 How To Use MMF in High-Risk Corneal Transplantation . . . . . . . . . 118

9.5.4 How To Use Rapamycin in High-Risk Corneal Transplantation . . . . . . . . . . . . . . . . . 118

9.5.5 How To Use Tacrolimus in High-Risk Corneal Transplantation . . . . . . . . . . . . . . . . . 118

9.5.6 Combination Therapies . . . . . . . . . . 1199.6 Conclusion . . . . . . . . . . . . . . . . . . . . . 119References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Chapter 10Trephination in Penetrating KeratoplastyBerthold Seitz, Achim Langenbucher,Gottfried O.H. Naumann

10.1 Introduction . . . . . . . . . . . . . . . . . . . 12310.2 Astigmatism and Keratoplasty . . . . 12410.2.1 Definition of Post-keratoplasty

Astigmatism . . . . . . . . . . . . . . . . . . . . 12410.2.2 Reasons for Astigmatism

After Keratoplasty . . . . . . . . . . . . . . . 12510.2.3 Prevention/Prophylaxis

of Astigmatism After Keratoplasty 12910.3 Trephination Techniques . . . . . . . . . 13010.3.1 Principal Considerations . . . . . . . . . 13110.3.2 Conventional Mechanical Trephines 13910.3.3 Nonmechanical Laser Trephination 14410.4 Concluding Remarks . . . . . . . . . . . . 148References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149

Contents XI

Chapter 11Infective Complications Following LASIKAdam Watson, Sheraz Daya

11.1 Introduction . . . . . . . . . . . . . . . . . . . 15311.2 Frequency and Presentation . . . . . . 15311.3 Characteristics . . . . . . . . . . . . . . . . . 15411.4 Differential Diagnosis . . . . . . . . . . . 15411.4.1 Diffuse Lamellar Keratitis

(DLK,“Sands of the Sahara”) . . . . . 15411.4.2 Steroid-Induced Intraocular

Pressure Elevation with Flap Oedema (Pseudo-DLK) . 155

11.5 Management . . . . . . . . . . . . . . . . . . . 15511.5.1 Flap Lift . . . . . . . . . . . . . . . . . . . . . . . 15511.5.2 Specimen Taking . . . . . . . . . . . . . . . . 15511.5.3 Treatment . . . . . . . . . . . . . . . . . . . . . . 15711.5.4 No Improvement . . . . . . . . . . . . . . . . 15711.6 Special Considerations . . . . . . . . . . . 15811.6.2 Fungal Keratitis . . . . . . . . . . . . . . . . . 15811.6.3 Viral Keratitis . . . . . . . . . . . . . . . . . . 15911.7 Visual Outcome . . . . . . . . . . . . . . . . . 15911.8 Management of Sequelae . . . . . . . . . 15911.9 Prevention . . . . . . . . . . . . . . . . . . . . . 160References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

Chapter 12Treatment of Adenoviral KeratoconjunctivitisJost Hillenkamp, Rainer Sundmacher,Thomas Reinhard

12.1 Introduction . . . . . . . . . . . . . . . . . . . 16312.1.1 Etiology and Clinical Course

of Ocular Adenoviral Infection . . . . 16312.2 Socioeconomic Aspect . . . . . . . . . . . 16612.3 Treatment . . . . . . . . . . . . . . . . . . . . . . 16612.3.1 Treatment of the Acute Phase . . . . . 16612.3.2 Treatment of the Chronic Phase . . . 16812.3.3 Prophylaxis . . . . . . . . . . . . . . . . . . . . 16912.4 Conclusion and Outlook . . . . . . . . . 17012.5 Current Clinical Practice

and Recommendations . . . . . . . . . . 170References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170

Chapter 13In Vivo Micromorphology of the Cornea:Confocal Microscopy Principles and Clinical ApplicationsRudolf F. Guthoff, Joachim Stave

13.1 Introduction . . . . . . . . . . . . . . . . . . . 17313.2 Principle of In Vivo Confocal

Microscopy Based on the Laser-Scanning Technique . 174

13.2.1 Slit-Scanning Techniques . . . . . . . . . 17513.2.2 Laser-Scanning Microscopy

and Pachymetry . . . . . . . . . . . . . . . . 17613.2.3 Fundamentals of Image Formation

in In Vivo Confocal Microscopy . . . 17913.3 General Anatomical

Considerations . . . . . . . . . . . . . . . . . 18013.4 In Vivo Confocal Laser-Scanning

Microscopy . . . . . . . . . . . . . . . . . . . . . 18113.4.1 Confocal Laser-Scanning Imaging

of Normal Structures . . . . . . . . . . . . 18213.5 Clinical Findings . . . . . . . . . . . . . . . . 19013.5.1 Dry Eye . . . . . . . . . . . . . . . . . . . . . . . . 19013.5.2 Meesmann’s Dystrophy . . . . . . . . . . 19013.5.3 Epithelium in Contact

Lens Wearers . . . . . . . . . . . . . . . . . . . 19213.5.4 Epidemic Keratoconjunctivitis . . . . 19513.5.5 Acanthamoeba Keratitis . . . . . . . . . 19613.5.6 Corneal Ulcer . . . . . . . . . . . . . . . . . . 19913.5.7 Refractive Corneal Surgery . . . . . . . 20013.6 Future Developments . . . . . . . . . . . . 20113.6.1 Three-Dimensional Confocal

Laser-Scanning Microscopy . . . . . . . 20113.6.2 Functional Imaging . . . . . . . . . . . . . 203References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206

XII Contents

Chapter 14Allergic Eye Disease: Pathophysiology,Clinical Manifestations and TreatmentBita Manzouri, Thomas Flynn,Santa Jeremy Ono

14.1 Introduction . . . . . . . . . . . . . . . . . . . 20914.2 Pathophysiology . . . . . . . . . . . . . . . . 21014.2.1 Type I Hypersensitivity . . . . . . . . . . 21014.2.2 Ocular Inflammatory Reaction:

Late Phase . . . . . . . . . . . . . . . . . . . . . 21114.2.3 Non-specific Conjunctival

Hyperreactivity . . . . . . . . . . . . . . . . . 21114.2.4 T-Cell-Mediated Hypersensitivity

in Allergic Eye Disease . . . . . . . . . . . 21214.3 Clinical Syndromes

of Allergic Eye Disease . . . . . . . . . . . 21214.3.1 Seasonal Allergic Conjunctivitis . . . 214

14.3.2 Perennial Allergic Conjunctivitis . . 21414.3.3 Vernal Keratoconjunctivitis . . . . . . 21414.3.4 Atopic Keratoconjunctivitis . . . . . . 21514.3.5 Giant Papillary Conjunctivitis . . . . 21714.4 Treatment of Allergic Eye Disease . 21714.4.1 Antihistamines . . . . . . . . . . . . . . . . . 21814.4.2 Mast Cell Stabilizing Agents . . . . . . 21814.4.3 Dual-Acting Agents . . . . . . . . . . . . . 21914.4.4 Non-steroidal Anti-inflammatory

Drugs (NSAIDs) . . . . . . . . . . . . . . . . 21914.4.5 Topical Corticosteroids . . . . . . . . . . 21914.4.6 Calcineurin Inhibitors . . . . . . . . . . . 22014.4.7 Future Drug Developments . . . . . . . 22014.5 Conclusion . . . . . . . . . . . . . . . . . . . . . 221References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . 225

Contents XIII

Daniel Böhringer, Dr.University Eye Hospital Killianstrasse 5, 79106 Freiburg, Germany

Marta Calatayud, MDDiagonal 419 6º1a, 08008 Barcelona, Spain

Claus Cursiefen, Dr.Department of OphthalmologyFriedrich-Alexander University Erlangen-NürnbergSchwabachanlage 6, 91054 Erlangen, Germany

M.K. Daly, MDCornea & External Diseases ServiceMoorfields Eye HospitalLondon, EC1V 2PD, UK

Sheraz Daya, MDEye Bank, Queen Victoria Hospital NHS TrustEast Grinstead, West Sussex, RH19 3DZ, UK

Harminder S. Dua, MD, PhDChair and Professor of Ophthalmology University of Nottingham, Eye ENT Centre Queens Medical CentreDerby Road, Nottingham, NG7 2UH, UK

Thomas Flynn, MRCOphth.Department of Ocular Immunology Institute of Ophthalmology University College London, Bath Street London, EC1V 9EL, UK

Javier Gaytan, MDInstituto de Microcirugia, Ocular de Barcelona Munner 10, 08022 Barcelona, Spain

Gerd Geerling, Prof. Dr.Department of Ophthalmology University of Würzburg Josef-Schneider-Strasse 11 97080 Würzburg, Germany

Oscar Gris, MDInstituto de Microcirugia, Ocular de Barcelona Munner 10, 08022 Barcelona, Spain

José L. Güell, MDAssociate Professor of OphthalmologyInstituto de Microcirurgia Ocular de Barcelona C. Munner, 10, 08022 Barcelona, Spain

Rudolf F. Guthoff, Prof. Dr.University Eye Hospital Rostock Doberaner Strasse 140 18057 Rostock, Germany

Dirk Hartwig, Dr.Institute of Immunology and Transfusion Medicine University of Lübeck, Ratzeburger Allee 160 23538 Lübeck, Germany

Jost Hillenkamp, Dr.Eye Hospital of the University of Regensburg Franz-Josef Strauss Allee 11, 93042 Regensburg Germany

Andy Hopkinson, PhDDivision of Ophthalmology and Visual Sciences, Eye ENT Centre Queens Medical Centre, University Hospital Derby Road, Nottingham, NG7 2UH, UK

Contributors

Friedrich E. Kruse, Dr.Department of Ophthalmology University of Erlangen-Nürnberg Schwabachanlage 6, 91054 Erlangen, Germany

Achim Langenbucher, Priv.-Doz. Dr.Department of Ophthalmology University of Erlangen-Nürnberg Schwabachanlage 6, 91054 Erlangen, Germany

D.F.P. Larkin, MD, MRCPF, FRCSCornea & External Diseases Service Moorfields Eye Hospital London, EC1V 2PD, UK

V. Senthil Maharajan, MS, FRCSDivision of Ophthalmology and Visual Sciences Eye ENT Centre Queens Medical Centre, University Hospital Derby Road, Nottingham, NG7 2UH, UK

Felicidad Manero, MDInstituto de Microcirugia Ocular de Barcelona Munner 10, 08022 Barcelona, Spain

Bita Manzouri, MRCOphth.Department of Ocular Immunology Institute of Ophthalmology University College London Bath Street, London, EC1V 9EL, UK

Gerrit R.J. Melles, MDNetherlands Institute for Innovative Ocular Surgery Laan Op Zuid 3903071 AA Rotterdam, The Netherlands

Gottfried O.H. Naumann, Prof. Dr.Department of Ophthalmology University of Erlangen-Nürnberg Schwabachanlage 6, 91054 Erlangen, Germany

Santa Jeremy Ono, PhDDepartment of Ocular Immunology Institute of Ophthalmology University College London 11–43 Bath Street, London, EC1V 9EL, UK

Thomas Reinhard, Prof. Dr.University Eye Hospital Killianstrasse 5, 79106 Freiburg, Germany

Alexander Reis, Priv.-Doz. Dr.Reis Medical Institution Est.Landstrasse 310, 9495 Triesen Principality of Liechtenstein

Berthold Seitz, Prof. Dr.Department of Ophthalmology University of Erlangen-Nürnberg Schwabachanlage 6, 91054 Erlangen, Germany

T.P.A.M. Slegers, MDDepartment of Ophthalmology University Medical Center Groningen 9700 RB Groningen, The Netherlands

Joachim Stave, Prof. Dr.University Eye Hospital Rostock Doberaner Strasse 140, 18057 Rostock Germany

Rainer Sundmacher, Prof. Dr.University Eye Hospital, Moorenstrasse 5 40225 Düsseldorf, Germany

Marta Torrabadella, MDPaseig Vall d’Hebron 119 08034 Barcelona, Spain

Adam Watson, FRANZCOCorneoplastic Unit Queen Victoria Hospital NHS Trust East Grinstead, West Sussex, RH19 3DZ, UK

XVI Contributors

Autologous Serum Eyedrops for Ocular Surface Disorders

Gerd Geerling, Dirk Hartwig

1

∑ The viability and function of the cornealand conjunctival epithelium is supportedby the antimicrobial, nourishing, mechani-cal and optical properties of tears, sincethese contain factors which promote prolif-eration, migration and differentiation of epithelial cells

∑ A lack of these epitheliotrophic factors,e.g. in aqueous tear deficiency, can compro-mise the ocular surface and result in seriousdisorders such as persistent epithelialdefects

∑ Pharmaceutical lubricants usually replacethe aqueous component of tears alone and may have little efficacy in improvingsurface disorders

∑ Serum has biomechanical and biochemicalproperties similar to normal tears

∑ In vitro cell culture experiments haveshown that the morphology, migration anddifferentiation of ocular surface epitheliaare better supported by serum than bypharmaceutical lubricants

∑ A number of protocols for the productionof serum eyedrops have been published,which vary considerably and can influencethe biochemical properties of this autolo-gous blood product

∑ Under EU legislation production of serumeyedrops requires a license to produceblood products by the appropriate nationalbody, which means that an extensive evalu-ation and documentation process must besuccessfully completed

∑ Alternatively the use of serum eyedrops isallowed on an intention to treat basis if pro-duction and application are all performedby the physician, i.e. in ophthalmic depart-

ments that have the necessary laboratoryfacilities and are able to admit the patientfor the duration of the treatment

∑ The production of serum eyedrops shouldbe well documented and measures for ap-propriate quality control (i.e. serological andmicrobiological tests) should be established

∑ The results of a considerable number ofclinical cohort studies have reported benefi-cial effects of its use for persistent epithelialdefects, severe dry eyes and other indica-tions

∑ The use of serum eyedrops implies the riskof transmission of infectious diseases, fromthe donor to other individuals involved in the production and application of theproduct

∑ In addition contamination of the dropperbottle and subsequent microbial keratitiscan occur

∑ Such complications can be largely avoidedby testing the patient for HIV, syphilis andhepatitis B and C prior to the blood dona-tion and testing every serum eyedrop batchfor bacterial contamination. In addition,the serum therapy can be combined withtopical antibiotics

∑ Due to these risks and the lack of ran-domised controlled trials, the use of serumeyedrops should still be considered experi-mental and informed consent should beobtained from every patient treated

∑ Alternative blood products with epithe-liotrophic potential such as plasma, plateletconcentrates or serum albumin should beevaluated since they are readily available asquality controlled blood derivatives fromblood banks on a routine basis

Core Messages

1.1Introduction

In recent years the use of eyedrops producedfrom autologous serum has gained wide accept-ance for the treatment of ocular surface disor-ders intractable to conventional medical thera-py. Such conditions include persistent epithelialdefects or severe dry eyes. Autologous serumwas first evaluated in 1984 by Fox et al. [9] insearch of an unpreserved lubricant, which wasnot available from pharmaceutical providers atthat time. However, it was Tsubota who repopu-larised their use when he described the epithe-liotrophic potential of serum for the ocular sur-face due to its content of growth factors andvitamins [46].

1.1.1The Rationale for Using Serum in Ocular Surface Disorders

The tear film supplies the ocular surface withmany nutrient and wound healing modulatingfactors such as fibronectin, vitamins or growthfactors, which support and modulate prolifera-tion, migration and differentiation of the con-junctival and corneal epithelium. These epithe-liotrophic factors are predominantly releasedinto the aqueous component of the tear film bythe main and accessory lacrimal glands as wellas conjunctival vessels, while glucose, elec-trolytes and amino acids are provided by theaqueous humour.

For example fibronectin, a disulphide glyco-protein that influences cell adhesion and migra-tion of the healing epithelium, predominantlyoriginates from plasma when conjunctivalblood vessels become more permeable duringan inflammatory reaction and the lacrimalgland itself secretes vitamins, neuropeptidesand growth factors such as substance P andepidermal growth factor (EGF) [36]. However,proteins of the aqueous tear film not only act asessential nutrients for the ocular surface epithe-lia, but also determine the biomechanical prop-erties of the tear lipid layer. Tear-lipocalin is anexample for this, which by acting as a transport

protein for retinol, supports goblet cell differen-tiation, but also reduces the surface tension of tears. As part of inflammatory processes,additional proteins, such as lactoferrin, serum-IgA and complement factors, are released intothe tears and support opsonisation and phago-cytosis of microbes by macrophages and lymphocytes. Tears thus not only have a lubri-cating and mechanical clearance function,but also epitheliotrophic and antimicrobialproperties.

If the carrier, i.e. the aqueous phase, or the epitheliotrophic factors of the tear film are diminished, the integrity of the surface epithelia can become disrupted and epithelialdefects evolve, which may persist and progress.Surgical attempts to rehabilitate the ocular sur-face in severe dry eyes also fail frequently [2, 22]unless sufficient substitute lubrication is pro-vided.

Summary for the Clinician

∑ The tear film has lubricant and nutrientproperties

∑ In severe dry eye it is not only the increasedbiomechanical stress, but also the lack ofepitheliotrophic factors that promotesdamage of the ocular surface

The ideal tear substitute would possess lubri-cant and nutrient properties. However – withfew exceptions – currently available productsare optimised for their biomechanical proper-ties only [49, 51]. Vitamin A, EGF and fibro-nectin have been used in vitro and in vivo toencourage epithelial wound healing, but due tostability concerns and limited clinical success,these single compounds have not become partof clinical routine management [15, 25, 29].Autologous blood offers a number of character-istic advantages:1. It contains a large number of substances also

present in tears, such as vitamin A, epithe-liotrophic and neurotrophic growth factors,immunoglobulins and fibronectin. Some ofthese factors are found in serum in higherconcentrations than in natural tears(Table 1.1) [20, 27, 29, 45].

2. Serum can be prepared as an autologousproduct and thus lacks antigenicity.

2 Chapter 1 Autologous Serum Eyedrops for Ocular Surface Disorders

3. Eyedrops from serum can be produced with-out preservatives and hence toxicity due toadditives is not an issue.

A wide range of quality controlled products canbe derived from full blood. These include notonly serum, but also, e.g. plasma, platelet richsuspensions or various protein fractions such asalbumin. Serum is the fluid component of fullblood that remains after clotting. It should notbe confused with plasma, which is obtainedwhen clotting is prevented by mixing a fullblood donation with an anticoagulant and re-moving all corpuscular elements by centrifuga-tion (see Sect. 1.5.3). Plasma thus does not con-tain the significant amount of platelet derivedgrowth factors such as EGF, PDGF and TGF-b,which are released into serum upon activationof the platelets during clotting. Platelet rich sus-pensions (platelet concentrates) are available asa standard blood product from blood banks,

while human serum albumin fractions can bepurchased from pharmaceutical companies.While all of these blood derived products arepotentially available as growth factor contain-ing solutions for topical application to the ocu-lar surface, serum has been predominantly usedin clinical studies.

The growth and migration promoting effectof serum on cell cultures in general and oncorneal epithelial cells is well documented [13,46]. In dose- and time-response experiments invitro, we found that serum maintained the mor-phology and supported proliferation of primaryhuman corneal epithelial cells better than un-preserved or preserved pharmaceutical tearsubstitutes [13]. It is also known that seruminduces mucin-1 expression in immortalised,conjunctival epithelial cells as a sign of higherdifferentiation [45] and that it increases tran-scription of RNA for NGF as well as TGF-b re-ceptors in human keratocytes [7]. The epithe-liotrophic properties are – again according to invitro tests – not reduced in patients sufferingfrom systemic autoimmune disease requiringsystemic immunosuppression [17].


∑ Plasma, platelet suspensions and albuminare available as blood derived productsfrom blood banks

∑ Serum is not a standard blood product,but can easily be produced on special request and currently is by far the most often used product in clinical studies

∑ Serum contains many epitheliotrophic factors also present in tears and can be prepared as an autologous unpreserved tear substitute that offers both lubricantand nutrient properties

1.1.2Legal Aspects

Autologous serum eyedrops are a blood prod-uct. The distribution of pharmaceuticals is reg-ulated by governmental laws in most countries.Although in the European Union the manufac-turing and distribution of pharmaceuticals isregulated by the individual country, several

1.1 Introduction 3

Table 1.1. Biochemical and biophysical propertiesof undiluted serum and normal, unstimulated human tears (EGF, epidermal growth factor; FGF, fibroblastgrowth factor; IGF, insulin like growth factor; NGF,neurotrophic growth factor; PDGF, platelet derivedgrowth factor; SP, substance P; TGF, transforminggrowth factor) [11, 15, 20, 48–52]

Tears Serum

pH 7.4 7.4

Osmolality 298±10 296

EGF (ng/ml) 0.2–3.0 0.5

TGF-b 2–10 6–33

NGF (pg/ml) 468.3±317.4 54.0

SP (pg/ml) 157.0±73.9 70.9±34.8

IGF-1 (ng/ml) 0.031±0.015 105

PGDF 0–1.33 15.5

Vitamin A (mg/ml) 0.02 46

Albumin (mg/ml) 0.023±0.016 35–53

Fibronectin (mg/ml) 21 205

Lactoferrin (ng/ml) 1,650±150 266

Lysozyme (mg/ml) 2.07±0.24 0.001

SIgA (mg/ml) 1,190±904 2,500

directives have been issued (1965/65, 1975/139,1975/318) which have been implemented in thelaws of each member state of the EU. Todayevery pharmaceutical product requires a mar-keting authorisation to be issued by a compe-tent authority of each state. Authorisation de-pends on the proof of efficacy in clinical trials,implementation of quality controls, reports ofadverse effects, proof of expert knowledge andother regulatory aspects. These criteria canprobably only be fulfilled by professional phar-maceutical manufacturers.

An exemption from the need to obtain mar-keting authorisation is granted if a physicianmanufactures a specific medical product byhimself or under his supervision with responsi-bility to treat his own patient on a named basis.This product has to be prepared according tothe doctor’s specifications and autologousserum eyedrops can therefore be produced onlyby the physician himself or by his staff. Howev-er, it remains the physician’s responsibility thatmanufacturing and application are performedcorrectly. Since even stricter regulations mayespecially exist for blood products in individualstates, every physician producing autologousserum eyedrops needs to inform himself aboutspecific national regulations.

In the United States, producers of drugs andmedical devices have to be registered with theFood and Drug Administration (FDA). Similarto EU regulations, registration is not necessary“for practitioners licensed by law to prescribe oradminister drugs or devices and who manufac-ture, prepare, propagate, compound, or processdrugs or devices solely for use in the course oftheir professional practice”.Again, special regu-lations on testing and approval of drugs by theFDA or for using blood products need to beevaluated by the practitioner.


∑ Serum eyedrops are a blood product thatcan be produced on a named patient basisaccording to the doctor’s specifications

∑ It remains the physician’s responsibilitythat manufacturing and application areperformed correctly

∑ Every patient should give their informedconsent before the production of autolo-gous serum eyedrops is initiated

∑ Quality control measures must be imple-mented for the production as well as theapplication

∑ Serum should only be applied under thesupervision of the prescribing doctor

1.2Production and Application

1.2.1Important Parameters of the Production Process

Although the complex composition of fullblood will certainly vary between individuals, itis also known that a number of production pa-rameters can significantly influence the bio-chemical composition of blood derived prod-ucts. These critical steps in the production ofserum eyedrops should therefore be standard-ised.

These include:∑ Clotting phase: duration and temperature∑ Centrifugation: centrifugal force

and duration∑ Dilution: dilution factor and diluent∑ Storage: container, temperature, duration

In the absence of any controlled clinical trialevaluating the impact of such differences in theproduction, in vitro models have been helpful inassessing a large number of protocol variations.The published clinical studies often fail to men-tion important parameters such as for how longthe blood was allowed to clot before centrifuga-tion. If the full blood sample is centrifugedbefore the clotting process is completed, therelease of platelet derived factors is reduced. Wehave established that the concentration of EGF,HGF, and TGF-b are higher after a 2-h clottingtime when compared with paired full bloodsamples that were allowed to clot only for 15 minand this was associated with a trend towardsbetter corneal epithelial cell proliferation [24].Thus we allow the full blood donation to clot forat least 2 h at room temperature. However, a


48-h period of storage at 4 °C, to allow trans-portation of full blood donations from periph-eral ophthalmic departments to a centralisedproduction unit, seems an equally acceptableprocedure [53].

The volume retrieved from a given blood do-nation as well as its biochemical compositionare also influenced by the centrifugation, whichis determined by the centrifugal g-force and thetime used to spin a sample. The g-force itself de-pends upon the revolutions of the rotor perminute (rpm) as well as on the diameter of therotor. Thus g-force and not rpm is the parame-ter that should be stated in a protocol. The g-force in the studies published so far – if men-tioned at all – probably varies by at least 1 log.A higher g-force not only helps to yield a largervolume of serum from a full blood sample(Fig. 1.1), but also reduces membranous plateletremnants in the supernatant, which in high con-

centrations have been shown to induce apopto-sis [5]. Also a higher g-force can result in a low-er concentration of TGF-b1. As TGF-b is able toslow down epithelial wound healing, Tsubotasuggested diluting the serum 1:5 with saline,which, however, reduces the concentration ofother growth factors, such as EGF, that areproven to support proliferation of corneal ep-ithelial cells. Combinations of 1,500 rpm – in anaverage size centrifuge equal to about 300 g – to4,000 g (ca. 5,000 rpm) for 5–20 min have beenused. A 15-min centrifugation at 3,000 g resultsin good separation of serum and blood clot,without inducing haemolysis [41].

It is obvious that the dilution of the obtainedserum sample to the final concentration in theeyedrops determines the concentration of ep-itheliotrophic factors. In clinical studies, 20%,33%, 50% or 100% have been used. Since theprotocols for the production of serum eyedropsalso varied in other parameters, there is no clearclinical evidence to favour any specific concen-tration. However, in vitro experiments showthat cell proliferation is best supported at a 20%concentration of serum. It was also shown thatthe type of diluent has an impact and that BSSrather than saline should be used [24].

In some indications, e.g. dry eye, autologousserum eyedrops are applied for many months.They are also usually produced without preser-vatives or stabilising additives to minimise therisk of drug induced toxicity. Since the produc-tion is labour intensive, a large number ofaliquot samples is prepared from a single blooddonation to keep the number of donation andprocessing efforts limited. To preserve the activ-ity of the biological substances thought to bebeneficial for the ocular surface, the drops canbe refrigerated or stored frozen. The concentra-tion of growth factors, vitamin A and fibro-nectin in pure and diluted serum was found toremain stable for at least 3 months if stored at–20 °C and for 1 month if stored at 4 °C. How-ever, it is known that many protein concentra-tions in tears are reduced if stored for severalweeks at 4 °C. While in developed countries ac-cess to freezers is rarely a problem, it seemspreferable to store unused daily dosage vials ofserum frozen [45, 38]. From this stock one con-tainer is then removed every morning and kept

1.2 Production and Application 5

Fig. 1.1. Demonstration of the influence of the g-force on the volume of serum obtained from 50 ml offull blood centrifuged 2 h after donation: A at 3,000 gfor 15 min; B at 500 g for 5 min

refrigerated at 4 °C until it is discarded at theend of the day. Alternatively dilution of theserum with chloramphenicol 0.5%, which hasfew toxic side effects, has also been advocated toallow the use of dropper bottles for up to 1 week.


∑ The list of protocol variations for the pro-duction of autologous serum eyedrops islong

∑ The biochemical composition of serumdepends on the time and conditions ofclotting, the centrifugation, dilution andstorage

∑ No controlled clinical trial has determinedso far which of the protocols publishedoffers the best epitheliotrophic support

∑ In the absence of such trials, the productionprotocol has been optimised in vitro

∑ The stock of serum eyedrops can be storedfrozen for several months to preserve the biologically active components of theproduct

1.2.2Current Standard Operating ProceduresUsed at the University of Lübeck

Following the principles of Good Manufactur-ing Practice and based on extensive evaluationin vitro, the following standard operating pro-cedures are currently used at the University ofLübeck (Fig. 1.2) [24, 19].

Patients are assessed for their suitability todonate according to the guidelines of the Bun-desärztekammer and Paul-Ehrlich Institute forblood donation and use of blood products. Thisrequires them to be in reasonably good health,with no significant cardiovascular or cere-brovascular disease, and free of bacterial infec-tion. Anaemia (Hb<11 g/dl) is a relative con-traindication. If only a small amount of blood(50–100 ml) is taken, mild anaemia or circulato-ry disorders need not be considered contraindi-cations. To minimise the danger of bacterialcontamination, no blood should be taken frompatients suffering from suspected septicaemia.To exclude transmission of infection, patientsmust be tested for hepatitis-B/-C, syphilis and

HIV serology (HbsAg; antibodies to HCV, HIV-I/-II, HIV-NAT, syphilis; HCV-NAT) beforeblood is donated for the production of eye-drops. A positive serology excludes the patientfrom the donation of autologous blood forserum eyedrop production. Prior to venisec-tion, the patient must be informed in writingabout the planned therapy, its experimental na-ture, the risks involved (e.g. bacterial contami-nation) and alternative methods of treatment.The patient’s consent should be obtained andkept with the notes.

Venipuncture is performed at the antecubitalfossa under aseptic conditions. Depending onthe expected duration of treatment, 100–200 mlof whole blood is collected into sterile contain-ers. For larger volumes a sterile blood packwithout anticoagulant can be used to collect upto 470 ml. A 100-ml donation of whole bloodwill yield 30–35 ml of serum, which diluted to20% is sufficient for at least 3 months of serumeyedrops 8 times daily. Larger volumes are rec-ommended in patients who require long-termtreatment, in order to minimise labour intensiveproduction. The containers are left standingupright for 2 h at room temperature to ensurecomplete clotting before they are centrifuged at3,000 g for 15 min. The supernatant serum isremoved under sterile conditions in a laminarair flow hood with sterile 50-ml disposable syringes. The volume retrieved is determinedand diluted 1:5 with sterile BSS. Gentle shakingensures homogenisation before portions of 2 mlare aliquoted through a 0.2-mm filter into steriledropper bottles. The effect of filtration has notbeen evaluated, but Fox recommends filtrationto remove fibrin strands, suspected to reducethe effect of serum eyedrops. The bottles aresealed and labelled with the name, date of birthof the patient, the date of production and the in-struction “Autologous blood serum for topicaluse in the eye. To be stored frozen and usedwithin 3 months after date of production. To bediscarded 24 hours after opening.” Two milli-litres of the solution is – as required by theEuropean Pharmacopoeia Addendum 2000 –sent for microbiological evaluation.

The product is available approximately 6 hafter venesection, but is only dispatched oncenegative serology and microbiology of donor


and product are confirmed. Usually the dropsare applied 8 times daily.A new bottle is openedeveryday. It is recommended to be stored at+4 °C and to be discarded after 16 h of use withregular household waste. The remaining bottlesare stored frozen (ideally at –20 °C) for up to3 months. If the domestic freezer has no ther-mometer, it is recommended to place one insideand control the temperature when taking a newvial out every day. If the temperature cannot beadjusted to about –20 °C, the dispensing doctormay consider recommendation of shorter stor-age episodes.

The costs of production – i.e. labour andconsumables alone – for a day’s dosage of serumeyedrops following this protocol are well below5 j. This is the approximate equivalent of onebottle of preserved pharmaceutical lubricant

and thus should be acceptable for the rare occa-sion where the ocular surface disease is so se-vere that the use of topical autologous serum isjustified [14].


∑ Every physician producing or prescribingautologous serum eyedrops should informhimself about specific national regulations

∑ Any production protocol should follow theprinciples of Good Manufacturing Practice

∑ The dropper bottles must be carefully labelled with the patient’s details and instructions for storage and use

∑ If produced without preservatives the dropsshould be kept frozen at –20 °C until the dayof use

1.2 Production and Application 7

Fig. 1.2. Standard manufacturing protocol for the preparation, storage and use of serum eyedrops. Parame-ters which influence the biochemical character of the product are also shown

1.2.3Quality Control

In order to guarantee quality, control measuresshould be initiated and if strictly interpretedserum eyedrops should be produced only bypersonnel supervised by the doctor directly incharge of the patient treated. Bacterial contam-ination of the product during the production aswell as from the application of serum eyedropsis a potential risk. Sterile manufacturing condi-tions, beginning with thorough skin disinfec-tion, are of the utmost importance. It is prefer-able that further processing is performed in aclosed system. To minimise the risk of infectionto third parties (e.g. production or nursingstaff), it is strongly recommended that thedonor is tested for HIV, HBV, HCV and syphilisbefore donation of larger amounts of blood arecollected for the production process itself. Forquality control purposes bacterial contamina-tion resulting from the production processneeds to be ruled out by a microbiologicalexamination of the product prior to initiation ofthe clinical application and a control systemmust be implemented to ensure that the productis only used when the microbiological and sero-logical tests are clear.

Prior to venipuncture and application, theidentity of the patient must be confirmed. Thepackaging and dropper bottles must be clearlylabelled with:∑ The patient’s name and date of birth∑ The name and address of the manufacturer∑ The date of manufacture and the date of ex-

piry∑ The instructions on how to store and use the

drops∑ A comment that the material is an autolo-

gous blood product, which is solely for appli-cation with the named patient

To minimise the variability of the product andto maximise the safety of its use, a written version of the standard operating procedures(SOP) should be established. Conscientiousdocumentation is indispensable for good med-ical and manufacturing practice. Each step rele-vant to manufacture as well as application (in-

cluding the dates of application and any un-wanted effect) should be recorded. From this itbecomes obvious that strict guidelines for goodmanufacturing, quality control and documenta-tion must be established and maintained priorand throughout the therapeutic use of auto-logous serum eyedrops. All steps of the produc-tion should be documented on a form which –together with the patient’s consent – is kept withthe patient’s notes.


∑ A written protocol of the standard operat-ing procedures should be established

∑ All production steps must be documentedon a SOP form

∑ Serum eyedrops should only be producedand released once the negative serology ofthe donor for hepatitis, syphilis and HIVand the negative microbiology of the prod-uct have been confirmed

1.3Clinical Results

Serum eyedrops have predominantly been usedfor persistent epithelial defects and severe dryeye, but also as a supportive measure in ocularsurface reconstruction and for a number ofother indications.

1.3.1Persistent Epithelial Defects

A persistent epithelial defect (PED) is defined asa defect of the corneal epithelium that – in theabsence of microbial keratitis – fails to healwithin the expected time course (e.g. 2 weeks)despite topical lubricants [26, 50]. A PED canoccur as a result of many different pathologies,including rheumatoid arthritis, neurotrophickeratopathy or dry eye [8]. “Success” of treat-ment is best defined as percentage of defectshealed in a given time or as total time to com-plete epithelial wound closure.


1.3.1.1Currently Available Published Data

In five prospective case series, 20% serumdiluted in 0.9% saline has been used 5–14 timesdaily for this indication (Table 1.2). In 1999Tsubota was the first to report a series of 16 eyes(15 patients) in whom the PEDs had persisteddespite medical treatment with lubricants or bandage contact lenses for a mean of7.2±9.4 months. Ten out of these 16 defectshealed completely within 4 weeks after initia-tion of therapy [11, 46]. Garcia-Jimenez reportedcomplete epithelial wound healing in 6 of 11 eyeswith persistent epithelial defects with healingbeginning within 3–4 weeks of treatment withserum eyedrops [1, 52]. In a group of 9 patientswith predominantly diabetic or postherpeticneurotrophic keratitis, all 12 epithelial defectshealed within 15.8±7.9 days and this was associ-ated in 9 eyes with an improvement of cornealsensitivity [27] (Fig. 1.3).

A different concentration of serum was usedin two other studies. Poon et al. substitutedunpreserved pharmaceutical lubricants with50–100% serum eyedrops and observed closureof a PED with a mean duration of 7.5±5.8(1–24) weeks in 9 of 15 eyes after 3.6±2.5 weeks(3 days–8 weeks) [33]. De Souza et al. treated 70

epithelial defects with undiluted serum hourlyin addition to routine medication, 45 of whichhad occurred early after penetrating kerato-plasty [8] and had persisted for a mean of15±17 days. Eighty-one percent of these defectswith a relative short history healed within14±12 days.

Healing generally starts within 2 weeks afterinitiation of the serum therapy [33]. So far nostudy has been able to show a correlation be-tween size or localisation of the defect with suc-cess or failure, but the older and deeper stromaldefects tended to heal less successfully. Also,when serum eyedrops are changed back topharmaceutical lubricants the epithelial defectsmay recur, as happened in 6 out of the 9 eyes inPoon’s and 9 out of 70 eyes in De Souza’s group.These figures are difficult to compare sincenone of the studies was placebo controlled andthe study population seems to differ significant-ly in terms of underlying pathogenesis andduration of the PED.


∑ Pathogenic factors that can be avoided ortreated, such as toxicity due to preservedeyedrops, steroids or active herpetic kerati-tis, should be ruled out before a PED istreated with serum eyedrops

1.3 Clinical Results 9

Table 1.2. Clinical studies using serum eyedrops to treat persistent epithelial defects. The production param-eters and results are given. Success is defined as percentage of eyes/patients with complete epithelialisation.Note that the scale used to measure these changes as well as the baseline level varied between the studies (NA,not applicable; NR, not reported; rpm, revelations per minute)

Author Con- Diluent Centri- Dura- Clott- Frequency Eyes Success centra- fugation tion ing of appli- objec-tion (g force) time cation (patients) tive

Alvarado 20% 0.9% NaCl 5000 RPM 10 min. NR NR 17 (14) 83%

De Souza 100% NA NR NR NR Hourly 70 (63) 81%

Garcia 20% 0.9% NaCl 5000 RPM 10 min. NR 10× 11 (11) 55%

Matsumoto 20% 0.9% NaCl 3000 RPM 10 min. NR 5–10× 14 (11) 100%

Poon 50–100% 0.5% 4000 RPM 10 min. 2 h 8× 15 (13) 60%chloram- (2200 G)phenicol

Tsubota 20% 0.9% NaCl 1500 RPM 5 min. NR 6–10× 16 (15) 63%

Young 20% 0.9% NaCl 1500 RPM 5 min. NR 6–14× 10 (10) 75%

∑ No signs of microbial keratitis should bepresent if serum application is consideredfor an epithelial defect, since the effect ofserum in this situation is unknown andmay support bacterial growth

∑ Twenty percent serum eyedrops are appliedapproximately every 2 h until the defect ishealed

∑ PEDs begin to heal generally within 2 weeksafter initiation of serum eyedrops

∑ Older and deeper stromal defects tend toheal less successfully

∑ If 20 % serum fails, a higher concentrationof serum may help to achieve epithelialisa-tion

∑ When serum eyedrops are changed back topharmaceutical lubricants the epithelialdefect may recur

1.3.2Dry Eye

Dry eye is a group of disorders, of diversepathogenesis, that share as common manifesta-tions signs and symptoms due to the interactionof both an abnormal tear film and an abnormalocular surface. It is subdivided into aqueous de-ficient and evaporative, i.e. lipid or mucin defi-cient dry eyes. Although it is believed to be oneof the most common ocular problems in theWestern world with an incidence of symptomsof dry eye in up to 14.6%, ocular surface changesare observed clinically in only 0.5%. Severeaqueous tear deficiency, however, can lead toblindness and serum eyedrops have been usedin this situation.


Fig. 1.3. Epithelial defect persisting unaltered for 2 weeks in the left eye of a female patient with severe aque-ous tear deficiency due to secondary Sjögren’s syndrome before (A, B) and 1 week after (C, D) treatment with20% autologous serum eyedrops 8 times daily. The defect started to heal within 2 days

A B

DC

In the dry eye “success of treatment” is moredifficult to define than in PEDs, and this can bedone either as subjective or objective improve-ment compared to baseline.“Subjective” successis determined as reduction of a score of symp-toms in a questionnaire of variable length.“Objective” success is either determined by re-duction of fluorescein or rose bengal positivestaining of the ocular surface or improvementof histologic parameters in impression cytol-ogy, although this is usually scored more or lessaccording to the subjective impression of anexaminer.

1.3.2.1Currently Available Published Data

Following the initial report of Fox in 1984, ittook 15 years until Tsubota in 1999 and subse-quently a number of studies reported on the useof serum eyedrops for dry eyes (Table 1.3). Foxtreated 30 eyes of 15 patients with 50% serum in0.9% NaCl and found that signs and symptomsimproved in all of the patients, until this med-ication was replaced by 0.5% serum or purediluent. Tsubota focussed on dry eyes due toSjögren’s syndrome. When treated with 20%serum 6–10 times daily for 4 weeks symptomsimproved by only 34%; however, fluoresceinand rose bengal staining decreased by 55% and68% of baseline, while tear break-up time re-mained unchanged [45]. Two groups of authorsdescribed similar findings in patients with dryeye due to graft-versus-host disease [32, 34] withsymptoms improving within days, but punctateepithelial staining improving only after months.Ogawa also reported that – although the aque-ous deficiency in his patients was less severe(£10 mm, Schirmer test) – in 50% symptomsrecurred while the patient continued to applyserum eyedrops and 43% required additionalpunctal occlusion. In a placebo-controlled pro-spective study of severe dry eyes with a meanSchirmer test score of less than 1 mm, 20%serum 6 times daily was not found to be signif-icantly more effective in improving symptomsand signs than 0.9% saline, which had beenused as diluent [40], although a trend towardsreduced fluorescein and rose bengal stainingwas observed after 2 months of treatment.

Two studies have reported the use of higherconcentrations of serum. Poon et al. found animprovement of subjective and objective crite-ria of severe dry eyes (Schirmer test <5 mm) inonly three out of eight eyes receiving 50%serum but all of three eyes receiving 100%serum [33]. Noble et al. compared the efficacy of3 months of autologous serum 50% diluted in0.9% saline in a prospective clinical crossovertrial against the previously used commercial lu-bricant and reported that 10 out 16 patients hadimproved symptoms [31], and that there wereimpression cytological findings in 6, no changein 10 and improvement in 9 of 25 treated eyes.

The efficacy seems to be dose dependentsince 94% of patients receiving eight applica-tions daily reported reduced symptoms com-pared to only 58% of those receiving four drops[39]. Overall the efficacy of serum eyedrops indry eyes varied between 30% and 100% forsymptomatic relief, between 39% and 61% forreduction of fluorescein and between 33% and68% for rose bengal positive staining. However,the variation in study population, productionand treatment protocol are again significant. Insome studies, serum was used as an additiverather than a substitute for lubricants and inothers therapeutic contact lenses or punctal oc-clusion were applied in addition to the serumeyedrop therapy. Comparison of the publisheddata is therefore difficult and it has to be con-cluded that no definite evidence supporting theuse of serum eyedrops in dry eyes is available sofar.


∑ Serum eyedrops should be reserved for themost severe cases of dry eye

∑ Punctal occlusion should be performedfirst

∑ Symptoms usually improve within days,but punctate epithelial staining maydecrease only after months of treatment

∑ The use of serum in dry eyes is not evidencebased. Using an optimised protocol for the production of serum eyedrops, a randomised controlled trial should be performed


12C

ha

pter 1A

utologous Serum Eyedrops for O

cular Surface Disorders

Table 1.3. Clinical studies using serum eyedrops to treat severe dry eyes. Success was defined either as number/percentage of all eyes/patients with improved/reduced mean baseline of objective [fluorescein (Fl) or rose bengal (RB) positive epitheliopathy or impression cytology (IPC)] or subjective (symptoms) scores.Note that the scale used to measure these changes as well as the baseline level varied between the studies (NR, not reported; NS, not significant; rpm, revelations perminute; replacement, frequency equivalent to previously applied pharmaceutical tear substitute)

Author Concen- Diluent Centri- Duration Clotting Frequency Eyes Success objective Success subjectivetration fugation time of application (patients)

Fox 33% 0.9% NaCl 500 g 10 min NR 2-hourly 30 (15) RB 41% 51%; 100%Noble 50% 0.9% NaCl NR NR 48–72 h Replacement 32 (16) IPC 36% 63%Ogawa 20% 0.9% NaCl 1,500 rpm 5 min NR ¥10 28 (14) Fl: 61%; RB: 40% 30%Poon 50–100% 0.5% chlor- 4,000 rpm 10 min 2 h ¥8 11 (9) Fl: 55%; RB: 45% 55%

amphenicol (2,200 g)Rocha 33% 0.9% NaCl 500 g 10 min NR Hourly 4 (2) 100% 100%Takamura 20% 0.9% NaCl 3,000 rpm 10 min NR ¥4–8 NR (26) “Improved” 77%Tananuvat 20% 0.9% NaCl 4,200 rpm 15 min NR ¥6 12 (12) Fl: 39%; RB: 33%; IPC 44% 36% (NS)Tsubota 20% NaCl 1,500 rpm 5 min NR ¥6–10 24 (12) Fl: 55%; RB: 68% 34%

1.3.3Other Indications

Other indications which reportedly have beentreated with autologous serum include recur-rent erosion syndrome, superior limbic kerato-conjunctivitis and as adjunctive therapy in sur-gical ocular surface reconstruction (Tables 1.4,1.5).

1.3.3.1Adjunctive Use in Ocular Surface Reconstruction

Absolute aqueous deficiency prevents a suc-cessful surgical ocular surface reconstruction.Tsubota used 20% autologous serum as adjunc-tive treatment in a prospective cohort study on14 eyes of 11 patients, in which due to Stevens-Johnson syndrome or ocular cicatricial pem-phigoid, the Schirmer test result with nasalstimulation was 0 mm. Surface reconstructionincluded a limbal stem cell graft, amnioticmembrane and/or penetrating keratoplasty.Within the short follow-up of 20 weeks, a stablecorneal epithelium was observed in 12 of the 14eyes [44] and these findings were confirmed byLagnado et al. [23]. Poon used 50% of serum intwo eyes undergoing keratoplasty for PEDs.Again a stable epithelium resulted. However,epitheliopathy recurred when the serum treat-ment was discontinued in these patients. In an-other study, Tsubota also observed that in fourchildren (mean age 9 years) with severe OSDand absolute dry eye due to Stevens-Johnsonsyndrome surface reconstruction failed despitethe use of autologous serum eyedrops [47].

1.3.3.2Recurrent Erosion Syndrome

Insufficient adhesion of the basal epithelial lay-ers to the underlying basement membrane isobserved following trauma or in corneal base-ment membrane dystrophy. This can lead torecurrent erosion syndrome (RES), which in-cludes repeated episodes of irritation, pain,epiphora and conjunctival hyperaemia. DelCastillo treated 11 patients with unilateral post-

traumatic RES with a mean of 2.2 recurrences/month in a prospective cohort study with unspecified, unpreserved lubricants and 20%serum eyedrops TDS for 3 months in a taper-ed fashion. During a mean follow-up of9.4±3.7 months the recurrence rate decreased to0.028/month. No information is given whetherprevious treatment modalities were suspendedfor the time of the serum application. Unfortu-nately the authors also do not state the durationof the history of RES, which may have beenrather short. Given the self-healing nature of thepost-traumatic variant of the condition, thesedata have to be taken with caution [4].

1.3.3.3Superior Limbic Keratoconjunctivitis

This is a rare, chronic, inflammatory diseasethought to result from a localised reduction ofgoblet cells and tear film deficiency at the 12:00limbus with subsequently reduced wettabilityand positive rose bengal staining of the cornealand conjunctival epithelium. Goto used 20%serum eyedrops as additional therapy 10 timesdaily for bilateral superior limbic keratocon-junctivitis (SLK) in a prospective cohort studyon 22 eyes. Within 4 weeks symptoms were im-proved in 9 of 11 and epitheliopathy in all pa-tients.Also, tear break-up time increased signif-icantly and conjunctival squamous metaplasiawas reduced. When the serum application wasdiscontinued, discomfort recurred [16].


∑ Autologous serum can be used in otherocular surface disorders such as superiorlimbic keratoconjunctivitis or if ocularsurface disease – e.g. due to dry eye or PED– becomes so severe that a surgical inter-vention is required

∑ However, the surface disease is likely torecur if the serum treatment is stopped


14C

ha

pter 1A

utologous Serum Eyedrops for O

cular Surface Disorders

Table 1.5. Clinical studies using serum eyedrops to treat other indications. Success is defined either as number/percentage of eyes/patients with improved/reducedmean baseline of objective [fluorescein (Fl) or rose bengal (RB) positive epitheliopathy or impression cytology (IPC)] or subjective (symptoms) score (NR, notreported; RoR, rate of recurrence)

Author Indication Concen- Diluent Centri- Duration Clotting Frequency Eyes Success Success tration fugation time of application (patients) objective subjective

Del Castillo Superior limbic 20% 0.9% NaCl 1,500 rpm 5 min NR ¥3 11 (11) RoR: 99% NRkeratoconjunctivitis

Goto Recurrent corneal 20% 0.9% NaCl 1,500 rpm 5 min NR ¥10 22 (11) Fl: 88%; 21%;erosion syndrome BR: 91%; 82%

IPC 100%

Table 1.4. Clinical studies using serum eyedrops for ocular surface reconstruction. Success was defined as number of all eyes with a stable postoperative ocularsurface score (h, hours; NR, not reported; rpm, revelations per minute)

Author Concen- Diluent Centri- Duration Clotting Frequency Eyes Success objectivetration fugation time of application (patients)

Lagnado 20% 0.9% 4,500 rpm 15 min 10–12 h 1–2 hourly 14 (14) 100%

Poon 50–100% 0.5% chloramphenicol 4,000 rpm (2,200 g) 10 min 2 h ¥8 2 (2) 100%

Tsubota 20% 0.9% NaCl 1,500 rpm 5 min NR 1/4 hourly 14 (11) 86%

1.4Complications

Potential unwanted effects of the use of autolo-gous serum as eyedrops include worsening ofthe initial problem, infection, immunologicalreactions and contact lens deposits.Most authorsmention no complications at all, but in fivepatients with dry eyes, discomfort or epithelio-pathy increased and eyelid eczema was reportedin two cases [34, 40].

1.4.1.1Risk of Infection

An infection in the context of topical serum ap-plication can occur either locally or systemical-ly. The risk for systemic transmission of an in-fectious disease arises only if serum of a donoraffected by a systemic infection, such as HIV,hepatitis or syphilis, is applied to a person oth-er than the donor. This can occur during theproduction or application of serum eyedrops,and a systemic as well as a topical route of entryof the infectious agent is possible, since trans-mission of HIV by a single serum droplet intoone eye has been reported at least in one case[6].Although the risk is small, the potentially fa-tal consequences make the tight quality controlas described under Sect. 1.2.3 mandatory.

Although it is known that serum has antimi-crobial properties, this has not been quantifiedfor diluted, cryopreserved serum so far. Sauer etal. observed that contamination of dropper bot-tles with Staphylococcus epidermidis occurredin 3 out of 40 bottles on the 7th day of use ofundiluted serum stored refrigerated and ap-plied from week dosage containers by trainedpersonnel in a hospital setting [35]. Lagnado etal., however, observed contamination of 18% ofcontainers at the end of the first day of applica-tion, and 43% of inpatients treated with 20%serum diluted with sterile saline (0.9%) had atone stage received serum from a day usagedropper bottle that at the end of the day wasfound to be contaminated. Again most of thesecontaminations were Staphylococcus epider-midis, but one case of S. aureus was also report-ed [23]. Since the indication for the use of serum

was either an epithelial defect or ocular surfacereconstruction surgery,all patients had receivedtopical antibiotics at the same time. This is like-ly to have reduced the rate of contamination ofthe serum containers and prevent cases ofmicrobial keratitis.

The incidence of dropper bottle contamina-tion and the risk of microbial keratitis is likelyto increase when the drops are used in a domes-tic setting by the patients themselves and if notopical antibiotic is applied, e.g. in patients withother indications than epithelial defects. Even if0.5% chloramphenicol was added as a preserva-tive to the dropper bottles, microbial keratitisevolved in 3 out of 13 eyes with PEDs treatedwith 50% or 100% serum [33]. Since laboratoryevidence suggests that dilution with an antibiot-ic may reduce the epitheliotrophic capacity ofserum eyedrops and since ocular surface dis-ease often requires long-term treatment, hospi-talisation of patients for this purpose is not asuitable option and storage of the serum prod-uct in day dosage vials seems preferable.

1.4.1.2Immunological Complications

In the first report by Fox in 1984 the authorsmentioned [9] that some users of serum – notFox himself – had encountered scleral vasculitisand melting in patients with rheumatoid arthri-tis, although the indication for the use of serumin these patients remains unknown. McDonnellreported one case of an immune complex depo-sition after hourly application of 100% serum.Poon observed the onset of one peripheralcorneal infiltrate and ulceration within 24 hafter initiation of serum drops [28, 33] andhypothesised that this could have been a resultof circulating antibodies which must also bepresent in serum eyedrops and could have re-acted with corneal antigens with a subsequentinflammatory response.

1.4.1.3Contact Lens Contamination

We have used serum eyedrops in combinationwith a soft, class IV hydrogel contact lens con-taining 45% ocufilcon D and 55% water (Bio-

1.4 Complications 15

medics 55) in six eyes (six patients) with persist-ent or recurrent epithelial defects.A contact lenswas applied if an epithelial defect recurred orprogressed while the eye was being treated withautologous serum. In five eyes the contact lenswas applied immediately after transplantationof an amniotic membrane. Three lenses in twoeyes showed substantial numbers of large de-posits on the anterior surface of the lens afterapplication of serum drops 8 times a day for18–78 days (Fig. 1.4). In addition all eyes re-ceived unpreserved ofloxacin (3 mg/ml) 4 timesdaily topically. All defects healed without signsof microbial keratitis or conjunctival hyper-aemia. However, since protein deposition mayinduce ocular surface inflammation, it is recom-mended to use silicone hydrogel lenses, if acombination with serum drops is necessary, be-cause this type of lens is less prone to accumu-late protein on the surface [21, 42, 54].


∑ Serum eyedrops are generally well toleratedwith little or no side effects

∑ However, all possible measures should betaken to exclude transmission of systemicdisease during the production or applica-tion of serum eyedrops

∑ Frequent monitoring of patients is recom-mended since in rare cases serious inflam-matory complications can result from theuse of autologous serum drops

∑ Serum eyedrops can be combined with other forms of wound healing supportingtherapy, without any serious adverse effectbeing reported so far

∑ If a contact lens is applied, a material with little tendency to accumulate surfacedeposits is recommended

1.5Alternative Blood Products for the Treatment of Ocular Surface Disease

Serum has to be prepared from an autologousblood donation for each patient individually.This is time and labour intensive and in manycountries no standard operating protocol hasyet been evaluated or approved by the licensingauthority. Other blood derived extracts, such asalbumin, plasma or platelet concentrates, arereadily available from pharmaceutical compa-nies or blood banks. They are quality controlledand might therefore be considered for treat-ment of ocular surface disorders.

1.5.1Umbilical Chord Serum

Recently umbilical chord serum has been pre-pared like autologous serum (5 min centrifuga-tion at 1,500 rpm), diluted to a 20% concentra-tion in 0.9% saline and used as an alternativetreatment for promoting corneal epithelialwound healing. In a prospective randomisedcontrolled clinical trial on 60 eyes, this led to ahigher rate of healing of persistent epithelial de-fects than autologous serum. However, the de-fects treated with autologous serum healedfaster than those treated with umbilical cordserum. This product is not autologous but allo-geneic and hence not only immunological prob-lems but also a higher risk of infection for therecipient may be expected. In many countries itis not supplied by a centralised blood serviceand thus it is more difficult to obtain. All con-


Fig. 1.4. Hydrogel bandage contact lens (45% ocufil-con D, 55% water) with contaminations in the righteye of an 88-year-old female patient after 78 days oftreatment with 20% serum 8 times daily. The patientpreviously had developed a recurrent corneal epithe-lial defect due to neurotrophic keratopathy secondaryto herpes zoster keratitis and severe aqueous tear de-ficiency due to secondary Sjögren’s syndrome whichpersisted, perforated and recurred despite repeatedmultilayer amniotic membrane grafts

cerns regarding quality control and microbio-logical testing are applicable [50].

1.5.2Albumin

One of the most abundant proteins in tears isalbumin is one of the most abundant (Table 1.1).It often acts as a carrier for other factors, such ashormones, including steroids, although nophysiological role for its presence in tears hasbeen described to date. Albumin is also used inmedicine for treating severe albumin deficien-cy, e.g. due to protein loss in extensive skinburns or liver dysfunction. Tsubota found in an animal experiment that the topical applica-tion of albumin reduces enzymes involved inapoptosis and improves epithelial cell viabilityand re-epithelialisation. When a 5% solution of recombinant albumin was applied 6 timesdaily, mean fluorescein and rose bengal, as wellas break-up time and symptom scores, im-proved significantly from baseline, but no com-ments are made as to which solvent was usedand whether any non-responders were foundwith this treatment. Although no complicationswere observed – as with all blood derived prod-ucts – a minute risk of transmission of viral orprion mediated disease cannot be ruled out ifthe albumin preparation was derived fromblood [37].

1.5.3Plasma and Platelets

Plasma is the cell free supernatant after cen-trifugation of full blood mixed with an antico-agulant. Plasma therefore contains only smallamounts of the growth factors present inplatelets since these are – due to the anticoagu-lant – not activated by blood clotting.

Platelets are a major source of growth factorsin serum. They can be obtained by means ofapheresis and stimulated with thrombin to re-lease their content. Following centrifugation theclear supernatant, which is free of any cellularremnants, can be resuspended with a buffer to aconcentration of choice and stored frozen for

further use. We have termed this product“platelet releasate” (PR). This blood productcontains large amounts of EGF and othergrowth supporting mediators, but little of theextracorpuscular factors such as fibronectin orvitamins.

Both products have so far only been tested invitro, but from these experiments it is clear thatplasma is not suitable to support proliferationand migration of epithelial cells. Althoughplatelet releasate also does not support migra-tion, it has a substantially stronger proliferativeeffect than serum and thus may be promisingfor the treatment of ocular surface disorders.However, this still has to be evaluated in a clini-cal trial [18, 19].


∑ Alternative blood products are routinelyproduced and quality controlled by bloodbanks. They are currently under investiga-tion

∑ Umbilical chord serum was found to bemore effective than autologous serum toheal epithelial defects, but it is more diffi-cult to obtain and supply is limited. Sinceevidence about this allogeneic treatmentmodality is even more scarce, it should cur-rently not be used to substitute autologousserum, unless the latter has failed to estab-lish a stable ocular surface

∑ Albumin may be one of the components ofserum that support epithelial wound heal-ing and can be used as a single compoundproduct without the need for an autologousblood donation. It has been reported to im-prove findings but not symptoms of ocularsurface disease in severe dry eyes

∑ Platelet releasate, but not plasma, may besuitable as an additional treatment modali-ty for ocular surface disease, but no clinicaldata are available yet

∑ Similar if not enforced legal guidelines andquality controls need to be applied to thesealternative blood products, since the risk of transmission of infection is due to theirallogeneic nature being theoreticallyincreased

1.5 Alternative Blood Products for the Treatment of Ocular Surface Disease 17

References

1. Alvarado Valero MC, Martinez Toldos JJ, BorrasBlasco J, Alminana Alminana A, Perez Ramos JM(2004) [Treatment of persistent epithelial defectsusing autologous serum application.]. Arch SocEsp Oftalmol 79(11):537–542

2. Chiou AG, Florakis GJ, Kazim M (1998) Manage-ment of conjunctival cicatrizing diseases and se-vere ocular surface dysfunction. Surv Ophthal-mol 43(1):19–46

3. Ciba-Geigy (1978) Wissenschaftliche TabellenGeigy. Ciba-Geigy, Basel

4. del Castillo JM, de la Casa JM, Sardina RC, Fer-nandez RM, Feijoo JG, Gomez AC et al. (2002)Treatment of recurrent corneal erosions usingautologous serum. Cornea 21(8):781–783

5. Dugrillon A, Lauber S, Nguyen XD, Klueter H(2002) Platelets applied to wounds and in tissueregeneration: induction of proliferation apopto-sis by platelet membranes. In: Mempel W,Schramm W (eds) Infusion therapy and transfu-sion medicine. Kluwer Academic/Plenum, Ams-terdam, pp 70–71

6. Eberle J, Habermann J, Gurtler LG (2000) HIV-1infection transmitted by serum droplets into theeye: a case report. AIDS 14(2):206–207

7. Ebner S, You L, Voelcker HE, Kruse FE (2001) Ef-fect of autologous serum on the healing of non-infectious corneal ulcers and expression ofgrowth factor receptors in the cornea. Ophthal-mologe 98:S27

8. Ferreira de Souza R, Kruse FE, Seitz B (2001) [Au-tologous serum for otherwise therapy resistantcorneal epithelial defects – prospective report onthe first 70 eyes]. Klin Monatsbl Augenheilkd218(11):720–726

9. Fox RI, Chan R, Michelson JB, Belmont JB,Michelson PE (1984) Beneficial effect of artificialtears made with autologous serum in patientswith keratoconjunctivitis sicca. Arthritis Rheum27(4):459–461

10. Fullard RJ, Tucker DL (1991) Changes in humantear protein levels with progressively increasingstimulus. Invest Ophthalmol Vis Sci 32:2290–2301

11. Garcia Jimenez V, Veiga Villaverde B, BaamondeArbaiza B, Cahue Carpintero I, Celemin VinuelaM, Simo Martinez R (2003) [The elaboration, useand evaluation of eye-drops with autologousserum in corneal lesions]. Farm Hosp 27(1):21–25

12. Geerling G, Honnicke K, Schroder C, Framme C,Sieg P,Lauer I et al. (1999) Quality of salivary tearsfollowing autologous submandibular gland trans-plantation for severe dry eye. Graefes Arch ClinExp Ophthalmol 237(7):546–553

13. Geerling G, Daniels JT, Dart JK, Cree IA, Khaw PT(2001) Toxicity of natural tear substitutes in afully defined culture model of human cornealepithelial cells. Invest Ophthalmol Vis Sci 42(5):948–956

14. Geerling G, Hartwig D (2002) [Autologous serumeye-drops for ocular surface disorders. A litera-ture review and recommendations for their appli-cation]. Ophthalmologe 99(12):949–959

15. Gordon JF et al. (1995) Topical fibronectin oph-thalmic solution in the treatment of persistentdefects of the corneal epithelium. Am J Ophthal-mol 119:281–287

16. Goto E, Shimmura S, Shimazaki J, Tsubota K(2001) Treatment of superior limbic keratocon-junctivitis by application of autologous serum.Cornea 20(8):807–810

17. Harloff S, Hartwig D, Dibbelt L, Wedel T, GeerlingG (2002) Korneale epitheliotrophe Kapazität vonSerum immunsuppressierter verglichen mit demgesunden Spender. Ophthalmologe 100:S152

18. Hartwig D, Harloff S, Liu L, Schlenke P, Wedel T,Geerling G (2004) Epitheliotrophic capacity of agrowth factor preparation produced from plateletconcentrates on corneal epithelial cells: a poten-tial agent for the treatment of ocular surface de-fects? Transfusion 44(12):1724–1731

19. Herminghaus P, Geerling G, Hartwig D, Wedel T,Dibbelt L (2004) [Epitheliotrophic capacity ofserum and plasma eyedrops: Influence of cen-trifugation]. Ophthalmologe 101(10):998–1005

20. Joh T, Itoh M, Katsumi K,Yokoyama Y, Takeuchi T,Kato T et al. (1986) Physiological concentrationsof human epidermal growth factor in biologicalfluids: use of a sensitive enzyme immunoassay.Clin Chim Acta 158(1):81–90

21. Jones L, Senchyna M, Glasier MA, Schickler J,Forbes I, Louie D et al. (2003) Lysozyme and lipiddeposition on silicone hydrogel contact lens ma-terials. Eye Contact Lens 29(1 Suppl):S75–79; dis-cussion S83–84, S192–194

22. Kao WWY, Kao CWC, Kaufman AH, KombrinckKW, Converse RL, Good WV et al. (1998) Healingof corneal epithelial defects in plasminogen- andfibrinogen-deficient mice. Invest Ophthalmol VisSci 39:502–508

23. Lagnado R, King AJ, Donald F, Dua HS (2004) Aprotocol for low contamination risk of autolo-gous serum drops in the management of ocularsurface disorders. Br J Ophthalmol 88(4):464–465

24. Liu L,Hartwig D,Harloff S,Herminghaus P,WedelT, Geerling G (2005) An optimized protocol forthe production of autologous serum eyedrops.Graefes Arch Clin Exp Ophthalmol 243:706–714


25. Lopez Bernal D, Ubels JL (1993) Artificial tearcomposition and promotion of recovery of thedamaged corneal epithelium. Cornea 12(2):115–120

26. Macaluso D, Feldman S (1997) Pathogenesis ofsterile corneal erosions and ulcerations. In:Krachmer J, Mannis M, Holland E (eds) Cornea.Mosby, St Louis, pp 204–209

27. Matsumoto Y, Dogru M, Goto E, Ohashi Y, KojimaT, Ishida R et al. (2004) Autologous serum appli-cation in the treatment of neurotrophic keratopa-thy. Ophthalmology 111(6):1115–1120

28. McDonnell PJ, Schanzlin DJ, Rao NA (1988) Im-munoglobulin deposition in the cornea after ap-plication of autologous serum. Arch Ophthalmol106:1423–1425

29. Nelson JD, Gordon JF (1992) Topical fibronectinin the treatment of keratoconjunctivitis sicca.Chiron Keratoconjunctivitis Sicca Study Group.Am J Ophthalmol 114(4):441–447

30. Ng V, Cho P, Mak S, Lee A (2000) Variability oftear protein levels in normal young adults: be-tween-day variation. Graefes Arch Clin ExpOphthalmol 238:892–899

31. Noble BA, Loh RS, MacLennan S, Pesudovs K,Reynolds A, Bridges LR et al. (2004) Comparisonof autologous serum eye drops with conventionaltherapy in a randomised controlled crossover tri-al for ocular surface disease. Br J Ophthalmol88(5):647–652

32. Ogawa Y,Okamoto S,Mori T,Yamada M,MashimaY, Watanabe R et al. (2003) Autologous serum eyedrops for the treatment of severe dry eye in pa-tients with chronic graft-versus-host disease.Bone Marrow Transplant 31(7):579–583

33. Poon AC, Geerling G, Dart JK, Fraenkel GE,Daniels JT (2001) Autologous serum eyedrops fordry eyes and epithelial defects: clinical and invitro toxicity studies. Br J Ophthalmol 85(10):1188–1197

34. Rocha EM, Pelegrino FS, de Paiva CS,Vigorito AC,de Souza CA (2000) GVHD dry eyes treated withautologous serum tears. Bone Marrow Transplant25(10):1101–1103

35. Sauer R, Bluthner K, Seitz B (2004) [Sterility ofnon-preserved autologous serum drops for treat-ment of persistent corneal epithelial defects].Ophthalmologe 101(7):705–709

36. Scardovi C, De Felice GP, Gazzaniga A (1993) Epi-dermal growth factor in the topical treatment oftraumatic corneal ulcers. Ophthalmologica 206:119–124

37. Shimmura S, Ueno R, Matsumoto Y, Goto E,Higuchi A, Shimazaki J et al. (2003) Albumin as atear supplement in the treatment of severe dryeye. Br J Ophthalmol 87(10):1279–1283

38. Sitaramamma T, Shivaji S, Rao GN (1998) Effect ofstorage on protein concentration of tear samples.Curr Eye Res 17(10):1027–1035

39. Takamura E, Shiozaki K, Hata H, Yukari J, Hori S(2002) Efficacy of autologous serum treatment inpatients with severe dry eye. In: Sullivan DA (ed)Lacrimal gland, tear film, and dry eye syndromes3. Kluwer Academic/Plenum, New York, pp 1247–1250

40. Tananuvat N, Daniell M, Sullivan LJ, Yi Q, McK-elvie P, McCarty DJ et al. (2001) Controlled studyof the use of autologous serum in dry eye pa-tients. Cornea 20(8):802–806

41. Thomas L (1998) Labor und Diagnose, 5th edn.TH-Books, Frankfurt/Main, p 1494

42. Tighe BJ, Jones L, Evans K, Franklin V (1998) Pa-tient-dependent and material-dependent factorsin contact lens deposition processes. Adv ExpMed Biol 438:745–751

43. Torsteinsdóttir L, Hakansson L, Hällgren R, Gud-björnsson B, Arvidson NG,Venge P (1999) Serumlysozyme: a potential marker of monocyte/macrophage activity in rheumatoid arthritis.Rheumatology 38:1249–1254

44. Tsubota K, Satake Y, Ohyama M, Toda I, Takano Y,Ono M et al. (1996) Surgical reconstruction of theocular surface in advanced ocular cicatricialpemphigoid and Stevens-Johnson syndrome [seecomments]. Am J Ophthalmol 122(1):38–52

45. Tsubota K,Goto E,Fujita H,Ono M,Inouc H,SaitoI et al. (1999) Treatment of dry eye by autologousserum application in Sjogren’s syndrome. Br JOphthalmol 83:390–395

46. Tsubota K, Goto E, Shimmura S et al. (1999) Treat-ment of persistent corneal epithelial defect byautologous serum application. Ophthalmology106(10):1984–1989

47. Tsubota K, Shimazaki J (1999) Surgical treatmentof children blinded by Stevens-Johnson Syn-drome. Am J Ophthalmol 128:573–581

48. Tsubota K, Goto E, Shimmura S, Shimazaki J(1999) Treatment of persistent corneal epithelialdefect by autologous serum application. Ophthal-mology 106(10):1984–1989

49. Ubels J,Williams K, Lopez Bernal D,Edelhauser H(1994) Evaluation of effects of a physiologic arti-ficial tear on the corneal epithelial barrier: elec-trical resistance and carboxyfluorescein perme-ability. In: Sullivan DA (ed) Lacrimal gland, tearfilm and dry eye syndromes. Plenum Press, NewYork, pp 441–452

50. Vajpayee RB, Mukerji N, Tandon R, Sharma N,Pandey RM, Biswas NR et al. (2003) Evaluation ofumbilical cord serum therapy for persistentcorneal epithelial defects. Br J Ophthalmol 87(11):1312–1316

References 19

51. Yokoi N, Komuro A, Nishida K, Kinoshita S (1997)Effectiveness of hyaluronan on corneal epithelialbarrier function in dry eye [see comments]. Br JOphthalmol 81(7):533–536

52. Young AL, Cheng AC, Ng HK, Cheng LL, LeungGY, Lam DS (2004) The use of autologous serumtears in persistent corneal epithelial defects. Eye18(6):609–614

53. Geerling G, Mac Lennan S, Hartwig D (2004) An-tologous serum eyedrops for ocular surface dis-orders. Br J Ophthalmol 88:1467–1474

54. Schrader S, Wedel T, Moll R, Geerling G (2005)Use of a combination of serum eyedrops and hy-drogel bandage contact lens for persistent epithe-lial defects. DOG-abstract, 2609-14.01


2.1Introduction

The first documented ophthalmologic applica-tion of the amniotic membrane was in the 1940’swhen it was used in the treatment of ocularburns [9, 54, 55]. Following initial reports, its usein ocular surgery, as indicated by reports in thescientific literature, abated until recently. It ap-

pears that during these ‘silent years’ it was beingused extensively in the Soviet Union and latter-ly in South America [13]. Its introduction toNorth America in the early 1990s heralded amassive surge in the ophthalmic applications ofthis membrane. The amniotic membrane is nowincreasingly being used in ocular surface sur-gery for a wide range of indications. There areover 500 publications in the scientific literaturewith most of them reporting success. However,

Controversies and Limitations of Amniotic Membrane in Ophthalmic Surgery

Harminder S. Dua, V. Senthil Maharajan, Andy Hopkinson

2

∑ The amniotic membrane is a useful adjunctin the management of many ocular condi-tions

∑ Several mechanisms of action have beenattributed to the membrane based on itsstructure and biochemical composition.Not all mechanisms are scientificallysubstantiated

∑ Many inter and intra donor variations in thestructure and function of the membranehave been demonstrated. Location in rela-tion to the placenta, duration of pregnancy,parity, gravidity, onset of labour and evenage and race of the donor are all variables

∑ Processing and preservation of the mem-brane can be accomplished by differentmethods. Different methods affect themembrane differently and can substantiallyalter the membrane. Mandatory quarantineof the membrane to rule out HIV contami-nation does not allow for use of fresh mem-branes.The potential risk of transmission of serious infections from one donor to anumber of recipients remains a concern

∑ Despite the vast literature on the use of themembrane, randomised controlled studiesare virtually none

∑ Lack of defined criteria of success makesevaluation of outcomes difficult.This is rendered more difficult due to impropercharacterisation of disease severity makingcomparisons between studies next toimpossible

∑ Although the beneficial effects of the mem-brane are emphasised in several consecu-tive case series, these studies often lackproper and adequate controls and it isimportant to bear in mind that existing, attimes simpler options do exist with equiva-lent or better efficacy

∑ Standardisation of the membrane suppliedfor widespread clinical use is an importantchallenge that lies ahead. Perhaps thegeneration of a ’synthetic membrane’ withknown quantities of desired ingredients,tailored to the intended clinical use, will bepossible in the future

Core Messages

the enthusiasm to extend its clinical applica-tions and indications is not matched by the sci-entific rigor that should be applied to any newproduct or technique that is being used so ex-tensively. There are thus several limitations ofthe amniotic membrane as applied to its oph-thalmic usage, which are not widely known. It istherefore important that these limitations arecarefully considered and all applications of themembrane be interpreted in the context of theselimitations. These limitations apply to the fol-lowing areas:1. Proposed mechanisms of action of the mem-

brane2. Intra and inter donor variations of the mem-

brane3. Processing and preservation of the mem-

brane4. Clinical studies and outcomes [definitions of

success and grading of disease severity]5. Efficacy of membrane in relation to other

established techniques and options


∑ The amniotic membrane has been used in ophthalmic surgery since the mid 1940s

∑ Many ophthalmic applications are pro-posed but not are supported by scientificevidence

∑ The major controversies and limitations of the membrane are in relation to its pro-posed mechanisms of action, inter and intradonor variations, its methods of processingand preservations, the evaluation of its outcomes against defined criteria and incomparison to existing techniques

2.2Proposed Mechanisms of Action of the Amniotic Membrane

Several mechanisms of action are attributed tothe membrane. These include: (a) promotes ep-ithelialisation, (b) inhibits scarring, (c) inhibitsvascularisation, (d) reduces inflammation, (e)provides a substrate for cell growth, (f) antimi-crobial effects and (g) as a biological bandage[3, 13]. Most of these are inferred from the struc-

tural and biochemical composition of the mem-brane often without any direct evidence.

To understand the basis of some of the pro-posed mechanisms of action of the AM it is use-ful to understand the structure and composi-tion of the membrane.

2.2.1Amnion Structure

The AM consists of five layers from within out-ward: (a) a single layer of highly metabolicallyactive, columnar to cuboidal epithelium; (b) athin basement membrane; (c) a compact layermade of reticular fibres virtually devoid of cells;(d) a loose network of reticulum containingfibroblasts, called the fibroblast layer; and (e) aspongy layer of wavy bundles of reticulumbathed in mucin, which forms the interface withthe chorion [4].

Matrix. Amniotic basal lamina contains largequantities of proteoglycans rich in heparin sul-phate.Amnion contains a large amount of colla-gen, hyaluronan and predominantly smallerproteoglycans such as biglycan and decorin,with decorin being more prominent of the two,and is located in close connection with the col-lagen fibrils [36]. Collagen types I, III, IV, V andVII [2, 25, 39, 62], laminin [2] and fibronectin[33] have been identified in amniotic basementmembrane and stromal amnion. Similaritiesbetween the lamini-1, laminin-5, fibronectinand type VII collagen components of the base-ment membranes of conjunctiva, cornea andamniotic membrane have been demonstrated[20]. The a-subchain components of collagenIV have been shown to be similar between am-niotic membrane and conjunctiva but differentbetween amniotic membrane and cornea [20].

2.2.2Amnion Composition

Some components of the membrane that arerelevant in the context of its mechanism ofaction, and that help understand its limitations,are mentioned below:

22 Chapter 2 Controversies and Limitations of Amniotic Membrane in Ophthalmic Surgery

Enzymes. Important amongst these are en-zymes involved in prostaglandin synthesis suchas phospholipases, prostaglandin synthase andcyclo-oxygenase [53, 57]. Prostaglandin dehy-drogenase, a prostaglandin-inactivating enzyme,has also been demonstrated [7]. Secretoryleukocyte protease inhibitor, a potent inhibitorof human leukocyte elastase, has been demon-strated in human amniotic fluid and in the am-niotic membrane. Its concentrations can be up-regulated by exposing amniotic cells to IL-1a,IL-1b, and TNFa [63].

Cytokines. Interleukins-6 and -8 are the pre-dominant cytokines associated with amnioncells [24, 47]. Expression of these cytokines wasincreased in the presence of IL-1b, TNFa andbacterial lipopolysaccharide. IL-10 and IL-1RA(receptor antagonist), both anti-inflammatorycytokines, have been shown in amnion epithe-lial and mesenchymal cells [22].

Growth Factors. Studies on human amnioticmembrane have revealed the presence of EGF,TGFa, KGF, HGF, bFGF, TGF-b1, and -b2 by RT-PCR for the mRNA and by ELISA for theprotein products [29]. TGF-b3 and growth fac-tor receptors KGFR and HGFR were also detect-ed by RT-PCR. A higher level of various growthfactors was found in amniotic membrane withepithelium than without epithelium, indicatingan epithelial origin for these growth factors [29, 45]. Neurotrophic factors like NGF (nervegrowth factor) have also been demonstrated inthe amniotic membrane and amniotic fluid [51,60].

Metalloproteases and Inhibitors of Metallopro-teases. Tissue inhibitors of metalloproteases(TIMPS) have been shown to be produced byboth amniotic epithelial cells and mesenchymalcells [22, 50]. The presence of tissue metallopro-teases (TMPS) has also been demonstrated inamniotic fluid and amniotic membrane, wherethey may play a role in the mechanisms of hu-man parturition and in the regulation of hostresponse to intrauterine infection [17, 41].

Controversies and Limitations. From the aboveit is obvious that there are several contradictory

components in the membrane. This is not sur-prising as in any biologically active tissue suchas the AM, balances and counterbalances for ac-tion of various molecules would be expected.However, when applied surgically to the ocularsurface or elsewhere only the presence of thedesirable ones for a particular set of action(s) isquoted with no regard to the opposing mole-cules. For example, the presence of prosta-glandins in the membrane would promote in-flammation but the presence of prostaglandininactivating enzyme and of secretory leukocyteprotease inhibitor would suppress inflamma-tion. The presence of anti-inflammatory cyto-kines such IL-1Ra and IL-10 would suppress in-flammation but the presence of IL-6 and IL-8would promote inflammation. In eyes that areinflamed due to injury, other proinflammatorycytokines such as IL-1a, IL-1b and TNFa couldalso promote both pro- and anti-inflammatorycytokines and enzymes.

Similarly, the presence of various growth fac-tors like EGF would support epithelial growthand TGF would support wound healing. Howev-er, TGF would itself promote scar tissue forma-tion and be contradictory to the ‘anti-adhesiveor scar suppressing’ action proposed for themembrane in preventing corneal and conjunc-tival cicatrisation. Likewise, the presence ofTIMPS would disfavour vascularisation but thepresence of TMPS would have the opposing ef-fect. AM has been used with success in somecases of ocular surface burns with limbal is-chaemia. One of the main concerns in such pa-tients is the ‘limbal ischaemia’ and proceduressuch as tenoplasty and conjunctival flaps wherepossible have been advocated in an attempt torestore limbal vascularisation. Would not appli-cation of the membrane, with its ‘inhibitor ofvascularisation effect’, have a contradictoryeffect on limbal ischaemia?

The mere presence of one or the other ‘factor’thus cannot be presented as evidence in supportof a particular mode of action for the mem-brane. How these various factors interplay, if atall, in the transplanted membrane to bringabout some of its attributed effects remains tobe elucidated.

Of the proposed mechanisms of action of themembrane the most likely manner in which it

2.2 Proposed Mechanisms of Action of the Amniotic Membrane 23

affects its beneficial effect is perhaps as a sub-strate or basement membrane transplant. Itprovides a favourable substrate by virtue of itsbasement membrane, for new epithelial cells tomigrate on, expand and adhere. Use of themembrane as a bandage to cover inflamed orexposed areas, due to injury or surgery, not onlyfavourably influences the healing process butalso has a dramatic favourable effect on thesymptom of pain and discomfort. It is our clini-cal experience that when denuded areas of theocular surface, particularly the cornea, are cov-ered by amniotic membrane, the levels of painand discomfort experienced by the patient aresignificantly reduced. This too could be as a re-sult of the mechanical or physical presence ofthe membrane. One study has shown that amni-otic fluid application to the corneal surface ofrabbits following excimer laser photokeratecto-my actually enhanced corneal sensitivity andnerve regeneration [32].


∑ The amniotic membrane is composed of asingle layer of epithelial cells, basementmembrane and avascular stroma

∑ Several growth factors, cytokines, proteasesand their inhibitors, antimicrobials, antian-giogenic factors and enzymes have beenidentified in these layers

∑ The mechanisms of action of the membraneare attributed to and inferred from its physical structure and its molecular con-stituents

∑ The fate of these substances after trans-plantation is unknown and the mere pres-ence of a factor or molecule does not implythat it is available in its active form aftersurgery

∑ Often both pro and anti factors are presentand a beneficial effect cannot be ascribed to any one of them without taking intoaccount the effect of the other

∑ The most uncontroversial mechanism of action is by its physical presence as a‘substrate’ transplant

2.3Intra and Inter Donor Variations of the Membrane

The general functional description of the am-nion is that of an epithelial lining which con-tributes to the homeostasis of amniotic fluid. Itis natural therefore that its physiological roleand some corresponding morphology willchange depending on the stage of gestation.This in fact has been amply demonstrated.Many changes in the biochemical compositionof the membrane are known to occur nearerterm and to be induced by labour, for exampleincreased apoptosis of amnion epithelial cellsoccurs just before commencement of labourand IL-6 and IL-8 are found in increased con-centrations in the amniotic fluid towards theend of pregnancy [24].

The amnion varies in histological appear-ance from conception to maturity and severaldifferent patterns are often noted even at term.There is an increase in prostaglandin synthasein the amnion at term and during labour [53].The epithelial morphology can change to that oflarge flat cells and some show distinct intercel-lular channels.Also, the amniotic epithelial cellsare columnar towards the placenta and cuboidalaway from it [43]. The apical surfaces of the am-niotic epithelial cells are covered by microvilli[8, 43, 61], the density of which varies duringpregnancy. An amorphous material of un-known substance is seen on the surface of thesemicrovilli at term [43]. In a recent elaboratestudy, using TGFb as a test molecule to illustrateinter donor variations such differences wereclearly demonstrated (see next section below).In a recent study, Fortunato et al. [16] havedemonstrated a racial disparity in the ability ofthe membrane to respond to infectious stimuli,suggesting that race may yet be another variableto contend with.

With increasing use of the membrane it isour experience that the thickness and trans-parency of the membrane varies at differentsites of the membrane. Generally, the mem-brane closer to the umbilical cord appearsthicker. This also affects the transparency ortranslucency of the membrane. The variation


between donors can be more pronounced. Theamnion can vary in thickness from 0.02 to0.5 mm [4, 8].

Controversies and Limitations. Many differ-ences, some yet unknown, can exist betweenamniotic membranes obtained from differentdonors. Racial variations too may exist. Thereare many other important variables such as theduration of gestation, trial of labour before cae-sarean section and perhaps parity and gravidi-ty of the donor. Increased prostaglandin andproinflammatory cytokines nearer term couldinfluence the effect of the membrane when ap-plied on the ocular surface. Even for the samedonor, it is the general practice to obtain nu-merous pieces of amnion for use in multipleoperations. Clearly therefore some pieces wouldbe from locations closer to the placenta and oth-ers distant to it. The thickness of these locationscan vary as can the morphology of the amniot-ic epithelium. The thickness could affect the in-tegration of the membrane with the ocular sur-face tissues and perhaps influence the ease withwhich the membrane ‘comes off ’ at some pointafter surgery. The transparency would naturallyaffect potential vision when applied on thecorneal surface. Current practice in procure-ment and supply of the membrane does not in-dicate these variables that could be important tothe outcome of its usage. Age, race, parity, gra-vidity, duration of gestation, whether trial oflabour was given or not and location of a partic-ular piece of membrane supplied need to beconsidered in evaluating outcomes and even-tually in bringing some semblance of standard-isation in the practice of amniotic membranetransplantation. This may not be practicallypossible but only by recording these variablescan we begin to understand whether and what effect they may have on the outcome oftransplantation. One point that comes out loudand clear from the above is that membranesused in transplantation are far from standard-ised across donors and even within the samedonor.


∑ The amniotic epithelial morphology variesfrom flat to cuboidal to columnar

∑ The thickness and transparency of themembrane is different at different parts ofthe membrane – thinner and clearer awayfrom the placenta, and in different donors

∑ The membrane undergoes considerablephysiological changes near term andduring labour

∑ Racial variations between donors can exist∑ Duration of pregnancy, trial of labour,

gravidity and parity can all influence thecomposition of the membrane

∑ Different pieces of the membrane from thesame donor can potentially have differenteffects

2.4Processing and Preservation of the Membrane

Many methods of preserving and storing amni-otic membrane for ocular and other uses havebeen described. Some of these are historical andothers popular and currently in vogue. Methodssuch as lyophilisation [6, 56], air drying [35, 46],glutaraldehyde and polytetrafluoroethylenetreatment [40] and irradiation [35, 46, 59] havebeen described but are not among the ones usedcommonly for ophthalmic use. Preservation byfreezing is the commonest mode of preserva-tion of the membrane before use. This involvesuse of two types of solutions, either DMSO inphosphate buffered saline [3, 52] or Eagle’s min-imum essential medium (MEM) and glycerol[26, 27]. Recently a freeze dried preparation ofthe membrane has been commercially intro-duced (Ambiodry) but not much is knownabout its usage. Furthermore different antibiot-ic cocktails, 0.5% silver nitrate [21] and 0.025%sodium hypochlorite solution [48, 49], havebeen used to render the membrane sterile. Inmany parts of the world ‘fresh membrane’(within days or weeks of donation) is still used.However, in most Western countries the use ofone or the other method of preservation ismandatory because of legislation requiring thatthe membrane be adequately screened for HIVcontamination. To this end the donor is tested at the time of delivery and 6 months later (to cover the window period of infection). All

2.4 Processing and Preservation of the Membrane 25

processed membranes are stored in quarantineuntil the second test is performed and reportednegative.

Controversies and Limitations. Several differ-ences can therefore occur between differentmembranes depending on whether they areused fresh or preserved and, in case of the latter,on the mode and duration of preservation. Mostmethods employed in the preservation of themembrane affect it in some manner.

Kruse et al. [30] demonstrated that cryop-reservation significantly impaired the viabilityand proliferative capacity of amniotic mem-brane and its cells. They concluded that amniot-ic membrane grafts function primarily as a ma-trix and not by virtue of transplanted functionalcells. Kubo et al. [31] have shown that after2 months of freezing, at least 50% of amnioticcells are viable and capable of proliferation. Af-ter 18 months of cryopreservation, they were notable to demonstrate a significant amount of cellsurvival. Our own studies show that cell viabilityis minimal, if at all, after 6 months of preserva-tion at –80 °C; however, at this time point themembrane continues to demonstrate manygrowth factors and cytokines (Hopkinson A etal., submitted for publication). Fujisato et al. [18]cross-linked amniotic membrane with chemicalmeans (glutaraldehyde) and with gamma-rayand electron beam. They showed that radiationcross-linked membranes degraded rapidly invitro compared to chemically cross-linkedmembranes. Hao et al. [22], who demonstratedthe presence of mRNA for both antiangiogenicand anti-inflammatory factors in amnioticmembrane, have suggested that amniotic mem-brane should be applied epithelial cell surfacedown in order to deliver a high concentration ofthese factors to the damaged ocular surface.Thiswould be applicable more to fresh rather than topreserved membranes if one were to accept (de-spite lack of any evidence in support) that theamniotic epithelial cells continue to produce thedesired ‘factors’ in biologically active forms aftertransplantation onto the ocular surface.

Hopkinson A et al. (submitted for publica-tion) studied extensively the growth factor TGF-b1 to determine intra- and intermembrane vari-ations. They showed that at both the gene and

protein level TGF-b1 is the highest expressedisoform of TGF-b, and that expression is lowerin AM than in chorion. In addition, they demon-strated considerable variations in TGF-b1 geneexpression between membranes, with expres-sion at different locations within a single mem-brane also appearing to vary. Another impor-tant observation was that maximal presence ofTGF-b1 was in the acellular spongy layer, whichsuggests that the spongy layer most likely acts asa depot for chorion-derived TGF-b1. It is possi-ble that the spongy layer acts as a physical bar-rier, preventing chorionic-TGF-b1 from diffus-ing into the AM during gestation. They alsoshowed that alterations in the method of han-dling the membrane could drastically alter theconcentration of TGF-b1. Any method of pro-cessing and storage of the membrane, therefore,that did not get rid of this layer would yield aproduct that would enhance wound healing andscarring and vice versa. This study has conclu-sively demonstrated that, not only for this onefactor but for other proteins as well, consider-able variations exist between membranes, atdifferent locations within the same membraneand can be profoundly affected by the methodof processing and storage of the membrane(Fig. 2.1). Clinically, such variation betweenmembranes is not considered prior to surgery,and therefore the effect on clinical efficacy isunknown. To determine the clinical significanceof such variables, further studies are required.

Koizumi et al. [28] cultivated rabbit limbaland corneal epithelial cells on denuded humanamniotic tissue. They demonstrated significant-ly improved growth of cells on membrane de-nuded of epithelial cells compared with intactmembrane. To the contrary, most of the mem-brane supplied in the USA is with the amnioiticepithelium in situ. It is claimed that ocular sur-face epithelial cells grow better on this surface.This controversy remains unresolved.Clinically,it is our experience that both membranes, withand without the amniotic epithelium in situ,seem to support growth of ocular surface cells.

The weight of the evidence available sup-ports the notion that viability of the tissue com-ponents of the amniotic membrane is not essen-tial for its biological effectiveness. However, theextent of the effectiveness could vary and result


in inconsistent results. Until the effect of thesedifferent methods of preservation and storageon the membrane has been evaluated and stan-dardised, success or failure of the membraneshould be qualified by the method of preserva-tion employed.

Despite the prolonged quarantine, the poten-tial risk of transmission of serious infectionsfrom one donor to a number of recipientsremains a concern. This concern is heightenedin countries where fresh membrane is usedthough mitigated to some extent by the fact thatonly a limited number of surgeries (recipients)are performed with a given donor.


∑ Many different methods of preservationand storage of the membrane exist

∑ Wet freezing in phosphate buffered salineor in minimum essential medium are cur-rently the two most popular methods

∑ A freeze-dried membrane has recently beenintroduced in the market and initial usageseems to suggest clinical efficacy

∑ The effect of these different methods on thestructural and biochemical composition ofthe membrane is not fully understood. Sucheffects could have a direct consequence onthe proposed mechanisms of action

∑ Studies have shown that preservation andprocessing can have profound effects on themembrane constituents

∑ Fresh, and to some extent even preserved,membranes that have been tested twice carry a potential risk of spread of seriousinfections

2.5Clinical Studies and Outcomes (Definitions of Success and Grading of Disease Severity)

Most published reports on use of the amnioticmembrane in ophthalmology have been consec-utive case series or retrospective studies. Ran-domised controlled studies are practically non-existent bar one or two [5]. This must be

2.5 Clinical Studies and Outcomes (Definitions of Success and Grading of Disease Severity) 27

Fig. 2.1 A, B. Two-dimensional gel electrophoresisof solubilised protein from transplant ready amnioticmembrane. Sixty micrograms of protein from twodifferent donor membranes (A, B) was separated on18-cm pH 3–11 IPG strips (Amersham Biosciences)and then on a 8–19% gradient polyacrylamide gel fol-lowed by silver staining of protein spots. Comparablespots of similar intensity representing similarities

(reference markers) between membranes are indicat-ed (arrows). Variation between membranes, repre-sented by spots detected in some membranes but notin others, is also indicated (A 1–4). Examples shownare of comparable zoomed areas of the whole gel, andrepresentative experiments of a total of 24 performedare shown

A B

considered a serious limitation of the use of theamniotic membrane. Another significant short-coming has been the lack of clear definitions ofsuccess or failure and the evaluation of the out-comes against such definitions. This limitationis to some extent matched by a similar lack ofconsistency in gradation and classification ofseverity of the disease that make up the majorindications for use of the membrane, for exam-ple ocular surface burns. These two factorscombine to make evaluation of the effects of themembrane difficult and in some instances ren-dering it even impossible to compare outcomesbetween different groups [12, 15, 23, 37].

We have used a pre-determined protocol todefine outcomes and propose this as one modelthat may be considered in evaluating the effica-cy of the membrane (Maharajan et al., submit-ted for publication). When the membrane wasused with the intention of it becoming incorpo-rated into the recipients tissue it was termed agraft and when the intention was for it to comeaway or be removed at a certain point followingsurgery, it was termed a patch. In our group ofpatients it was used primarily as a graft or apatch with four objectives: (a) to establish ep-ithelial cover in an area where none existed, (b)to prevent corneal perforation in eyes at riskdue to stromal melting, (c) to limit scarringwhere the clinical likelihood was high or wherescarring (symblepharon/adhesions) previouslyexisted, and (d) to limit inflammation and neo-vascularisation. The outcome was evaluatedagainst both the intended purpose of the mem-brane, patch or graft, and whether the objectivewas achieved or not. Three outcomes were thusdefined, success, partial success and failure.

Criteria for Success or Failure (Maharajan et al.,submitted for publication). The purpose ofAMT was to act as a patch, graft or both. AMTwas carried out to fulfil one or more of the fol-lowing objectives: (a) to establish epithelial cov-er in an area where none existed, (b) to preventcorneal perforation in eyes at risk due to stro-mal melting, (c) to limit scarring where the clin-ical likelihood was high or where scarring (sym-blepharon/adhesions) previously existed, and(d) to limit inflammation and neovascularisa-tion. Three outcome measures were applied: (1)

success: when the membrane served the pur-pose that was intended, i.e., acted as a patch orgraft and the objective was achieved; (2) partialsuccess: (a) when the membrane did not servethe purpose that was intended, i.e. acted as apatch when intended as a graft or vice versa butthe objective was achieved, for example re-ep-ithelialisation of a persistent epithelial defect(PED) occurred, (b) when the membrane (as apatch) did not persist long enough but the ob-jective was nevertheless achieved, for exampleepithelialisation continued till the defect wasclosed, (c) when multiple objectives were setand not all were realised; (3) failure: when theobjective was not achieved even though the pur-pose may have been achieved, for example if the


Fig. 2.2. A Ocular surface reconstruction with al-lolimbal transplantation and use of two membranes.The inner 9-mm disc acts as a graft and the outerlarger membrane as a patch. The outer membraneprevents conjunctival epithelial cells mixing with thetransplanted limbus-derived cells growing on the in-ner membrane (author’s technique: H.S. Dua). B Theouter membrane has cut through sutures and retract-ed, exposing the inner membrane. This was consid-ered as a partial success

A

B

membrane was intended as a patch and acted assuch for the expected duration but the PED didnot heal.

When the above criteria were applied to 74procedures involving use of the amniotic mem-brane, failure of the procedure was observed in44% of patients where the membrane was usedin the presence of stem cell deficiency, in 33% ofprocedures where the membrane was used inthe absence of stem cell deficiency and in 44%of patients where the membrane was used forconjunctival reconstruction. This clearly illus-trates that a significant proportion of failurescan occur and these should be appropriatelyrecognised and documented in order to refineand temper the vastly expanding indications foruse of the membrane (Figs. 2.2, 2.3).

Controversies and Limitations. Lack of ran-domised controlled trials, lack of clearly definedcriteria of success and failure against intendeduse of the membrane and inadequate gradationor classification of disease severity all con-tribute significantly to highlight a serious limi-tation surrounding the clinical use of the mem-brane, ascertaining its efficacy for differentindications and comparing and evaluating out-comes of use.


∑ The membrane can be used to serve eitheras a patch, when the membrane is removedafter some time or is expected to fall off,or as a graft (including carrier of ex vivoexpanded cells) when the membrane isincorporated into the host tissue

∑ Important objectives of use of the mem-brane are to provide epithelial cover, arrestmelting, limit scarring, limit inflammationand neovascularisation

∑ Outcomes should be clearly defined,for example as success, partial success or failure

∑ Proper gradation or classification ofclinical conditions is necessary for properevaluation of the membrane

∑ Randomised controlled studies are requiredto scientifically evaluate the efficacy of themembrane

2.5 Clinical Studies and Outcomes (Definitions of Success and Grading of Disease Severity) 29

Fig. 2.3. A Corneal scarring and vascularisation de-spite allolimbal transplantation and use of amnioticmembrane in a case of stem cell deficiency. Thispatient was considered a failure. B Failed amnioticmembrane transplant in a patient of acute ocular sur-face burn. C Amniotic membrane and corneal stro-mal melting in a patient of bullous keratopathy treat-ed with amniotic membrane transplantation. Thepicture was taken on day 7 after surgery

A

B

C

2.6Efficacy of Membrane in Relation to Other Established Techniques and Options

Another interesting point of note is that manyreports of the success of the use of the mem-brane for one indication or the other do notcompare it with valid controls or even againststandard existing techniques for the same indi-cation. This leads to the erroneous message thatthe membrane should be used for a particularindication without flagging that an existing pro-cedure or alternative to the membrane may beequally effective or in some cases better. Thefollowing illustrations emphasise this point:

When the advocated use of the membrane torepair leaking trabeculectomy blebs was stud-ied against the standard conjunctival advance-ment, the latter was found to be more reliable[5]. In some studies where the membrane hasbeen used to repair failing blebs, antimetabo-lites too have been used as adjuncts. In the ab-sence of any controls it is impossible to assessthe contribution of the membrane to any suc-cess, which may equally have been due to theantimetabolite [13, 19].

Its use in pterygium surgery too is not fullyclarified. It may be a useful alternative but autol-ogous conjunctival grafts seem to have a bettersuccess rate than amniotic membrane grafts.Prabhasawat et al. [44] found that in pterygiumsurgery the recurrence rate was higher withamniotic membrane compared to autologousconjunctival grafts. Ma et al. [34] found no dif-ference between the amniotic membrane, mito-mycin C or autologous conjunctival grafts in themanagement of pterygium, but recommend useof the membrane.

An important case in point is the reportedsuccess of the membrane in treating patientswith partial stem cell deficiency. The membranecan offer help in instances where a fibrovascularpannus has to be excised, but where abnormalconjunctiva derived epithelium has encroachedonto the corneal surface the membrane is oftennot required though reportedly used and advo-cated. Tseng have shown good results with useof the membrane following superficial debride-

ment of abnormal cells from the surface of thecornea in cases of partial stem cell deficiency [1, 38, 58]. In a technique now established assequential sector conjunctival epitheliectomy(SSCE), Dua has shown [10, 11, 14] that the sameeffect can be achieved with simple debridementwithout the use of the membrane. The studiesthat advocated the use of the membrane for thisindication did not include any controls wheresimple debridement was undertaken withoutamniotic membrane transplant. The SSCE tech-nique can be taken as akin to controls for thesestudies and illustrates that proposed indicationsof the membrane are not always substantiatedscientifically.

Similarly, the use of the membrane for the in-dication of painful bullous keratopathy [42]needs to be evaluated against the alternative of anterior stromal puncture and temporaryplacement of a bandage contact lens. The latteris a simple outpatient procedure with compara-tively little expense. In an ongoing stratified (ac-cording to pain score) randomized controlledstudy we have performed the two different pro-cedures in 25 patients and thus far no significantdifference between amniotic membrane trans-plantation and anterior stromal puncture isemerging.

Controversies and Limitations. Many proposedindications of the membrane are not foundedon hard evidence. In some instances use of themembrane may be an option but not necessari-ly a better option. In some cases it has distinctdisadvantages over existing techniques. Longerfollow-up studies are needed to ascertain dura-tion of its efficacy.


∑ Alternative options exist for many ofthe clinical indications for use of the membrane

∑ In most instances the membrane has notbeen compared directly with these optionsto evaluate its superiority or otherwise

∑ Clear examples of the failure of the membrane exist where other options have succeeded


∑ In some cases where the membrane may be just as good as an existing technique,the associated disadvantages of an inter-ventional procedure requiring use of anoperative room should be considered

This chapter has been written to put in perspec-tive the widespread use of the membrane andhighlight areas which need to be addressed byfurther studies and continued critical analysis.That is not to say that the membrane does nothave its uses. It is certainly a useful option inmany clinical situations and in this context isthe subject of another chapter in this text. Stan-dardisation of the transplant ready amnioticmembrane (TRAM) in relation to donor vari-ables and processing and preservation vari-ables; proper categorisation of the extent andseverity of the diseases for which it is used; anddefining criteria of success and failure to evalu-ate outcomes will go a long way in putting the amniotic membrane on a sound scientificfooting. Perhaps the generation of a ‘syntheticmembrane’ with known quantities of desiredingredients, tailored to the intended clinical use,will be possible in the future.

References

1. Anderson DF, Ellies P, Pires RT, Tseng SC (2001)Amniotic membrane transplantation for partiallimbal stem cell deficiency. Br J Ophthalmol85:567–575

2. Aplin JD, Campbell S, Allen TD (1985) The extra-cellular matrix of human amniotic epithelium:ultrastructure,composition and deposition. J CellSci 79:119–136

3. Azuara-Blanco A, Pillai CT, Dua HS (1999) Amni-otic membrane transplantation for ocular surfacereconstruction. Br J Ophthalmol 83:399–402

4. Bourne GL (1960) The microscopic anatomy ofthe human amnion and chorion. Am J Obstet Gy-necol 79:1070–1073

5. Budenz DL, Barton K, Tseng SC (2000) Amnioticmembrane transplantation for repair of leakingglaucoma filtering blebs. Am J Ophthalmol 130:580–588

6. Burgos H, Sergeant RJ (1983) Lyophilized humanamniotic membranes used in reconstruction ofthe ear. J R Soc Med 76:433

7. Cheung PY, Walton JC, Tai HH, Riley SC, ChallisJR (1990) Immunocytochemical distribution andlocalization of 15-hydroxyprostaglandin dehy-drogenase in human fetal membranes, decidua,and placenta. Am J Obstet Gynecol 163:1445–1449

8. Danforth DN, Hull RW (1958) The microscopicanatomy of the fetal membranes with particularreference to the detailed structure of the amnion.Am J Obstet Gynecol 75:536–550

9. de Rotth A (1940) Plastic repair of conjunctivaldefects with fetal membranes. Arch Ophthalmol23:522–525

10. Dua HS (1998) The conjunctiva in corneal epithe-lial wound healing. Br J Ophthalmol 82:1407–1411

11. Dua HS (2001) Sequential sector conjunctival ep-itheliectomy. In: Holland EJ, Mannis MJ (eds) Oc-ular surface disease, medical and surgical man-agement, Chap. 14. Springer, New York, pp 168–174

12. Dua HS, Azuara-Blanco A (2000) Discussion onamniotic membrane transplantation for acutechemical or thermal burns. Ophthalmology 107:990

13. Dua HS, Gomes JAP, King AJ, Maharajan VS(2004) The amniotic membrane in ophthalmolo-gy. Surv Ophthalmol 49:51–77

14. Dua HS, Gomes JAP, Singh A (1994) Corneal ep-ithelial wound healing. Br J Ophthalmol 78:401–408

15. Dua HS, King AJ, Joseph A (2001) A new classifi-cation of ocular surface burns. Br J Ophthalmol85:1379–1383

16. Fortunato SJ, Lombardi SJ, Menon R (2004)Racial disparity in membrane response to infec-tious stimuli: a possible explanation for observeddifferences in the incidence of prematurity. Am JObstet Gynecol 190:1557–1562

17. Fortunato SJ, Menon R (2002) Screening of novelmatrix metalloproteinases (MMPs) in human fe-tal membranes. J Assist Reprod Genet 19:483–486

18. Fujisato T, Tomihata K, Tabata Y, Iwamoto Y, Bur-czak K, Ikada Y (1999) Cross-linking of amnioticmembranes. J Biomater Sci Polym Ed 10:1171–1181

19. Fujishima H, Shimazaki J, Shinozaki N, Tsubota K(1998) Trabeculectomy with the use of amnioticmembrane for uncontrollable glaucoma. Oph-thalmic Surg Lasers 29:428–433

20. Fukuda K, Chikama T, Nakamura M, Nishida T(1999) Differential distribution of subchains ofthe basement membrane components type IVcollagen and laminin among the amniotic mem-brane, cornea, and conjunctiva. Cornea 18:73–79

21. Haberal M, Oner Z, Bayraktar U, Bilgin N (1987)The use of silver nitrate-incorporated amnioticmembrane as a temporary dressing.Burns 13:159–163

References 31

22. Hao Y, Ma DH, Hwang DG (2000) Identificationof antiangiogenic and anti-inflammatory pro-teins in human amniotic membrane. Cornea19:348–352

23. Joseph A, Dua HS, King AJ (2001) Failure of amni-otic membrane transplantation in treatment ofacute chemical burns. Br J Ophthalmol 85:1065–1069

24. Keelan JA, Sato T, Mitchell MD (1997) Interleukin(IL) -6 and IL-8 production by human amnion:regulation by cytokines, growth factors, gluco-corticoids, phorbol esters, and bacterial lipo-polysaccharide. Biol Reprod 57:1438–1444

25. Keene DR, Sakai LY, Lunstrum GP, Morris NP,Burgeson RE (1987) Type VII collagen forms anextended network of anchoring fibrils. J Cell Biol104:611–621

26. Kim JC, Tseng SCG (1995) The effects on inhibi-tion of corneal neovascularization after humanamniotic membrane transplantation in severelydamaged rabbit corneas. Korean J Ophthalmol9:32–46

27. Kim JC, Tseng SCG (1995) Transplantation of pre-served human amniotic membrane for surfacereconstruction in severely damaged rabbitcorneas. Cornea 14:473–484

28. Koizumi NJ, Fullwood NJ, Bairaktaris G et al.(2000) Cultivation of corneal epithelial cells onintact and denuded human amniotic membrane.Invest Ophthalmol Vis Sci 41:2506–2513

29. Koizumi NJ, Inatomi TJ, Sotozono CJ, FullwoodNJ, Quantock AJ, Kinoshita S (2000) Growth fac-tor mRNA and protein in preserved human amni-otic membrane. Curr Eye Res 20:173–177

30. Kruse FE, Joussen AM, Rohrschneider K, You L,Sinn B, Baumann J, Volcker HE (2000) Cryopre-served human amniotic membrane for ocularsurface reconstruction. Graefes Arch Clin ExpOphthalmol 238:68–75

31. Kubo M, Sonada Y, Muramatsu R, Usui M (2001)Immunogenicity of human amniotic membranein experimental xenotransplantation. InvestOphthalmol Vis Sci 42:1539–1546

32. Lee HS, Kim JC (1996) Effect of amniotic fluid incorneal sensitivity and nerve regeneration afterexcimer laser ablation. Cornea 15:517–524

33. Linnala A, Balza E, Zardi L, Virtanen I (1993) Hu-man amnion epithelial cells assemble tenascinsand three fibronectin isoforms in the extracellu-lar matrix. FEBS Lett 317:74–78

34. Ma DH, See LC, Liau SB, Tsai RJ (2000) Amnioticmembrane graft for primary pterygium: compar-ison with conjunctival autograft and topical mit-omycin C treatment. Br J Ophthalmol 84:973–978

35. Martinez Pardo ME, Reyes Frias ML, RamosDuron LE, Gutierrez Salgado E, Gomez JC, MarinMA, Luna Zaragoza D (1999) Clinical applicationof amniotic membranes on a patient with epider-molysis bullosa. Ann Transplant 4:68–73

36. Meinert M, Eriksen GV, Petersen AC, Helmig RB,Laurent C, Uldbjerg N, Malmstrom A (2001) Pro-teoglycans and hyaluronan in human fetal mem-branes. Am J Obstet Gynecol 184:679–685

37. Meller D, Pires RT, Mack RJ, Figueiredo F, Heili-genhaus A, Park WC, Prabhasawat P, John T,McLeod SD, Steuhl KP, Tseng SC (2000) Amnioticmembrane transplantation for acute chemical orthermal burns. Ophthalmology 107:980–990

38. Meller D, Tseng SC (2000) Amniotic membranetransplantation with or without limbal allograftsin corneal surface reconstruction in limbal defi-ciency. Ophthalmologe 97:100–107

39. Modesti A, Scarpa S, D’orazi G, Simonelli L,Caramia FG (1989) Localization of type IV and Vcollagens in the stroma of human amnion. ProgClin Biol Res 296:459–463

40. Muralidharan S, Gu J, Laub GW, Cichon R,Daloisio C, McGrath LB (1991) A new biologicalmembrane for pericardial closure. J BiomedMater Res 25:1201–1209

41. Park KH, Chaiworapongsa T, Kim YM, Espinoza J,Yoshimatsu J, Edwin S, Gomez R, Yoon BH,Romero R (2003) Matrix metalloproteinase 3 inparturition, premature rupture of the mem-branes, and microbial invasion of the amnioticcavity. J Perinat Med 31:12–22

42. Pires RT, Tseng SC, Prabhasawat P, Puangsrichar-ern V, Maskin SL, Kim JC, Tan DT (1999) Amniot-ic membrane transplantation for symptomaticbullous keratopathy. Arch Ophthalmol 117:1291–1297

43. Pollard SM, Aye NN, Symonds EM (1976) Scan-ning electron microscope appearances of normalhuman amnion and umbilical cord at term. Br JObstet Gynaecol 83:470–477

44. Prabhasawat P, Barton K, Burkett G, Tseng SC(1997) Comparison of conjunctival autografts,amniotic membrane grafts, and primary closurefor pterygium excision. Ophthalmology 104:974–985

45. Rao CV, Carman FR Jr, Chegini N, Schultz GS(1984) Binding sites for epidermal growth factorin human fetal membranes. J Clin EndocrinolMetab 58:1034–1042

46. Rao TV, Chandrasekharam V (1981) Use of dryhuman and bovine amnion as a biological dress-ing. Arch Surg 116:891–896


47. Reisenberger K, Egarter C, Knofler M, Schiebel I,Gregor H, Hirschl AM, Heinze G, Husslein P(1998) Cytokine and prostaglandin productionby amnion cells in response to the addition of dif-ferent bacteria. Am J Obstet Gynecol 178:50–53

48. Robson MC, Krizek TJ (1973) The effect of humanamniotic membranes on the bacteria populationof infected rat burns. Ann Surg 177:144–149

49. Robson MC, Samburg JL, Krizek TJ (1972) Quan-titative comparison of biological dressings. SurgForum 23:503–505

50. Rowe TF, King LA, MacDonald PC, Casey ML(1997) Tissue inhibitor of metalloproteinase-1and tissue inhibitor of metalloproteinase-2 ex-pression in human amnion mesenchymal and ep-ithelial cells. Am J Obstet Gynecol 176:915–921

51. Sakuragawa N, Elwan MA, Uchida S, Fujii T,Kawashima K (2001) Non-neuronal neurotrans-mitters and neurotrophic factors in amniotic ep-ithelial cells: expression and function in humansand monkey. Jpn J Pharmacol 85:20–23

52. Shimazaki J, Yang HY, Tsubota K (1997) Amnioticmembrane transplantation for ocular surface re-construction in patients with chemical and ther-mal burns. Ophthalmology 104:2068–2076

53. Smieja Z, Zakar T, Walton JC, Olson DM (1993)Prostaglandin endoperoxide synthase kinetics inhuman amnion before and after labor at term andfollowing preterm labor. Placenta 14:163–175

54. Sorsby A, Haythorne J, Reed H (1947) Furtherexperience with amniotic membrane grafts incaustic burns of the eye. Br J Ophthalmol 31:409–418

55. Sorsby A, Symmons HM (1946) Amniotic mem-brane grafts in caustic burns of the eye (burns ofsecond degree). Br J Ophthalmol 30:337–345

56. Steinkogler FJ, Haddad R (1985) Conjunctival re-placement – comparative animal experimentstudy. Klin Monatsbl Augenheilkd 187:359–360

57. Toth P, Li X, Rao CV (1993) Expression of hCG/LHreceptor gene and its functional coupling to theregulation of cyclo-oxygenase-1 and -2 enzymesin human fetal membranes. Placenta 14:A78

58. Tseng SC, Prabhasawat P, Barton K et al. (1998)Amniotic membrane transplantation with orwithout limbal allografts for corneal surface re-construction in patients with limbal stem celldeficiency. Arch Ophthalmol 116:431–441

59. Tyszkiewicz JT, Uhrynowska-Tyszkiewicz IA,Kaminski A, Dziedzic-Goclawska A (1999) Am-nion allografts prepared in the central tissuebank in Warsaw. Ann Transplant 4:85–90

60. Uchida S, Inanaga Y, Kobayashi M, Hurukawa S,Araie M, Sakuragawa N (2000) Neurotrophicfunction of conditioned medium from humanamniotic epithelial cells. J Neurosci Res 62:585–590

61. van Herendael BJ, Oberti C, Brosens I (1978) Mi-croanatomy of the human amniotic membranes.A light microscopic, transmission, and scanningelectron microscopic study. Am J Obstet Gynecol131:872–880

62. Yurchenco PD, Ruben GC (1987) Basement mem-brane structure in situ: evidence for lateral asso-ciations in the type IV collagen network. J CellBiol 105:2559–2568

63. Zhang Q, Shimoya K, Moriyama A, Yamanaka K,Nakajima A, Nobunaga T, Koyama M, Azuma C,Murata Y (2001) Production of secretory leuko-cyte protease inhibitor by human amniotic mem-branes and regulation of its concentration in am-niotic fluid. Mol Hum Reprod 7:573–579

References 33

Transplantation of Limbal Stem Cells

Harminder S. Dua

3

∑ Most self renewing tissues are served by apopulation of stem cells

∑ Potency and plasticity are two importantcharacteristics of stem cells.They may alsohave the potential to transdifferentiate

∑ Stem cells usually reside in a defined ’niche’.The corneal epithelial stem cells are be-lieved to be located in the limbal palisades

∑ Clinical and laboratory evidence stronglysupports the notion that corneal epithelialstem cells are located at the limbus but nomarker yet exists that can positively identi-fy a limbal stem cell

∑ Limbal stem cell deficiency can be congeni-tal or acquired. Ocular surface burns, im-mune mediated ocular surface diseasesand chronic inflammation are importantcauses of limbal stem cell deficiency

∑ The effects of limbal stem cell deficiencycan range from mild, such as loss of limbalanatomy or conjunctivalisation of the pe-ripheral cornea, to severe, such as cornealinvasion by a thick fibrovascular pannus orpersistent epithelial defects with stromalmelts

∑ The diagnosis of limbal stem cell deficiencyis essentially clinical, but impression cytol-ogy may help. Presence of goblet cells onthe cornea is diagnostic

∑ Limbal stem cell deficiency can be unilater-al or bilateral, partial or total

∑ Mild cases of partial deficiency can betreated by sequential sector conjunctivalepitheliectomy

∑ Total unilateral cases can be treated withauto-limbal transplantation

∑ Bilateral cases often require allo-limbaltransplantation from living related or cadaver donors

∑ Auto-limbal and living related donor trans-plantation should be avoided in the pres-ence of active inflammation. Auto-limbaltransplantation should be avoided in unilat-eral manifestation of a systemic disease

∑ Amniotic membrane transplant can becombined with any of the limbal transplantprocedures

∑ Allografts usually require long-term systemic (and/or topical) immunosuppres-sion

∑ All associated pathology such as lid malpo-sitions, trichiasis, secondary glaucoma andcataracts should ideally be managed priorto considering limbal transplantation,as far as clinically possible

∑ Buccal mucosa grafts help restore somemoisture to a dry ocular surface. Living tissue transplants usually do not survive in a dry environment

∑ Long-term outcomes of auto-limbal trans-plants are far better than those of allo-limbal transplants

∑ Ex vivo expansion of limbus derived epithe-lial cells as a sheet on different substratescan also be used in ocular surface recon-struction with good results but is also sub-ject to immune rejection

Core Messages

3.1Introduction

In this chapter the general characteristics ofstem cells (SC) and their niche are first de-scribed. The evidence supporting the existenceof SC at the corneoscleral limbus, both clinicaland scientific, is then explored providing thescientific basis for the transplantation of thelimbus in the management of limbal SC defi-ciency. A brief account of the causes and effectsof SC deficiency is provided as background tothe indications and different techniques of lim-bal SC transplantation. The surgical techniquesare elaborated together with postoperativemanagement and any adjunctive proceduresthat may complement limbal transplantation.

Ex vivo expanded limbal stem cells on amni-otic membrane or other substrates can also beused in ocular surface reconstruction. This con-stitutes a distinct method of putative limbal stemcell transplantation and is the subject of anotherchapter in this book. This technique of preparingthe tissue construct and stem cell transplantationis therefore omitted from this chapter.

3.2Stem Cells

3.2.1Definition

Stem cells are progenitor cells that are responsi-ble for cellular replacement and tissue regener-ation. They are the ultimate source cells fromwhich arise almost all other cells that constitutea given organ served by the SC. SC can be foundin both embryonic and adult tissues and repre-sent only a very small proportion (0.01–10%) ofthe total cell mass [1, 32, 44].

3.2.2Characteristics of Stem Cells

Stem cells are poorly differentiated or undiffer-entiated, long-lived, slow cycling but highlyclonogenic cells that have a high capacity for

self renewal and an increased potential for er-ror-free division [65–67]. They have the abilityto proliferate indefinitely [32] and generally livefor the duration of the organ(ism) in which theyreside. A constant pool is maintained by differ-ent strategies of cell division. The most accept-ed strategy is that of asymmetric cell divisionwhereby one daughter cell stays back in thestem cell niche and the other follows the path of proliferation and differentiation, acquiringfunctional characteristics of the tissue or organ.The same balance can be maintained if the twodaughter cells from one SC proceed down thepath of differentiation and the two daughtercells of another SC stay back in the niche as stemcells [57, 61].

The daughter cell(s) that step outside thestem cell pool are destined to divide and differ-entiate with the acquisition of features thatcharacterise the specific tissue. Such a cell iscalled a ‘transient amplifying cell’ (basal cornealepithelial cells) and is less primitive than itsparent stem cell. It is believed in some quartersthat there exists a window of opportunity dur-ing which some of these cells (‘transient cells’)[54, 55] can revert to the SC pool as SC. Transientamplifying cells divide more frequently thanstem cells but have a limited proliferative poten-tial and are considered the initial step of a path-way that results in terminal differentiation.They differentiate into ‘postmitotic cells’ (wingcells) and finally to ‘terminally differentiatedcells’ (superficial squamous cells). Both postmi-totic and terminally differentiated cells are in-capable of cell division. All cells except stemcells have a limited life span and are destined todie [45, 66, 84].

Potency and plasticity are two key attributesof SC. SC have the potential to give rise to differ-ent cell lineages. This potency is, however, notuniform and there exists amongst SC a hierar-chy of potential. SC can be totipotent, pluripo-tent, multipotent, or unipotent. The zygote thatcan form the entire embryo and part of the pla-centa is an example of a totipotent cell. Cells ofthe inner cell mass from which most tissues thatarise from the three germ layers can be derived,but not components of the placenta, are consid-ered pluripotent. Most tissue specific SC aremultipotent, capable of producing lineages that

36 Chapter 3 Transplantation of Limbal Stem Cells

can differentiate in two, three or more differentcell types with functional attributes of the organin which they reside. Some SC, in their naturalenvironment, may have only limited potentialwith the ability to generate only one specific celltype. These SC are labelled unipotent SC orcommitted progenitors.SC of the epidermis andthe corneoscleral limbus are considered to beexamples of this category [1, 3, 68].

The ‘plasticity’ of SC refers to their ability totransdifferentiate. Some SC when relocated to adifferent site (tissue) can assume the role thatsupports the structure and function of the newsite, thus aiding in regeneration, repair andmaintenance of the cell population at the newsite. Stem cell potential and plasticity are bothmore pronounced in embryonic SC comparedto adult SC. Embryonic SC have virtually an un-limited potential for self-renewal and differenti-ation. Given the right microenvironment andthe right signals, adult and embryonic SC can(theoretically) be made to follow a desired pathof differentiation or propagated indefinitely, inan undifferentiated state. Herein lies the im-mense therapeutic potential of SC [3].

3.2.3The Stem Cell ’Niche’

The microenvironment in which the SC reside isreferred to as their ‘niche’ [60]. SC are usuallyconfined to their ‘niche’ where the microenvi-ronment supports and maintains the stemnessof SC and affords a degree of protection. In sol-id organs, where cell migration commences atone point and progresses until the cells are shedat a distant point(s), the SC niche is usually lo-cated at the point of commencement. The ‘niche’represents the collective influence of other localmatrix cells, the extracellular matrix, its vascu-larity, basement membrane characteristics andprevalent growth factors and other cytokines. Inthe intestinal mucosa, for example, it is believedthat the pericryptal fibroblasts/subepithelialmyofibroblasts may serve as niche cells [60, 79]and in the epidermis, beta 1 integrin mediatedadhesion to its ligand, type IV collagen, isshown to influence behaviour of epidermal SC[44]. The niche also affords protection to the all-important SC [79, 95].


∑ Stem cells:– Undifferentiated– Long lived– Slow cycling– Clonogenic– Asymmetric division– Potency: usually pluripotent

or multipotent– Plasticity: transdifferentiation– Niche: SC microenvironment

∑ SC progeny:– ‘Transient cells’– Transient amplifying cells –

basal epithelium– Postmitotic cells – wing cells– Terminally differentiated cells –

superficial squamous cells

3.3Limbal Stem Cells

3.3.1The Clinical Evidence

During corneal epithelial wound healing andnormal epithelial turnover, cell migration andmigration of sheets of epithelium [20] havebeen shown to occur in a centripetal mannerfrom the corneoscleral limbus towards the cen-tre of the cornea [4, 5, 51]. Large corneal epithe-lial wounds, where the wound edge is closer tothe limbus, heal at a faster rate than smallerwounds [59]. Repeated denudation of the cen-tral corneal epithelium shows that the healingrate of the second wound is more rapid thanthat of the first. This suggests that rapidly divid-ing younger cells of the periphery have movedto more central areas after the first trauma andrespond readily to the second [80].

Human corneal epithelial defects with par-tial limbal involvement demonstrate a preferen-tial circumferential migration of a population ofcells along the limbus, from both ends of theremaining intact limbus [21] (Fig. 3.1A). Com-plete epithelial cover for the corneal surface isnot established until limbal re-epithelializationis first complete, suggesting that the circumfer-

3.3 Limbal Stem Cells 37

entially migrating population of cells probablyrepresents in part the healing response of limbalstem cells. In patients with limbal abnormali-ties, alternating columns of normal and fluores-cein staining cells often corresponding to limbalpalisades – columnar keratopathy – have beennoted to extend from the limbus towards the centre in radial or curvilinear rows [22](Fig. 3.1B). The palisades of Vogt and the inter-

palisade rete ridges provide a unique structureto the limbus (Fig. 3.2). The structure of the pal-isades and the rete ridges, their vascularity andpigmentation are all analogous to repositoriesof stem cells in the monkey palm epidermis [9, 86, 88]. It has also been reported thathemidesmosomes of peripheral cells of normaland healing mouse corneas are arranged in ra-dial rows, leading to the interpretation that this


Fig. 3.1. A Healing of corneal epithelial wound in-volving the limbus showing a preferential circumfer-ential migration of tongue-shaped sheets of limbalepithelial cells arising from either end of the remain-ing intact epithelium. (Slit lamp anterior segmentphotograph with fluorescein dye) (with permissionfrom Br J Ophthalmol: Dua et al. 2001; 85:1379–1383).

B ‘Columnar keratopathy’ is the name given by theauthor to this presentation of alternating columns offluorescein stained epithelium and normal cornealepithelium. These correspond to the limbal palisadesand represent an early sign of limbal stem cell defi-ciency

A B

Fig. 3.2. Slit lamp photograph of the limbus showing the palisade (of Vogt) structure with: A pigment columnsmigrating into peripheral cornea and B fluorescein staining of columnar migration in response to a centralabrasion

A B

orientation represents centripetal migration ofepithelial cells [5]. Very recently, a uniqueanatomical structure, termed the limbal epithe-lial crypt [27], has been discovered at the pe-ripheral end of the interpalisade rete ridges,numbering approximately five to seven per hu-man cornea. This has features consistent withthose of a SC repository or ‘niche’.

3.3.2The Scientific Evidence

Basic research has identified a number of char-acteristics that are unique to the limbal basalepithelial cells and set them apart from the rest:Mitosis rates are highest at the limbus, both inthe normal physiological state and followingstimulation [31, 37, 8]. Limbal epithelial cellshave the greatest proliferative potential in vitro,compared to any other part of the cornea[28–30]. Limbal basal cells lack the epithelialcell differentiation cytokeratin CK3 [18, 56, 73,101]. Impression cytology examination of thehuman limbus shows that, morphologically, thelimbal cells are smaller, more densely packedand have a greater nucleus to cytoplasm ratiocompared to adjacent corneal and conjunctival

cells (H.S. Dua, unpublished observations,Fig. 3.3A, B).

Several other attributes that are unique to thelimbal epithelium (Table 3.1) have also been de-scribed. These include the presence of alpha-enolase [101, 102], EGF receptors [100, 103], pig-ment [9], cytokeratin profile (CK3/12 negative)[7, 73], presence of vimentin [53–55], CK19 andspecific basement membrane characteristics[34, 35, 85]. Vimentin and CK19 positive, CK3negative clusters of cells with unique electronmicroscopic morphology have been demon-strated [54, 55]. Connexin 43 (Cx43), a gap junc-tion protein, has been noted in human cornealbut not limbal basal epithelium [12,58,96]. It hasbeen proposed that absence of Cx43 segregatescells from adverse events generated in neigh-bouring cells and helps preservation of SC intheir microenvironmental niche [96].

Zhao et al. [98] have recently reported thatlimbal epithelial cells cultured in the presence ofmitogens express neural progenitor markers,specifically nestin. A transcriptional factor, p63involved in morphogenesis, has been proposedto identify keratinocyte stem cells at the limbus[63], but its role as a marker of limbal SC is con-troversial [25, 46]. Similarly, well defined mark-ers of haematopoietic SC, namely CD34 and

3.3 Limbal Stem Cells 39

Fig. 3.3. A Impression cytology specimen of the hu-man limbus from an eye bank donor eye. The limbalcells are smaller, tightly packed and show a greaternuclear-cytoplasmic ratio. B Montage of the humanlimbus, peripheral cornea and conjunctiva

A

B

CD133, have failed to demonstrate any uniquesubpopulation of cells at the limbus [25, 47]. AnATP-binding cassette transporter protein,ABCG2, is believed to be a marker of a side pop-ulation of cells that have the ability to effluxHoechst 3342 dye [99]. Side population cells thatcontain this transporter protein are believed tobe stem cells [36]. Limbus epithelial cells havebeen shown to express ABCG2 [94] and thesemay represent the subpopulation that containthe stem cells. The limbal epithelial crypt re-cently demonstrated by Dua et al. [27] containscells that predominantly stain positive forABCG2, indicating that the crypt may providethe niche for corneal epithelial SC (Fig. 3.4).

The above data strongly supports the notionthat progenitor cells exist at the corneosclerallimbus. Whether these are truly SC as defined inother organ systems remains to be established.There is evidence to suggest that SC or progeni-tor cells for the conjunctival epithelium residemaximally in the fornices and for goblet cells andperhaps for conjunctival epithelium may also bescattered throughout the epithelial surface.


∑ Evidence for corneal epithelial (limbal)stem cells:

∑ Clinical:– Unique palisade architecture– Centripetal migration from limbus– Circumferential migration along limbus– Pigment and other deposits migrating in

columnar manner from limbus– Larger corneal epithelial wounds

(closer to limbus) heal faster– Second wounds heal faster– Relative resistance of limbus epithelium

to denudation– Columnar keratopathy– Limbal deficiency allows conjunctivali-

sation of cornea and persistent epithelialdefects

∑ Scientific:– Different morphology of limbal cells– Increased hemidesmosomes at limbus

basal epithelium


Table 3.1. Differences between epithelial cells of the limbus and central cornea (CK, cytokeratin; CX,connexin; EGFR, epithelial growth factor receptor)

Limbus Central cornea

CK 5/14+ve CK 5/14–veCK 3/12–ve CK 3/12+veCK 19+ve CK 19–veP63+ve P63–ve but see results in this paperCX 43–ve CX 43+Vimentin+ve Vimentin–veIntrinsic melanogenesisCytochrome oxidase and ATpase+veAlpha-enolase+ve Alpha-enolase+veBeta-1-integrin+ve Beta-1-integrin+veEGFR+ve (strong) EGFR+veABCG2+ve ABCG2–ve

Fig. 3.4. Limbal epithelial crypt: representing a solidcord of cells extending from the undersurface of alimbal palisade. These cells are positive for the puta-tive stem cell marker ABCG2. Haematoxylin stainedcryo section, ¥100

– Increased mitosis rates at limbus– Increased proliferative potential

of limbal basal cells– Absence of cytokeratin 12 in limbal

basal cells– Absence of gap junctions in limbal

basal cells– Presence of certain enzymes such

as alpha-enolase and ABCG2– Different basement characteristics

at limbus compared to central cornea– Presence of limbal epithelial crypts

(niche)

3.4Limbal Stem Cell Deficiency

3.4.1Causes of Limbal Stem Cell Deficiency

Stem-cell deficiency can be congenital or ac-quired. Congenital SC deficiency occurs as a re-sult of hereditary aplasia of limbal stem cells asoccurs in aniridia and congenital erythrokera-todermia. More often though, stem cell defi-ciency is acquired as a result of extraneous in-sults that acutely or chronically destroy limbalstem cells. These include chemical or thermalinjuries, ultraviolet and ionising radiation,Stevens-Johnson syndrome, advanced ocularcicatricial pemphigoid, multiple surgery orcryotherapy, contact lens wear, or extensive/chronic microbial infection such as trachoma.Keratitis associated with multiple endocrine de-ficiencies, neurotrophic (neural and ischaemic)keratopathy and chronic limbitis also lead even-tually to SC deficiency but are less common [13,18, 24, 33, 41, 42].


∑ Causes of limbal stem cell deficiency:∑ Congenital: aniridia, erythrokeratodermia

Acquired:– Chemical and thermal burns– Chronic inflammatory disorders– Progressive cicatrisation conditions –

OCP, SJS– Prolonged contact lens wear– Multiple ocular surface surgery

– Medicamentosa including preservatives– Idiopathic

3.4.2Effects of Limbal Stem Cell Deficiency(Modified from Dua et al. [25])

The hallmark of limbal stem cell deficiency is‘conjunctivalisation’ of the cornea and the mostsignificant clinical manifestation is a persistentcorneal epithelial defect.

The clinical symptoms of limbal deficiencymay include decreased vision, photophobia,tearing, blepharospasm, and recurrent episodesof pain (epithelial breakdown), as well as a his-tory of chronic inflammation with redness.

Depending on the extent of limbal involve-ment, SC deficiency can be partial or total. Par-tial SC deficiency may vary in extent to involvethe pupillary area, when intervention is usuallyrequired, or exclude the visual axis when noneor minimal intervention with topical medica-tion may be required. Further, partial SC defi-ciency may vary in severity from mild, whenonly an abnormal epithelial sheet covers a vari-able area of the cornea, to severe when a part ofthe cornea, usually including the pupillary area,is covered by a thick fibrovascular pannus.

The clinical features of SC deficiency, frommild to severe, include the following [13, 14, 18,21, 22, 24, 43, 64, 88]: (a) loss of limbal anatomy,(b) irregular, thin epithelium, (c) stippled fluo-rescein staining of the area covered by abnor-mal epithelium, (d) unstable tear film, (e) fila-ments and erosions, (f) superficial and deepvascularisation, (g) persistent epithelial defectsleading to ulceration, melting and perforation,(h) fibrovascular pannus, and (i) scarring, kera-tinisation and calcification.1. Loss of limbal anatomy: The normal limbal

architecture with rows of palisades and theperilimbal vascular arcade is usually best de-fined at the superior and inferior limbus. Thearchitecture may vary depending on age ofthe individual. With increasing age the defi-nition of palisades becomes less distinctnasally and temporally. Pigmentation of thelimbal palisades is a feature in some races.Alterations in limbal anatomy include con-

3.4 Limbal Stem Cell Deficiency 41

tiguous or patchy fluorescein staining ofconjunctiva derived cells at the limbus andextending onto the peripheral cornea, seg-mental limbal hyperaemia indicating chron-ic inflammation, thickening of limbal epithe-lium, vascularisation of peripheral corneaand scarring (Figs. 3.1B, 3.5A, B).

2. Irregular, thin epithelium: When the initialinjury is mild and superficial or the diseaseprocess leading to stem cell deficiency isslowly progressive, loss of a segment oflimbal epithelium may occur without signif-icant damage to the substratum. A sheet ofconjunctival/metaplastic epithelium conse-quently covers the cornea without any no-table vascularisation. This epithelium is usu-ally thin and irregular as can be seen by thepooling of fluorescein dye at the junction ofthe abnormal and remaining normal epithe-lium (Fig. 3.6, see also Fig. 3.11A) [14].

3. Stippled fluorescein staining of the area cov-ered by abnormal epithelium: The abnormalconjunctival/metaplastic epithelium readilytakes up fluorescein dye [43], allowing easyvisualisation of the abnormal cells and theirpattern of distribution. The abnormal fluo-rescein-staining ‘conjunctivalised’ epitheli-um may take on the pattern of columns,whorls or wedges with the broad base to-wards the limbus and the narrow curvingapex toward the corneal centre (Figs. 3.5A,3.6) [22].

4. Unstable tear film: The abnormal epitheliumdemonstrates a rapid tear film break up time

over it and areas of negative and positive flu-orescein staining.

5. Tags of loose epithelium, filaments with mu-cus and recurrent erosions are other featuresassociated with the abnormal epithelial cov-er on the cornea.

6. Superficial and deep vascularisation: Inmoderate to severe cases of stem cell defi-ciency, superficial and/or deep vascularisa-tion of the cornea occurs. It is largely re-stricted to the area of stem cell deficiencyand may affect a segment of the limbus or theentire circumference may become involved(Fig. 3.7).

7. Persistent epithelial defects (Fig. 3.8): Chron-ic non-healing ulceration of the cornealepithelium or cycles of repeated breakdown


Fig. 3.5. A Signs of mild limbal stem cell deficiency – peripheral conjunctivalisation highlighted with fluo-rescein staining. The junction of corneal and conjunctival phenotypes of epithelia is marked with arrows.B Peripheral vascularisation with loss of limbal architecture

A B

Fig. 3.6. Peripheral conjunctivalisation with poolingof dye and stippled staining of the abnormal epitheli-um, between 12 and 3 o’clock

followed by healing, associated with a chron-ic low grade inflammation, is a feature of lim-bal stem cell deficiency. These defects are li-able to lead to deep stromal infiltrates thatmay or may not be related to infection. Theedges of the epithelial defect have a distinctrolled-up or heaped appearance. Over time,progressive melting of the corneal stromawith perforation can occur.

8. Fibrovascular pannus: In moderate to severecases of stem cell deficiency, epithelial coverof the denuded cornea is associated with en-croachment of fibrovascular tissue of vary-

ing thickness (Figs. 3.7, 3.8) [49]. This tissuesupports the thickened multilayered con-junctiva derived epithelium.

9. Scarring, keratinisation and calcification:The end stage of the aftermath of limbal stemcell deficiency, whatever the cause, is scar-ring and eventually calcification of theaffected tissue. Usually by this stage the in-flammation has subsided and the eye is com-paratively comfortable. In patients who haveassociated severe dry eyes the covering ep-ithelium becomes totally or partially kera-tinised (Fig. 3.9A, B).

3.4 Limbal Stem Cell Deficiency 43

Fig. 3.7. Superficial and deep vascularisation with afibrovascular pannus encroaching on the cornea fol-lowing chemical burn in which 9.5 clock hours of thelimbus and 60% of the conjunctiva were involved(clinical grade – 9.5/60%) (with permission from Br JOphthalmol: Dua et al. 2001; 85:1379–1383)

Fig. 3.8. Persistent epithelial defect and fibrovascu-lar pannus on cornea related to total stem cell defi-ciency following unilateral alkali (cement) burn(clinical grade 12/65%) (with permission from Br JOphthalmol: Dua HS and Azuara-Blanco A 2000;84:273–278)

Fig. 3.9. A Right eye of patient with 10 clock hours oflimbus and 70% conjunctival involvement followinga chemical (alkali) burn. B Left eye of same patientwith 12 clock hours of limbs and 90% conjunctival in-

volvement (clinical grade 10/70% RE, 12/90% LE).Scarring, vascularisation, adhesions and some kera-tinisation are present. The lids on both sides were alsoseverely damaged

A B


∑ Effects of limbal stem cell deficiency– Mild Æ severe– Loss of limbal anatomy– Conjunctival epithelial ingress onto

cornea – stippled fluorescein staining– Columnar keratopathy– Unstable tear film over affected area– Frank conjunctivalisation– Corneal vascularisation –

superficial and deep– Fibrovascular pannus covering corneal

surface– Persistent epithelial defect– Stromal melting– Perforation, scarring, calcification– Keratinisation

3.4.3Diagnosis of Stem Cell Deficiency

The diagnosis of stem cell deficiency remainsessentially clinical. On slit lamp biomicroscopicexamination, the conjunctivalised cornea pres-ents a dull and irregular reflex. The epitheliumis of variable thickness and translucent toopaque. Conjunctival epithelium on the corneaappears to be more permeable than corneal ep-ithelium and takes up fluorescein stain in a stip-pled or punctate manner. In cases of partialconjunctivalisation of the cornea, fluorescein

dye tends to pool along the junction of thesheets of corneal and conjunctival epithelial cellphenotypes.At this junction, the corneal epithe-lial sheet shows tiny processes or undulationsthat give the junction its characteristic appear-ance.

Loss of architecture of the limbal palisades ofVogt and vascularisation are other common fea-tures. When damage is extensive, vascularisa-tion occurs in the form of fibrovascular pannus,which increases the thickness of the affectedarea of the cornea. However, the underlyingcorneal stroma may be considerably thinned bythe initial insult of disease process.

The presence of goblet cells on impressioncytology specimens taken from the corneal sur-face or in biopsy specimens of the fibrovascularpannus covering the cornea is pathognomonicof conjunctivalisation of the cornea (Fig. 3.10A)[25, 69]. Biopsy specimens also demonstrate amultilayered, at times keratinised epitheliumoverlying dense fibrous and vascular tissue(Fig. 3.10B). Intraepithelial lymphocytes, whichare a feature of conjunctival epithelium, are alsoseen on conjunctivalised corneal epithelium.These are predominantly CD8+/*HML-1 + cells(cytotoxic T lymphocytes expressing the hu-man mucosal lymphocyte antigen) [23, 25]. Fea-tures of squamous metaplasia or loss of corneaspecific cytokeratins (CK 3/12) on immunohis-tology are other effects noted on biopsy speci-mens.


Fig. 3.10. A Impression cytology from surface of cornea with stem cell deficiency and a fibrovascular pannusshowing goblet cells. PAS stain, ¥400. B Biopsy of fibrovascular pannus showing multilayered epithelium,vascularisation and intraepithelial lymphocytes along the basal layers

A B


∑ Diagnosis of limbal stem cell deficiency– Essentially clinical– Impression cytology – goblet cells on

cornea pathognomic– Biopsy – multilayered epithelium,

intraepithelial lymphocytes, vessels– Vimentin and CK 19 positive cells in

central cornea (normally present in peripheral cornea and limbus)

3.5Limbal Transplant Surgery

3.5.1Principles

Management of stem cell deficiency can be con-sidered in the following steps:

After Acute Injury. When a patient presentsafter an acute insult it should be ascertainedwhether the involvement of the limbus is partialor total. This can be done by use of fluoresceinstain and slit lamp examination. If partial,appropriate medication required for the under-lying cause and to control inflammation shouldbe initiated. The eye should be examined at 24-or 48-h intervals and the process of re-epithe-lialisation observed. If this is occurring from theremaining intact limbal epithelium [21], thisshould be encouraged and any attempt at re-ep-ithelialisation from the conjunctival epitheliumshould be discouraged by sequential sectoralconjunctival epitheliectomy (SSCE, see below)[14, 15]. If total, allow the cornea to be covered byconjunctival epithelium, if possible, before con-templating surgical intervention. This may takeseveral days. The guiding principle should bethat corneal epithelial cover for cornea and con-junctival epithelial cover for conjunctiva is theideal end result but conjunctival epithelial cov-er for cornea is preferable to no epithelial coverto cornea.

In Established Cases. The principles underly-ing surgical procedures involving limbal stemcells are firstly to expand the corneal epithelial

sheet derived from any existing sector of limbusin the affected eye. This can be achieved bySSCE (see below) [14, 15], especially if the corneais partially covered by a layer of thin, metaplas-tic, conjunctivalised epithelium. If no healthysector of limbus is available in the affected eyeand if the other eye is normal with a positivelydocumented absence of involvement in theoriginal injury, autologous limbal transplanta-tion should be considered. If the other eye isalso affected or the underlying condition is asystemic illness such as Stevens-Johnson syn-drome, allografts from a living related donor orfrom a cadaver donor should be considered. Inthe acute stage of limbal stem cell deficiency, forexample acute chemical burns, auto-limbal orliving related donor limbal transplants shouldbe avoided at all costs. The chances of the trans-planted material becoming caught up in the in-flammatory and scarring process are high withloss of a valuable resource for future recon-struction.Use of auto or living related donor tis-sue, if available, should be attempted in quieteyes. All the above procedures can be comple-mented with amniotic membrane transplanta-tion. Penetrating keratoplasty may be combinedwith or following any of these procedures.

Limbal transplantation involves taking alamellar strip of limbal tissue, usually withsome adjacent peripheral cornea and/or con-junctiva and transplanting it to a suitably pre-pared bed in the host eye. Sutures are usuallyrequired to keep the donor graft in place.

3.5.2Preoperative Considerations

All associated lid abnormalities, intraocularpressure problems and presence of cataractshould ideally be dealt with prior to undertakingocular surface restorative surgery. Symble-pharon correction with amniotic membrane orbuccal mucosa graft should also precede stemcell grafting. At times, if a corneal graft proce-dure is being contemplated at the time of stemcell grafting, it can be combined with cataractextraction and lens implantation.When an intu-mescent cataract is associated with raised pres-sure, corneal grafting may become a necessity if

3.5 Limbal Transplant Surgery 45

a dense fibrovascular pannus or corneal scarprecludes visualisation of the interior of the eye.

Patients with limbal SC deficiency and con-junctivalised corneal surface tend to manifestpersistent chronic inflammation. Stem cellgrafts do not perform well in the presence of in-flammation and can be destroyed by the inflam-matory and scarring processes. Ideally inflam-mation should be controlled and the eyerendered as quiet as possible with the use oftopical and systemic steroids or other immuno-suppressants which may become necessary insome conditions such as Stevens-Johnson syn-drome and ocular cicatricial pemphigoid.

Most stem cell grafts do not survive in a dry(eye) environment. At times the injurious insultresulting in stem cell deficiency also results in asevere dry eye state. In such situations, if topicallubricants including autologous serum drops,punctal occlusion and buccal mucosa grafts donot restore adequate moisture to the ocular sur-face, a keratoprothesis procedure should beconsidered.


∑ Treatment algorithm∑ General principles:

– Manage underlying factors, e.g.,chronic inflammation, contact lens wear,topical medications

– Topical lubrication– All associated problems, e.g., raised

pressure, conjunctival adhesions,lid malpositions, should be addressedbefore undertaking ocular surfacereconstruction

– Limbal transplants do not perform wellin dry eyes

∑ In acute limbus injury:– If partial, i.e. some limbus is surviving –

allow corneal epithelialisation to occurfrom limbus derived cells – SSCE

– If total:a) Allow conjunctival epithelium

to grow onto corneab) Transplant sheet of ex vivo expanded

limbal epithelial cellsc) Avoid use of autologous or living

related donor tissue until acute inflammation is well under control

∑ In established cases:– Treat eye lid problems, glaucoma and

conjunctival adhesions first– Partial or total– Partial:

a) Visual axis not involved: sympto-matic, lubricants of SSCE

b) Visual axis involved: SSCEc) Dense fibrovascular pannus:

sector limbal transplantTotal:a) Unilateral: auto-limbal transplantb) Ex vivo expansion of autologous

limbal cellsc) Bilateral: allo-limbal transplantd) Ex vivo expansion of cells (living

related, living non-related, cadaver)e) Amniotic membrane and autologous

serum drops as adjunctsf) Allo-transplants require systemic

immunosuppression

3.6Surgical Techniques

3.6.1Sequential Sector ConjunctivalEpitheliectomy (SSCE) [14, 15] (Figs. 3.11, 3.12)

In cases with partial, mild to moderate conjunc-tivalisation of the cornea, without significant fi-brovascular pannus, removal of the conjuncti-valised epithelium is all that is required. Thiscan be achieved at the slit lamp under topicalanaesthesia, using a crescent blade or a surgicalknife. It is important to remove all conjunctivalepithelium, especially along its line of contactwith the remaining corneal epithelium. Follow-ing removal of conjunctival epithelium from thecorneal surface, it is important to closely moni-tor the patient to ensure that the denuded sur-face is re-epithelialised by cells derived from theremaining corneal epithelial sheet, i.e. limbalderived cells and not by conjunctival cells. Thiscan be effected by repeatedly debriding (se-quential epitheliectomy) any conjunctival ep-ithelium that encroaches upon the limbus until


the limbus and corneal surface is re-populatedby limbal epithelium derived cells.

In cases where only 1 or 2 clock hours of lim-bal epithelium is surviving, it may be appropri-ate to attempt re-epithelialisation of the visualaxis only, with limbal derived cells. An area cor-responding to the visual axis is debrided off itsconjunctival epithelial cover and re-epitheliali-sation with limbal derived cells is achieved. This

has the theoretical advantage of not overstress-ing the small remaining sector of limbal ‘stem’cells. This technique of SSCE can also be useful-ly combined with limbal transplant to allowcells derived from transplanted limbal tissue(auto or allo) to re-populate the host cornealsurface without ‘contamination’ from conjuncti-val epithelium (see below).

3.6 Surgical Techniques 47

Fig. 3.11 A–D. Sequential sector conjunctival ep-itheliectomy (SSCE, H.S. Dua). A Conjunctivalisationof the cornea involving the visual axis followingchemical injury. The demarcation between the twophenotypes of cells is clearly visible. B Appearanceimmediately after removing the abnormal epithelium(epitheliectomy). C The corneal epithelial sheet is mi-grating across the surface but the conjunctival epithe-

lium too has started to re-encroach on the cornea.D After complete healing, the visual axis is now covered by healthy corneal epithelium. A new line ofcontact between conjunctival and corneal epitheliumis established (fluorescein stained anterior segmentphotographs). The patient’s vision improved from3/18 to 6/9 (with permission from Br J Ophthalmol:Dua HS 1998; 82:1407–1411)

A B

DC

3.6.2Auto-limbal Transplantation (Figs. 3.13, 3.14)

In patients where total stem cell deficiency affects only one eye, an auto-limbal transplantprocedure is the ideal option [6, 19, 42, 48, 49, 88,89]. It is important, however, to be absolutelycertain that the donor eye was not involved atthe time of the initial injury. In unilateral mani-festations of systemic diseases,harvesting tissuefrom the apparently normal eye is not recom-mended.

The surgical technique consists of the follow-ing steps (the author’s [19] modified techniqueis described): (a) a 16-mm Flieringa ring is su-tured in place when the procedure is to be com-bined with a corneal graft (and lens extractionwith implant). A 360° peritomy is first per-formed in the recipient eye. (b) The fibrovascu-lar pannus covering the corneal surface is dis-sected off at a suitable plane. Any bleedingpoints are individually cauterised with lightdiathermy. (c) The donor tissue consisting ofcorneal-limbal-conjunctival explants is har-vested from the contralateral normal eye. Twoexplants, corresponding to 2 clock hours(11–1 o’clock and 5–7 o’clock) and consisting of avery narrow strip (1 mm or less) of peripheralcornea, limbus and 3 mm of bulbar conjunctiva,are harvested. The conjunctival area to be re-moved is marked with a surgical marker pen.The conjunctiva is incised superficially with apair of scissors and dissected in a superficialplane up to the limbus.An angled bevelled blade


Fig. 3.12. A Conjunctivalisation of the superiorcornea involving the visual axis. B, C After SSCE with-out and with fluorescein stain, respectively. The visu-al axis is now clear

A

B

C

Fig. 3.13 A, B. Diagrammatic representation of autologous limbal transplantation. A Positioning of explantson recipient limbus at the 12 and 6 o’clock positions without or with (B) a corneal graft (with permission fromBr J Ophthalmol: Dua HS, Azuara-Blanco A 2000; 84:273–278)

is used to (lamellar) dissect the correspondinglimbal area extending into peripheral cornea tojust inside (central) to the vascular arcade. (d)Suitable beds may be prepared at the superiorand inferior limbus of the recipient eye by usingthe excised explants as templates to mark thearea to be prepared. This is not always essential.(e) The donor tissue is then sutured onto the re-cipient eye with two interrupted 10-0 nylon su-tures at the corneal margin and two along thescleral edge of the explant. Care should be takennot to bury the knots in the explant tissue as thiscould strip the explants off when attempting toremove the sutures in the postoperative period.At times the knots may be left unburied to facil-itate removal. The conjunctiva of the recipienteye is then approximated to the donor conjunc-tiva with interrupted 8-0 Vicryl sutures (ab-sorbable), taking a bite into episclera. (f) Whena penetrating keratoplasty is also required, thisis performed after the limbal explants are firstsutured in place. (g) A bandage contact lens isplaced on the cornea and subconjunctival an-tibiotics and corticosteroids are injected at theend of the procedure.

Adjunctive use of amniotic membrane canbe made either as a graft to provide a suitablebed for limbal explant derived epithelial cells togrow on the cornea and/or as a patch to preventconjunctival epithelial cells from extendingonto the cornea and admixing with the limbalexplant derived cells (see below).

3.6.3Allo-limbal Transplantation

3.6.3.1Living Related Donor

When a living related donor, who is tissuematched to the recipient, is available, tissue isharvested from one donor eye and used on therecipient eye exactly as described above forauto-limbal transplantation [10, 70].


Fig. 3.14 A–C. Auto-limbal transplantation. A Pre-operative persistent epithelial defect following an al-kali (cement) burn as shown in Fig. 3.8. B Postopera-tive status after auto-limbal transplants at the 6 and12 o’clock positions. The patient’s eye is stable over5 years postoperatively (with permission from Br JOphthalmol: Dua HS, Azuara Blanco A 2000;84:273–278. C Donor site for autologous limbal trans-plant, stained with fluorescein. Note that the centraledge of the removed tissue needs to extend just cen-tral to the limbal vascular arcade

A

B

C

3.6.3.2Cadaver Donor

In most instances, limbal tissue is obtainedfrom cadaver donor eyes [16, 41, 42, 48, 81–83, 88,89, 91]. In such an event, tissue matching is notusually practical. In the author’s protocol, a pairof ‘fresh’ donor eyes is used within 48 h of death.Donor eye retrieval should be done within 24 hof death and surgery within the next 24 h.Donor age of less than 50 years is preferred.‘Fresh’ and ‘young’ donor eyes are preferred be-cause the success of the procedure depends onthe transplantation of healthy limbal stem cells.

The surgical technique consists of the follow-ing steps (the author’s technique [16] is de-scribed) (Fig. 3.15): Donor limbus tissue is pre-pared before the patient is anaesthetised. (a)The donor eyeball is inflated with air (1–2 ml),injected through the stump of the optic nerve,to make the globe firm. (b) The globe is fixed ona Tudor Thomas stand. A vacuum (or manual)trephine with a diameter 3 mm smaller than the

corneal diameter (i.e., average of vertical andhorizontal corneal diameter) is used to trephinethe central donor cornea to one-fourth to one-fifth of the stromal depth (approximately150 mm). Proper centration is important to en-sure that a uniform width of peripheral corneais obtained. (c) Superficial lamellar dissectionof the peripheral cornea is then carried out us-ing an angle bevelled blade, and extended intothe sclerocorneal junction and 1 mm beyond,into sclera. Approximately 1–2 mm of donorconjunctiva, if present, is maintained. The dis-sected tissue is divided at one point and exci-sion completed with a curved scissors, by cut-ting along the outer circumference of thedissected tissue. The limbal tissue to be graftedthus consists of an open ring of peripheralcorneal and limbal epithelium (and conjuncti-val epithelium at places), and superficialcorneal, limbal and scleral stroma. (d) Prepara-tion of the recipient eye is similar to that de-scribed for auto-limbal transplantation exceptthat a ‘bed’ is not prepared to receive the limbal


Fig. 3.15 A–E. Diagrammatic representation of allo-limbal transplantation. A Injection of air to firm thedonor globe. B Harvesting the limbal circumferencefrom the donor globe. C Positioning of explants on re-cipient limbus without C or with D a corneal graft. As

the donor explant is placed slightly peripheral to therecipient limbus, more than one donor may be re-quired as shown. E Combined allo-limbal transplantand corneal graft (with permission from Br J Oph-thalmol: Dua HS, Azuara-Blanco A 1999; 83:414–419

A

D E

B C

ring explant. The ‘open ring’ of donor tissue isplaced on the host limbus and sutured with in-terrupted 10-0 nylon sutures at the corneal andscleral margin. Six to eight sutures are firstpassed along the inner (corneal) edge of thedonor tissue and partial thickness of hostcorneal stroma.A similar number of sutures arethen passed directly opposite to the inner su-tures, along the outer (scleral) edge of the donortissue. These are anchored to the superficialsclera of the host. The tension on these suturesdetermines the final tension on the inner su-tures. The knots are trimmed and buried. (e)This method invariably leaves a small gap (ap-proximately 5–8 mm) between the two ends ofthe donor tissue ring (superiorly). This is filledwith a piece of donor limbal tissue, cut to size,harvested from the other eye of the same donor.This piece usually requires a couple of addition-al sutures along either edge. (f) The host con-junctiva is approximated to the scleral edge ofthe transplanted limbal ring with interrupted 8-0 Vicryl sutures (absorbable). (g) A penetrat-ing keratoplasty if required at the time of sur-gery is performed after the limbal ring is su-tured in place. The donor graft for penetratingkeratoplasty (usually 7–7.5 mm) is obtainedfrom the central cornea of the donor wholeglobe. (h) A bandage contact lens is placed onthe cornea and subconjunctival antibiotics and

corticosteroids are injected at the end of theprocedure.

Adjunctive use of amniotic membrane canbe made either as a graft to provide a suitablebed for limbal explant derived epithelial cells togrow on the cornea and/or as a patch to preventconjunctival epithelial cells from extendingonto the cornea and admixing with the limbalexplant derived cells (see below).

3.6.4Adjunctive Surgery

3.6.4.1Amniotic Membrane Grafts

The amniotic membrane serves as a useful ad-junct to stem cell grafting [2, 17, 26, 50, 77, 90]. Itis commonly deployed to provide a suitablesubstratum for the transplanted limbal graftderived epithelial cells to migrate on and formadhesion complexes. After excision of the fibro-vascular tissue, if the underlying host bed isfound to be irregular and scarred, use of a 9- or10-mm disc of amniotic membrane, epithelialside up, can provide a suitable substratum forthe transplanted limbal derived epithelial cellsto migrate upon. The amniotic membrane canalso be deployed as a biological bandage to the


Fig. 3.16 A, B. Use of double amniotic membrane toprevent admixture of conjunctival and corneal ep-ithelium on the corneal surface. A The inner mem-brane disc is sutured with the epithelial surface up toact as a graft and substrate for the cells to grow on,

and the outer membrane, epithelial side down, acts asa patch. B Regenerating cells from the peritomisedconjunctiva are seen growing on the outer mem-brane. In the absence of the outer membrane (patch)these would have encroached on the corneal surface

A B

denuded corneal stroma, allowing epithelialisa-tion to occur beneath it whilst trapping inflam-matory cells and downregulating inflammationand scarring at the same time. Two amnioticmembranes, one inner one serving as a graftand one outer membrane serving as a patch, canbe simultaneously applied. The outer mem-brane is sutured such that its edges are tuckedunder the peritomised conjunctiva. Conjuncti-va derived epithelium then grows on the outermembrane and is prevented from admixingwith limbus-derived epithelium that is spread-ing onto the corneal surface. This technique wasdeveloped by the author and is regularly em-ployed [26] (Fig. 3.16). It avoids the need forSSCE postoperatively. The outer membrane fallsoff or can be removed in 10–14 days. For furtherdetails on amniotic membrane, see Chap. 2 onamniotic membrane transplantation.

3.6.4.2Ex Vivo Expansion of Limbal Cells

Limbal ‘stem cell’ transplantation can also becarried out by ex vivo expansion of limbal ep-ithelial cells, either directly or on a substrate offibrin, collagen or amniotic membrane [26, 62,74, 75, 87].

3.6.4.3Corneal Grafts

Lamellar or full thickness corneal grafts can becombined with auto- or allo-limbal transplanta-tion. This may be necessary when the corneadamage is severe and when it is considered thatthe host corneal bed will not support a healthyepithelium despite use of an amniotic mem-brane. In general terms, a definitive cornealgraft for visual purposes should be deferred un-til ocular surface epithelial integrity has beenrestored by limbal transplantation. However, ina recent study using limbal tissue remainingafter keratoplasty from organ cultured cor-neoscleral discs, we have shown that such tissue(where the death to enucleation time and thetime lapse between enucleation to placement inorgan culture is short and where the donor isrelatively young) retains good proliferative ca-pacity for up 30 days in storage (V. Shanmu-

ganathan, submitted to Br J Ophthalmol 2005).This offers the opportunity to carry HLA typingand matching and also allows for depletion ofantigen presenting Langerhans cells. It shouldtherefore be possible to use organ culture pre-served corneoscleral discs for simultaneousallo-limbal transplant and keratoplasty withreduced risk of immune mediated rejection.

3.6.5Postoperative Treatment

Topical preservative-free antibiotic drops suchas chloramphenicol 0.5% are used four times aday for the first month. Topical preservative-free steroid drops such as prednisolone acetate1% are used four times a day for the first8–12 weeks, and slowly tapered during the ensu-ing weeks. A low dose of topical corticosteroids(one drop per day) is maintained unless eleva-tion of intraocular pressure occurs. Autologousserum eyedrops (20%) [52, 92, 93] are givenhourly until the epithelialsation is complete,usually in 7–10 days. Preservative free artificialtears are then instituted. It is important to close-ly monitor the re-epithelialisation process untilcompleted. Any attempt by conjunctiva derivedcells to encroach onto the corneal surfaceshould be thwarted by SSCE, until the periphery(limbus) of the host cornea is re-epithelialisedby limbus derived cells.

All patients undergoing allo-limbal trans-plantation also need systemic immunosuppres-sion. Besides steroids, azathioprine, cyclosporinA, rapamycin, mycophenolate mofetil andtacrolimus (FK506, Prograf) have been used [11,72, 78, 97]. Theoretically, immunosuppressionshould be continued almost indefinitely. Theauthor has used cyclosporin A and of late FK506up to 18 months postoperatively [78]. Attemptsto reduce or stop the drug have resulted in lim-bal and/or corneal graft rejection episodes. For-tunately, the dose required to prevent or controlrejection episodes is very low (2–8 mg/day,maintaining a blood trough level of 1–12 mg/l).Serious side effects, though they occur, are notvery common, but require constant monitoringof patients and measures of kidney and liverfunctions. It is good practice to involve a clinical


immunologist or a physician versed in im-munosuppressive therapies in the managementand monitoring of these patients.

Successful stem cell grafting is a team effortinvolving the corneal and oculoplastic surgeonsin close cooperation with the clinical immunol-ogist. Often multiple surgical procedures are re-quired and visual outcome, though useful fromthe patients’ viewpoint, may be limited. Thethreat of limbal graft rejection is real and con-siderable. The all-important question of the du-ration of systemic immunosuppression remainsto be answered. Not all long-term DNA trackingstudies on recipient eyes have been able to showpresence of donor derived cells even in the pres-ence of a healthy corneal surface [38–40, 71, 76].This would suggest that restoration of a normalsurface and ‘microenvironment’ may allow hoststem cells, either surviving limbal stem cells orbone marrow derived stem cells, to repopulatethe surface. In this situation long-term im-munosuppression would not be a necessity. Onthe other hand, long-term follow-up studieshave also demonstrated that the outcome ofallo-limbal transplant is not as good as that ofauto-limbal transplant. This may reflect chronic‘immune mediated’ damage and attrition of thetransplanted limbal stem cells or the relative‘freshness’ of auto grafts, conferring upon thema survival advantage.

References

1. Alison MR, Poulsom R, Forbes S, Wright NA(2002) An introduction to stem cells. J Pathol197:419–423

2. Azuara-Blanco A, Pillai CT, Dua HS (1999) Amni-otic membrane transplantation for ocular surfacereconstruction. Br J Ophthalmol 83:399–402

3. Bishop AE, Buttery LD, Polak JM (2002) Embry-onic stem cells. J Pathol 197:424–429

4. Buck RC (1979) Cell migration in repair of mousecorneal epithelium. Invest Ophthalmol Vis Sci18:767–784

5. Buck RC (1985) Measurement of centripetal mi-gration of normal corneal epithelial cells in themouse. Invest Ophthalmol Vis Sci 26:1296–1299

6. Chiou AG-Y, Florakis GJ, Kazim M (1998) Man-agement of conjunctival cicatrizing diseases andsevere ocular surface dysfunction. Surv Ophthal-mol 43:19–46

7. Cooper D, Schermer A, Sun TT (1985) Classifica-tion of human epithelia and their neoplasms us-ing monoclonal antibodies to keratins: strategies,applications, and limitations. Lab Invest 52:243–256

8. Cotsarelis G, Cheng SZ, Dong G, Sun TT, LavkerRM (1989) Existence of slow-cycling limbal ep-ithelial basal cells that can be preferentially stim-ulated to proliferate: implications on epithelialstem cells. Cell 57:201–209

9. Davanger M, Evensen A (1971) Role of the peri-corneal papillary structure in renewal of cornealepithelium. Nature 229:560–561

10. Daya SM, Holland EJ, Mannis M (2001) Living re-lated conjunctival limbal allograft. In: Holland EJ,Mannis MJ (eds) Ocular surface disease, medicaland surgical management, Chap 17. Springer, NewYork, pp 201–207

11. Djalilian AR, Nussenblatt RB, Holland EJ (2001)Immunosuppressive therapy in ocular surfacetransplantation. In: Holland EJ, Mannis MJ (eds)Ocular surface disease, medical and surgicalmanagement, Chap 22. Springer, New York, pp243–252

12. Dong Y, Roos M, Gruijters T, Donaldson P, Bulli-vant S, Beyer E (1994) Differential expression oftwo gap junction proteins in corneal epithelium.Eur J Cell Biol 64:95–100

13. Dua HS (1995) Stem cells of the ocular surface:Scientific principles and clinical applications. Br JOphthalmol 79:968–969

14. Dua HS (1998) The conjunctiva in corneal epithe-lial wound healing. Br J Ophthalmol 82:1407–1411

15. Dua HS (2001) Sequential sector conjunctival ep-itheliectomy. In: Holland EJ, Mannis MJ (eds) Oc-ular surface disease, medical and surgical man-agement, Chap 14. Springer, New York, pp 168–174

16. Dua HS, Azuara-Blanco A (1999) Allo-limbaltransplantation in patients with limbal stem-celldeficiency. Br J Ophthalmol 83:414–419

17. Dua HS, Azuara-Blanco A (1999) Amniotic mem-brane transplantation. Br J Ophthalmol 83:748–752

18. Dua HS, Azuara-Blanco A (2000) Limbal stemcells of the corneal epithelium. Surv Ophthalmol44:415–425

19. Dua HS,Azuara-Blanco A (2000) Autologous lim-bal transplantation in unilateral stem cell defi-ciency. Br J Ophthalmol 84:273–278

20. Dua HS, Forrester JV (1987) Clinical patterns ofcorneal epithelial wound healing. Am J Ophthal-mol 104:481–489

21. Dua HS, Forrester JV (1990) The corneosclerallimbus in human corneal epithelial wound heal-ing. Am J Ophthalmol 110:646–656

References 53

22. Dua HS, Gomes JA, Singh A (1994) Corneal ep-ithelial wound healing. Br J Ophthalmol 78:401–408

23. Dua HS, Gomes JAP, Jindal V, Appa SN, Schwart-ing R, Eagle RC Jr, Donoso LA, Laibson PR (1994)Mucosa specific lymphocytes in the human con-junctiva, corneoscleral limbus and lacrimalgland. Curr Eye Res 13:87–93

24. Dua HS, Saini JS, Azuara-Blanco A, Gupta P(2000) Limbal stem cell deficiency: Concept, aeti-ology, clinical presentation, diagnosis and man-agement. Ind J Ophthalmol 48:83–92

25. Dua HS, Joseph A, Shanmuganathan VA, Jones RE(2003) Stem cell differentiation and effects ofdeficiency. Eye 17:877–885

26. Dua HS, Gomes JAP, King AJ, Maharajan S (2004)The amniotic membrane in ophthalmology. SurvOphthalmol 49:51–77

27. Dua HS, Shanmuganathan VA, Powell-RichardsAO, Tighe PJ, Joseph A (2005) Limbal epithelialcrypts – A novel anatomical structure and a puta-tive limbal stem cell niche. Br J Ophthalmol (inpress)

28. Ebato B, Friend J, Thoft RA (1987) Comparison ofcentral and peripheral human corneal epitheliumin tissue culture. Invest Ophthalmol Vis Sci28:1450–1456

29. Ebato B, Friend J, Thoft RA (1988) Comparison oflimbal and peripheral human corneal epitheliumin tissue culture. Invest Ophthalmol Vis Sci29:1533–1537

30. Eggli P, Boulton M, Marshall J (1989) Growthcharacteristics of central and peripheral bovinecorneal epithelial cells in vitro. Graefes Arch ClinExp Ophthalmol 227:263–270

31. Friedenwald JSB, Buschke W (1944) Some factorsconcerned in the mitotic and wound-healing ac-tivities of the corneal epithelium. Trans Am Oph-thalmol Soc 42:371–383

32. Fuchs E, Segre JA (2000) Stem cells: a new leaseon life. Cell 100:143–155

33. Fujishima H, Shimazaki J, Tsubota K (1996) Tem-porary corneal stem cell dysfunction after radia-tion therapy. Br J Ophthalmol 80:911–914

34. Gipson IK, Grill SM, Spurr SJ, Brennan SJ (1983)Hemidesmosome formation in vitro. J Cell Biol97:849–857

35. Gipson IK, Kiorpes TC (1982) Epithelial sheetmovement: protein and glycoprotein synthesis.Dev Biol 92:259–262

36. Goodell MA, Rosenzweig M, Kim H, Marks DF, DeMaria M, Paradis G, Grupp SA, Ssieff CA, Mulli-gan RC, Johnson RP (1996) Isolation and func-tional properties of murine hematopoietic stemcells that are replicating in vivo. J Exp Med183:1797–806

37. Hanna C, O’Brien JE (1960) Cell production andmigration in the epithelial layer of the cornea.Arch Ophthalmol 64:536–539

38. Henderson TR, Coster DJ, Williams KA (2001)The long term outcome of limbal allografts: thesearch for surviving cells. Br J Ophthalmol85:604–609

39. Henderson TR, Findlay I, Matthews PL, Noble BA(2001) Identifying the origin of single cornealcells by DNA fingerprinting: part I – applicationto limbal allografting. Cornea 20:400–403

40. Henderson TR, Findlay I, Matthews PL, Noble BA(2001) Identifying the origin of single cornealcells by DNA fingerprinting: part II – applicationto limbal allografting. Cornea 20:404–407

41. Holland EJ (1996) Epithelial transplantation forthe management of severe ocular surface disease.Trans Am Ophthalmol Soc 94:677–743

42. Holland EJ, Schwartz GS (1996) The evolution ofepithelial transplantation for severe ocular sur-face disease and a proposed classification system.Cornea 15:549–556

43. Huang AJ, Tseng SC (1991) Corneal epithelialwound healing in the absence of limbal epitheli-um. Invest Ophthalmol Vis Sci 32:96–105

44. Janes SM, Lowell S, Hutter C (2002) Epidermalstem cells. J Pathol 197:479–491

45. Jones PH, Watt FM (1993) Separation of humanepidermal stem cells from transit amplifying cellson the basis of differences in integrin functionand expression. Cell 73:713–724

46. Joseph A, Powell Richards AOR, Shanmu-ganathan VA, Dua HS (2004) Epithelial cell char-acteristics of cultured human limbal explants. BrJ Ophthalmol 88:393–398

47. Joseph A, Hossain P, Jham S, Jones RE, Tighe P,McIntosh RS, Dua HS (2003) Expression of CD34and L-selectin on human corneal keratocytes.Invest Ophthalmol Vis Sci 44:4689–4692

48. Kenyon KR, Rapoza PA (1995) Limbal allografttransplantation for ocular surface disorders.Ophthalmology 102 (Suppl):101–102

49. Kenyon KR, Tseng SC (1989) Limbal autografttransplantation for ocular surface disorders.Ophthalmology 96:709–22; discussion 722–723

50. Kim JC, Tseng SCG (1995) Transplantation of pre-served human amniotic membrane for surfacereconstruction in severely damaged rabbitcorneas. Cornea 15:473–484

51. Kinoshita S, Friend J, Thoft RA (1981) Sex chro-matin of donor corneal epithelium in rabbits. In-vest Ophthalmol Vis Sci 21:434–441

52. Lagnado R, King AJ, Donald F, Dua HS (2003) Aprotocol for low contamination risk of autolo-gous serum drops in the management of ocularsurface disorders. Br J Ophthalmol 88:464–465


53. Lauweryns B, van den Oord JJ, Volpes R, Foets B,Missotten (1991) Distribution of very late activa-tion integrins in the human cornea. An immuno-histochemical study using monoclonal antibod-ies. Invest Ophthalmol Vis Sci 32:2079–2085

54. Lauweryns B, van den Oord JJ, Missotten L (1993)The transitional zone between limbus and pe-ripheral cornea. An immunohistochemical study.Invest Ophthalmol Vis Sci 34:1991–1999

55. Lauweryns B, van den Oord JJ, De Vos R, Missot-ten L (1993) A new epithelial cell type in the hu-man cornea. Invest Ophthalmol Vis Sci 34:1983–1990

56. Lavker RM, Dong G, Cheng SZ, Kudoh K, Cot-sarelis G, Sun TT (1991) Relative proliferativerates of limbal and corneal epithelia. Implicationsof corneal epithelial migration,circadian rhythm,and suprabasally located DNA-synthesizing ker-atinocytes. Invest Ophthalmol Vis Sci 32:1864–1875

57. Leblond CP (1981) The life history of cells in re-newing systems. Am J Anat 160:114–158

58. Matic M, Petrov IN, Chen S, Wang C, Dimitrije-vich SD, Wolosin JM (1997) Stem cells of thecorneal epithelium lack connexins and metabo-lite transfer capacity. Differentiation 61:251–260

59. Matsuda M, Ubels JL, Edelhauser HF (1985) Alarger corneal epithelial wound closes at a fasterrate. Invest Ophthalmol Vis Sci 26:897–900

60. Mills JC, Gordon JI (2001) The intestinal stem cellniche: there grows the neighborhood. Proc NatlAcad Sci USA 98:12334–12336

61. Morrison SJ, Shah NM,Anderson DJ (1997) Regu-latory mechanisms in stem cell biology. Cell88:287–298

62. Pellegrini G, Traverso CE, Franzi AT (1997) Long-term restoration of damaged corneal surfaceswith autologous cultivated corneal epithelium.Lancet 349:990–993

63. Pellegrini G, Dellambra E, Golisano O, MartinelliE, Fantozzi I, Bondanza S (2001) p63 identifieskeratinocyte stem cells. Proc Natl Acad Sci USA98:3156–3161

64. Pfister RR (1994) Corneal stem cell disease: con-cepts, categorization, and treatment by auto- andhomotransplantation of limbal stem cells. CLAO J20:64–72

65. Potten SC,Loeffler M (1987) Epidermal cell prolif-eration. I. Changes with time in the proportion ofisolated, paired and clustered labeled cells insheets of murine epidermis. Virchows Arch 53:286–300

66. Potten SC, Loeffler M (1990) Stem cells: attrib-utes,cycles, spirals,pitfalls and uncertainties.Les-son for and from the crypt. Development 110:1001–1020

67. Potten CS, Morris RJ (1988) Epithelial stem cellsin vivo. J Cell Sci 10 (Suppl):45–62

68. Poulsom R, Alison MR, Forbes SJ, Wright NA(2002) Adult stem cell plasticity. J Pathol 197:441–456

69. Puangsricharern V, Tseng SC (1995) Cytologicevidence of corneal diseases with limbal stem celldeficiency. Ophthalmology 102:1476–1485

70. Rao SK, Rajgopal R, Sitalakshmi G, PadmanabhanP (1999) Limbal allografting from related livedonors for corneal surface reconstruction. Oph-thalmology 106:822–828

71. Reinhard T, Spelsberg H, Henke L, Kontopoulos T,Enczmann J, Wernet P, Berschick P, SundmacherR, Bohringer D (2004) Long-term results of allo-geneic penetrating limbo-keratoplasty in totallimbal stem cell deficiency. Ophthalmology 111:775–782

72. Reis A, Reinhard T, Voiculescu A et al. (1999)Mycophenolate mofetil versus cyclosporin A inhigh risk penetrating keratoplasty. Br J Ophthal-mol 83:1268–1271

73. Schermer A, Galvin S, Sun TT (1986) Differentia-tion-related expression of a major 64K cornealkeratin in vivo and in culture suggests limbal lo-cation of corneal epithelial stem cells. J Cell Biol103:49–62

74. Schwab IR (1999) Cultured corneal epithelia forocular surface disease. Trans Am Ophthalmol Soc97:891–986

75. Schwab IR, Reyes M, Isseroff RR (2000) Success-ful transplantation of bioengineered tissue re-placements in patients with ocular surface dis-ease. Cornea 19:421–426

76. Shimazaki J, Kaido M, Shinozaki N, Shimmura S,Munkhbat B, Hagihara M, Tsuji K, Tsubota K(1999) Evidence of long term survival of donorderived cells after limbal allograft transplanta-tion. Invest Ophthalmol Vis Sci 40:1664–1668

77. Shimazaki J,Aiba M, Goto E, Kato N, Shimmura S,Tsubota K (2002) Transplantation of human lim-bal epithelium cultivated on amniotic membranefor the treatment of severe ocular surface disor-ders. Ophthalmology 109:1285–90

78. Sloper ML, Powell RJ, Dua HS (2001) Tacrolimus(FK506) in the management of high-risk cornealand limbal grafts. Ophthalmology 108:1838–1844

79. Spradling A, Drummond-Barbosas D, Kai T(2001) Stem cells find their niche. Nature 414:98–104

80. Srinivasan BD, Eakins KE (1979) The reepithelial-ization of rabbit cornea following single and mul-tiple denudation. Exp Eye Res 29:595–600

References 55

81. Sundmacher R, Reinhard T (1996) Central corne-olimbal transplantation under systemic cyclo-sporin A cover for severe limbal stem cell insuffi-ciency. Graefes Arch Clin Exp Ophthalmol 234(Suppl):122–125

82. Sundmacher R, Reinhard T (1998) Homologouslamellar central limbokeratoplasty in severe lim-bal stem cell deficiency. Klin Monatsbl Augen-heilkd 213:254–255

83. Tan DTH, Ficker LA, Buckley RJ (1996) Limbaltransplantation. Ophthalmology 103:29–36

84. Thoft RA, Friend J (1983) The X,Y, Z hypothesis ofcorneal epithelial maintenance. Invest Ophthal-mol Vis Sci 24:1442–1443

85. Thoft RA,Wiley LA, Sundarraj N (1989) The mul-tipotential cells of the limbus. Eye 3:109–113

86. Townsend WM (1991) The limbal palisades ofVogt. Trans Am Ophthalmol Soc 89:721–756

87. Tsai RJ, Li LM, Chen JK (2000) Reconstruction ofdamaged corneas by transplantation of autolo-gous limbal epithelial cells. N Eng J Med 3431:86–93

88. Tseng SC (1989) Concept and application of lim-bal stem cells. Eye 3:141–157

89. Tseng SCG, Tsubota K (1997) Important conceptsfor treating ocular surface and tear disorders.AmJ Ophthalmol 124:825–835

90. Tseng SCG, Prabhasawat P, Barton K, Gray T,Meller D (1998) Amniotic membrane transplan-tation with or without limbal allografts forcorneal surface reconstruction in patients withlimbal stem cell deficiency. Arch Ophthalmol116:431–441

91. Tsubota K, Toda I, Saito H, Shinozaki N, Shimaza-ki J (1995) Reconstruction of the corneal epitheli-um by limbal allograft transplantation for severeocular surface disorders. Ophthalmology 102:1486–1496

92. Tsubota K, Satake Y, Ohyama M, Toda I, Takano Y,Ono M, Shinozaki N, Shimazaki J (1996) Surgicalreconstruction of the ocular surface in advancedocular cicatricial pemphigoid and Stevens-John-son syndrome. Am J Ophthalmol 122:38–52

93. Tsubota K, Satake Y, Shimazaki J (1996) Treat-ment of severe dry eye. Lancet 348:123

94. Watanabe K, Nishida K, Yamato M, Umemato T,Sumide T, Yamamoto K, Maeda N, Watanabe H,Okano T, Tano Y (2004) Human limbal epitheli-um contains side population cells expressing theATP-binding cassette transporter ABCG2. FEBSLett 565:6–10

95. Watt FM, Hogan BLM (2000) Out of Eden: stemcells and their niches. Science 287:1427–1430

96. Wolosin JM, Xiong X, Schutte M, Stegman Z,Tieng A (2000) Stem cells and differentiationstages in the limbo-corneal epithelium. ProgRetin Eye Res 19:223–255

97. Xu KP, Wu Y, Zhou J, Zhang X (1999) Survival oflimbal stem cell allografts after administrationof cyclosporin A. Cornea 18:159–165

98. Zhao X, Das AV, Thoreson WB, James J, WattnemTE, Rodriguez-Sierra J (2002) Adult corneal lim-bal epithelium: a model for studying neural po-tential of non-neural stem cells/progenitors. DevBiol 250:317–331

99. Zhou S, Schuetz JD, Bunting KD, Colapietro AM,Sampath J, Morris JJ, Lagutina I, Grosveld GC,Osawa M, Nakauchi H, Sorrentino BP (2001) TheABC transporter Bcrp1/ABCG2 is expressed in awide variety of stem cells and is a molecular de-terminant of the side-population phenotype. NatMed 7:1028–1034

100. Zieske JD (1994) Perpetuation of stem cells in theeye. Eye 8:163–169

101. Zieske JD, Bukusoglu G, Yankauckas MA (1992)Characterization of a potential marker of cornealepithelial stem cells. Invest Ophthalmol Vis Sci33:143–152

102. Zieske JD, Bukusoglu G,Yankauckas MA,WassonME, Keutmann HT (1992) Alpha-enolase is re-stricted to basal cells of stratified squamous ep-ithelium. Dev Biol 151:18–26

103. Zieske JD,Wasson M (1993) Regional variation indistribution of EGF receptor in developing andadult corneal epithelium. J Cell Sci 106:145–152


4.1Introduction

The stem cells of the corneal epithelium seem tobe located in the limbal basal layer and are theultimate source of constant corneal epithelialrenewal. A new strategy of treating limbal stemcell deficiency is to transplant a bioengineeredgraft by expanding limbal epithelial stem cellsex vivo on amniotic membrane.

Ex vivo expansion of limbal epithelial stemcells has been developed to circumvent poten-tial complications both for the recipient and thedonor arising from conjunctival limbal auto-graft transplantation. This technique expandslimbal epithelial progenitor cells from a smallbiopsy using a 3T3 fibroblast feeder layer [13] oramniotic membrane. Using the amniotic mem-brane to constitute such a composite graft, suc-cessful reconstruction of the normal cornealsurface has been achieved in several humanstudies with partial or total limbal stem celldeficiencies [8, 9].

As previously described, amniotic mem-brane, the innermost layer of the fetal or placen-tal membrane, consists of an epithelial mono-layer, a thick basement membrane, and anavascular stroma (Fig. 4.1). With appropriateprocurement and preservation, amniotic mem-brane can be used as a biological substrate with-out viable and proliferative active cells [8]. It isthus non-immunogenic and therefore does notrequire immunosuppression when used fortransplantation. The amniotic membrane stro-ma also contains quantities of growth factors,various antiangiogenic and anti-inflammatoryproteins and natural inhibitors to various pro-teases.

Limbal Stem Cell Culture

José L. Güell, Marta Torrabadella, Marta Calatayud, Oscar Gris,Felicidad Manero, Javier Gaytan

4

|

∑ The stem cells of the corneal epitheliumseem to be located in the limbal basal layerand are the ultimate source of constantepithelial renewal

∑ Ex vivo expansion of limbal epithelial stemcells has been developed mainly to circum-vent potential complications, for both therecipient and the living donor, arising fromstandard conjunctival limbal transplanta-tion

∑ With the appropriate procurement andpreservation, human amniotic membranecan be used as a biological substrate with-out viable and proliferative active cells butwith the advantages of basement mem-brane and as a source of beneficial bio-logical factors

∑ To examine the epithelial phenotype, im-munostaining techniques are used with apanel of monoclonal antibodies to mucinsand keratins

∑ The preparation of human amniotic mem-brane and the culture of explanted tissuemust be performed by specially trainedpersonnel in a stem-cell management laboratory

∑ Preliminary clinical experience is highly en-couraging, not only regarding final clinicaloutcome but also because of the flexibilityof the technique, the increase in possiblecases to be treated and the reduced needfor systemic immunosuppression

∑ Ocular surface reconstruction is becominga much more staged than artistic approach,and we are of the opinion that its final clini-cal results will improve strongly in the nearfuture especially with the help of the newcell bioengineering technique

Core Messages

Recently, Grueterich et al. identified the cul-ture environment that will favor the mainte-nance of the stem cell containing the limbalepithelial phenotype [9]. This is achieved byculturing limbal explants on an intact humanamniotic membrane, which retains the devital-ized amniotic epithelium, without the use of a3T3 fibroblast feeder layer. The expanded ep-ithelium on intact amniotic membrane adopts alimbal epithelial phenotype whereas that ondenuded amniotic membrane reveals a cornealepithelial phenotype.

4.2Epithelial Phenotype

To examine the epithelial phenotype, immunos-taining techniques are used with a panel ofmonoclonal antibodies to mucins and keratins.In normal ocular surface epithelia, AE5 anti-

body, which recognizes K3 keratin, stains thesuprabasal limbal epithelium and the full thick-ness of the central corneal epithelium, but notthe conjunctival epithelium. AE5 antibodystains suprabasal human limbal epithelial cells(HLEC) cultured on amniotic membrane for13–21 days. Immunostaining for K12 keratin by AK2 antibody is also positive for limbalsuprabasal epithelial cells and for the full thick-ness of the corneal epithelium, but negative forthe conjunctival epithelium in vivo. HLEC onamniotic membrane are negative for AK2. K14keratin is expressed in the basal and suprabasalcell layers of the conjunctival limbal and pe-ripheral corneal epithelium, but is found pre-dominantly in the basal epithelial cells of thecentral corneal epithelium. HLEC cultured onamniotic membrane showed full thicknessstaining to K14 keratin after 13–21 days of cul-turing. MUC5AC (Mucina 5AC) recognizes con-junctival goblet cell secreting mucins and stainsconjunctival goblet cells in vivo. MUC5AC donot stain any cells cultured on amniotic mem-brane. Collectively, these results indicate thatthe resultant phenotype of HLEC grown on am-niotic membrane retains a limbal origin, is pre-dominantly basal epithelial cells, and remainsundifferentiated (Table 4.1) [9].

4.3Preparation of Human Amniotic Membrane

Amniotic membrane tissue can be obtained,processed, and preserved frozen as reported byLee and Tseng [9] at the Eye Bank. The amniot-ic membrane measuring 5¥5 cm, after thawingand washing, is tightened on a 3-cm cultureplate with the basement membrane side up.

58 Chapter 4 Limbal Stem Cell Culture

Fig. 4.1. Human amniotic membrane. Cross-sectionreveals a cuboidal epithelial monolayer (1), a thickbasement membrane (2) and an avascular stroma (3)

Table 4.1. Immunostaining expression of keratins K3, K12 and K14, and MUC5AC of corneal epithelium, con-junctival epithelium, limbal epithelium and HLEC grown on amniotic membrane cultures

Immunostaining Corneal Conjunctival Limbal epithelium HLECepithelium epithelium

K3 keratin Positive Negative Positive suprabasal Positive suprabasal

K12 keratin Positive Negative Positive suprabasal Negative

K14 keratin Positive basal Positive Positive Positive

MUC5AC Negative Positive Negative Negative

4.4Culture of Explanted Tissue

A piece of limbal tissue measuring 1¥2 mm con-taining epithelial cells and part of the cornealstroma is obtained for ex vivo culture. The tis-sue source is the contralateral eye in autologoustransplantation or a living donor in relatedallotransplantation. Cadaveric donor is also asource of limbal tissue for conditions in whichboth eyes are affected and no living donor isavailable.

The obtained tissue is placed with Ham’s F12medium containing 50 mg/ml gentamicin and1.25 mg/ml amphotericin B until it is processed.Limbal tissue is exposed for 5 min to Dispase II(1.2 U/ml in Mg2+ and Ca2+ free Hank’s balancedsalt solution, HBSS) at 37 °C under humidified5% CO2. The explants are then cultured inDMEM medium, which is a 1:1 mixture ofDMEM and Ham’s F12 medium containing5 ng/ml epithelial growth factor (EGF), 5 mg/mlinsulin, 5 mg/ml transferrin, 5 ng/ml sodiumselenite, 0.5 mg/ml hydrocortisone, 30 ng/mlcholera toxin A,0.5% dimethylsulfoxide (DMSO),50 mg/ml gentamicin, 1.25 mg/ml amphotericinB and 5% autologous serum, at 37 °C under 5%CO2 and 95% humidity. The medium is renewedevery 2–3 days [2]. For allogeneic related trans-plantation, donor serum is used and for allo-geneic non-related transplantation AB testedblood bank serum is employed. The limbal ep-ithelial cell explants are plated onto the base-ment-membrane side of the amniotic mem-brane, placed in the center. The extent of eachoutgrowth is monitored with a phase contrastmicroscope. During the expansion phase thelimbal epithelial outgrowth exhibits a compactand uniform cell layer (Fig. 4.2).

The culture is maintained for 2–3 weeks, bywhich time the epithelial cells have grown andspread to form a cell layer that covers an area2–3 cm in diameter. Every week bacteriologicaltesting is performed to assess microorganismcontamination. The mycoplasma content testand Gram’s test are performed 24 h beforetransplantation.

4.5Tissue Procurement

The safety of biological medicinal products re-lies on rigorous control of each of their com-ponents. Human tissue should be handled according to European Commission Directive2003/63/EC [8]. Corneoscleral tissue from hu-man donor eyes is obtained after proper in-formed consent in the case of a living donor orfrom an authorized Eye Bank for cadavericdonors. Human amniotic membrane should beobtained after elective cesarean delivery whenblood-borne microorganisms such as humanimmunodeficiency virus, hepatitis virus type Band C, and syphilis have been excluded by sero-

4.5 Tissue Procurement 59

Fig. 4.2 A, B. Ex vivo expansion of limbal stem cellsfrom limbal biopsy. A Culture plate with the tightenedamniotic membrane and a limbal biopsy placed in thecenter (arrows). B Phase-contrast microscopy of theexpanded cells reveals a monolayer of epithelial cellsof small and uniform size, ¥400

A

B

logic tests. Hepatitis virus type C and humanimmunodeficiency virus should additionally beexcluded by means of PCR.


∑ Immature corneal cells are located in theepithelial basal layer of the limbus

∑ Ex vivo expansion of limbal epithelial cellsis a technique that avoids potential compli-cations to the donor and the recipient eye

∑ Immunohistochemical features allow us todistinguish immature from mature cells onthe cultured cells

∑ Human amniotic membrane provides the limbal epithelium with an adaptedmicroenvironment and a solid basal layer

∑ Amniotic membrane epithelium favorslimbal stem cell growth

∑ Biological medicinal products rely onrigorous control of each of their compo-nents and should be handled according tothe specific laws of each country

4.6Preliminary Clinical Experience

In standard ocular surface reconstruction, am-niotic membrane transplantation in the corneaand conjunctiva provides a basement mem-brane, especially when the ocular surface ishighly irregular, and a source of biological fac-tors to enhance and improve reepithelialization[3–7]. Once the amniotic membrane is sutured,we may proceed with our standard limbal trans-plantation over it.

Because sclerocorneal limbal tissue is highlyvascularized with a high antigenic weight, weneed to prescribe long-term systemic immuno-suppression in most limbal transplantation cas-es (cadaver or relative donor).

Limbal stem cell culture is a new techniquewith obvious advantages over the other tech-niques used to restore ocular surface. The aim isto cause as little damage to the ocular surface aspossible.Ex vivo expansion of corneal stem cellsis a technique described previously [7, 8, 11, 13],and has been developed by several groups. Itbasically consists of taking a small piece of thedonor eye limbus, expanding the cells in the lab-

oratory over a piece of amniotic membrane andusing them in a pathologic eye to restore thecorneal epithelium. We prefer amniotic mem-brane because its basal membrane is very simi-lar to corneal epithelial basal membrane [7, 11],and the epithelium of the amniotic membranecontains several growth factors that promotecellular proliferation [7]. The main advantage isthat in nearly all cases we avoid immunosup-pression, and cause minimal injury to the donoreye, so there is no risk of iatrogenous limbal de-ficiency in that eye.

Limbal deficiency must have been confirmedpreviously by clinical examination, impressioncytology and, if possible, immunofluorescenceon the affected eye.

The process starts with, after informed con-sent, the extraction of a limbal biopsy from thedonor eye (the healthy eye of the same patientor a relative or cadaver donor eye). It must bedone in the operating room,under sterile condi-tions and with great care being taken to avoidthe conjunctival tissue, because we could ex-pand other cells than corneal epithelial stemcells. The size of the biopsy must be as small aspossible, but large enough to ensure cellulargrowth in culture (2 mm2), with a depth of100 mm, without blood vessels or conjunctivaltissue. Once we have the biopsy, it must be takento the laboratory as soon as possible, where it istreated as mentioned above, and then we needto wait for 2 or 3 weeks, depending on the cellu-lar growth observed, before we can implant theamniotic membrane with the expanded cells onthe eye. Every process which involves any ma-nipulation of the tissue must be done understerile conditions.

4.6.1Principles for Taking the Biopsy

1. Avoid or eliminate the conjunctiva. Conjunc-tival epithelial cells grow easily in cultureand interfere with corneal epithelium prolif-eration.

2. Perform the biopsy under sterile conditions.3. Take a biopsy as small as 1–2 mm2 and

100 mm in depth.


4. Take a biopsy from the superior limbus ifpossible, to ensure immature cells are includ-ed on it (the superior limbus contains morestem cells) [13]

We usually implant the tissue after a lamellarkeratectomy is performed to eliminate any re-mains of conjunctival tissue over the cornea, su-turing the graft, with the small biopsy included,over the corneal surface, then placing a thera-peutic contact lens over the graft to preserve itand avoid blinking-trauma during the first2 weeks. Postoperative treatment consists oftopical steroids and antibiotics, autologousserum drops to promote epithelial growth, andimmunosuppressive agents if the biopsy is per-formed on a relative or unknown donor.

In our series of seven patients transplanted,the only complication detected in the immedi-ate postoperative period was corneal meltingwith perforation, which was solved with pene-trating keratoplasty (PKP).

The most important purpose of this tech-nique is to expand immature cells from the pro-liferative compartment of the limbus and keepthem active for as long as possible, which en-sures a healthy ocular surface, and allows us to perform other procedures to restore visualacuity such as PKP or lamellar keratoplasty

4.6.2Advantages of Limbal Stem Cell Culture

1. In many cases our patient has enoughhealthy limbal area in at least one of the eyesfor a small biopsy to be taken but not for alarge (90°) standard biopsy to be done forstandard transplantation.

2. Many close relatives (HLA matched if possi-ble) will be amenable to a small biopsy but notto a large one with the known associated risks.In both (1) and (2), we might eliminate orstrongly reduce systemic immunosuppres-sion

3. In those cases where the donor tissue is froma cadaver, the main advantage of the cultureis the density of viable cells at the time oftransplantation compared with the freshoriginal piece of tissue.

On the other hand, one limitation is preopera-tive cell type determination. To assess limbalstem cell transplant success we do not manipu-late the culture. Rather we implant the centralarea, where the graft shows the best growth, andwe verify the particular biological and im-munohistochemical characteristics of the cul-tured cells with the remaining piece of tissueafter the surgery. In all cases limbal stem cellimmunophenotyping is demonstrated.


∑ Amniotic membrane transplantation on thecornea and conjunctiva contributes, instandard ocular surface reconstruction,as a basement membrane and as a source of biological factors

∑ In most standard limbal transplantationcases, we all need to prescribe long-termsystemic immunosuppression

∑ The use of amniotic membrane as our “cell transporter” was decided upon because of our wide and long-term experi-ence of using it in our standard ocular surface reconstructive techniques

∑ The most important purpose of our tech-nique is to expand immature cells from theproliferative compartment of the limbusand keep them active for as long as possible

∑ We must keep in mind the main conceptualadvantages of limbal stem cell culture inclinical practice

4.7Case Report

A 31-year-old Caucasian male with a story ofbilateral caustication came to us in July 2002. Atthe first exploration of the eyes in early Septem-ber 2002, the right eye showed visual acuity(VA), light perception, diffuse corneal conjunc-tivalization with central perforation and exten-sive superior symblepharon; in the left eye theVA was count fingers 10 cm, and diffuse cornealconjunctivalization with extensive inferiorsymblepharon was seen. Our case notes thenread as follows:12.09.02 Emergency PKP plus amniotic mem-

brane as a patch OD

4.7 Case Report 61

10.12.02 Inferior symblepharon surgery OD25.02.03 Superior symblepharon surgery OD26.08.03 Inferior symblepharon surgery OS08.03.04 Impression cytology OS (Fig. 4.5):(Fig. 4.3) only 2 h (10–12 o’clock) with healthy

limbal tissue; healthy limbal areabiopsy for culture on amniotic mem-brane

25.03.04 Lamellar corneal scar dissection andcultured cells on amniotic membranesutured on the cornea – conjunctivalsurface OS

29.04.04 PKP + standard amniotic membrane (Fig. 4.6) transplantation + lateral temporary

tarsorrhaphy19.10.04 Sectorial conjunctival epitheliectomy(Fig. 4.7) as described by Dua [7] (area from

6 o’clock to 8 o’clock) and selectivesuture removal


Fig. 4.3. Specimen from a limbal deficiency eye. Pre-operative image, Papanicolaou’s stain

Fig. 4.4 A, B. Slit-lamp appearance of both eyes inMarch 2004. A left eye; B white cornea with vessels,right eye

A

Fig. 4.5 A, B. Operating room impression cytologyof the left eye. Note the presence in all the specimens ofgoblet cells characteristic of conjunctival epithelium

A

B

B

16.11.04 Best corrected visual acuity (BCVA)OS 20/40 posterior subcapsularcataract and healthy ocular surface(Fig. 4.8)

The patient will receive cataract surgery on OSand right eye ocular surface reconstructionwith a limbal biopsy of his left eye.

4.7 Case Report 63

Fig. 4.6 A, B. Slit-lamp appearance of the left eye inMarch 2004

A

B

Fig. 4.7 A, B. Slit-lamp appearance of the left eye inSeptember 2004. Note the diffuse epithelial toxicity(late fluorescein staining)

A

B

Fig. 4.8 A, B. Impression cytology of the left eye. Note the presence of a normal corneal epithelium and theabsence of globet cells

A B

4.8Future Standard Staging Approach for Ocular Surface Reconstruction

1. IOP control: shunt tube if necessary2. Lids and conjunctival “cul de sac” recon-

structive surgery3. Limbal reconstruction: limbal stem cell cul-

ture will definitely improve the actual resultsof corneal reconstruction. We must remem-ber that the presence of new conjunctivalcells is also necessary in some cases such aspemphigoid or Stevens-Johnson syndrome

4. PKP or deep anterior lamellar keratoplasty ifnecessary

5. Keratoprothesis in those cases where stage 3and 4 fails

References


2. Dua HS, Saini JS, Azuara-Blanco A, Gupta P(2000) Limbal stem cell deficiency: concept, aeti-ology, clinical presentation, diagnosis and man-agement. Ind J Ophthalmol 48:83–92

3. Dua HS (1998) The conjunctiva and corneal ep-ithelial wound healing. Br J Ophthalmol 82:1407–1411

4. Gris O, Guell JL, del Campo Z (2000) Limbal-con-junctival autograft transplantation for the treat-ment of recurrent pterygium. Ophthalmology107: 270–273

5. Gris O,Wolley-Dod C, Güell JL et al. (2002) Histo-logical findings after amniotic membrane graft inthe human cornea. Ophthalmology 109:508–512

6. Gris O, del Campo Z, Wolley-Dod C et al. (2003)Conjunctival healing after amniotic membranegraft over ischemic sclera. Cornea 22(7):675–678

7. Grueterich M, Espana E, Tseng SC (2002) Con-nexin 43 expression and proliferation of humanlimbal epithelium on intact and denuded amniot-ic membrane. Invest Ophthalmol Vis Sci 43:63–71

8. Koizumi N, Inatomi T, Suzuki T et al. (2001) Cul-tivated corneal epithelial stem cell transplanta-tion in ocular surface disorders. Ophthalmology108:1569–1574

9. Kruse FE, Joussen AM, Rohrschneider K et al.(2000) Cryopreserved human amniotic mem-brane for ocular surface reconstruction. GraefesArch Clin Exp Ophthalmol 238:68–75

10. Lee SH, Tseng SCG (1997) Amniotic membranetransplantation for persistent epithelial defectswith ulceration. Am J Ophthalmol 123:303–312

11. Meller D, Pires RTF, Tseng SC (2002) Ex vivopreservation and expansion of human limbal ep-ithelial stem cells on amniotic membrane cul-tures. Br J Ophthalmol 86:463–471

12. Rama P, Bonini S, Lambiase A et al. (2001) Autol-ogous fibrincultured limbal stem cells perma-nently restore the corneal surface of patients withtotal limbal stem cell deficiency. Transplantation72:1478–1485

13. Tsai RJF, Li L-M, Chen J-K (2000) Reconstructionof damaged corneas by transplantation of autolo-gous limbal epithelial cells. N Engl J Med 343:86–93


5.1Introduction

In the past few years, deep anterior lamellar ker-atoplasty (DALK) has seen renewed interest asan alternative to conventional penetrating ker-atoplasty. The introduction of several new dis-section techniques, the optical visualisation of

dissection depth during surgery, and the avail-ability of various lasers may provide new possi-bilities for the management of anterior cornealdisorders. In fact, we may be currently witness-ing the most dramatic change in the concept ofkeratoplasty from being a conventional pene-trating procedure towards that of custom-madecorneal tissue replacements. Below, the mostrecent developments are summarised and themost important issues concerning DALK arediscussed.

5.2Main Drawbacks of Conventional DALK

Since the early 1900s, penetrating keratoplastyhas been the procedure preferred by most sur-geons for the treatment of corneal disorders. Onaverage, the overall clinical result of penetratingkeratoplasty seems rather poor and is oftencomplicated by a high degree of astigmatism,suture related complications, and incompletewound healing. Despite its advantages, i.e. lessrisk of intraocular complications and allograftrejection, a lamellar procedure is often consid-ered troublesome, mainly because of the risk ofperforation during surgery, the development ofinterface haze, and the time-consuming, tedioussurgical technique.

5.3Different Concepts

To overcome these problems, several techniquesare now available with which to perform a con-trolled deep anterior lamellar keratoplasty pro-

Deep Anterior Lamellar Keratoplasty

Gerrit R.J. Melles

5

|

∑ With current techniques, the clinical out-come of a lamellar keratoplasty may besimilar to that after penetrating keratoplasty

∑ Compared to conventional penetrating keratoplasty, new lamellar keratoplastytechniques provide safer “closed-system”surgeries with less morbidity and betterclinical outcomes

∑ The scope of surgical tools for lamellar keratoplasty has been expanded to providefeasible and adjustable techniques

∑ Since postoperative treatment may be the most challenging aspect in kerato-plasty surgery, careful patient selection and psychologic preparation are importantin achieving good results

∑ Especially in the patient population eligiblefor lamellar keratoplasty, the proceduremay have major advantages for both thesurgeon and the patient, such as longergraft survival, less aftercare and less dependency on the health care system

∑ Lamellar keratoplasty surgery may be slowly tending towards the use of custom-made transplants

Core Messages

cedure at a planned corneal depth with minimalrisk of perforation and interface haze develop-ment.∑ The creation of an optical interface by filling

the anterior chamber with air allows the sur-geon to visually control the dissection depthduring the entire surgery. This allows thesurgeon to choose the desired dissectiondepth. In cases in which a penetratingkeratoplasty may have a bad prognosis, forexample a patient with Down syndrome orrecurrent herpes simplex virus keratitis, thesurgeon may aim for a relatively shallow dis-section depth to complete the procedurewith a minimal risk of perforation.

∑ Instead of removing the anterior corneal tis-sue layer by layer, the cornea may be dissect-ed to the desired depth at the first go. Thissaves time, and if a perforation occurs, it is inthe very early phase of the surgery, so theprocedure can be quickly converted to apenetrating keratoplasty.

∑ Instead of performing a blunt dissection,several instruments and medical devices aswell as various lasers are currently availablewith which to obtain a regular and smoothdissection plane.


∑ New lamellar keratoplasty techniques allowfor a single stromal dissection at an optical-ly controlled depth

∑ Improved instrument designs and variouslasers allow the creation of a smooth recipi-ent stromal bed

5.4Important Preoperative Considerations

Careful patient selection and patient prepara-tion may in part determine whether the out-come of a deep anterior lamellar keratoplastyprocedure is considered succesful both by thesurgeon and the patient. The main parametersfor patient selection are:∑ Endothelial cell density. The good condition

of the recipient endothelium is a prerequisitefor a lamellar procedure. Although relativelyinfrequent, corneas with an anterior corneal

disorder may also have a compromised en-dothelium, due to combined disease, the dis-ease itself or prolonged medication (for ex-ample, combined dystrophies or herpessimplex virus keratitis/endotheliitis).

∑ Penetrating keratoplasty bad prognosis. Inpatients in whom a conventional penetratingprocedure is contraindicated, complete visu-al rehabilitation is often not the primarygoal. In these cases, a relatively shallow dis-section depth (70–80%) may be considered,which greatly reduces the risk of perforationduring surgery, for example, extreme kerato-conus in Down syndrome or long-standingrecurrent herpes keratitis.

∑ Patient age. Young patients with an isolatedcorneal disorder like keratoconus may bene-fit the most from a lamellar procedure. Theendothelial cell loss after lamellar kerato-plasty has been found to show a similar pat-tern to that of the physiological cell loss in avirgin cornea, which may significantly im-prove the long-term expectation for graftsurvival. Since the integrity of the globe isbetter preserved in a lamellar procedure andthe risk of wound rupture or dehiscence maybe relatively low, a lamellar keratoplasty maygive fewer restrictions to sports and otherdaily activities that are relatively importantto young people. With a lamellar procedure,the risk of allograft rejection may also begreatly reduced, which may give fewer re-strictions to people travelling to or living incountries with a less sophisticated healthcare system.

∑ Atopic constitution. The long-term results ofany type of keratoplasty procedure may berelatively poor in patients with atopic diseaseor a concurrent facial skin disorder. In ourseries, the occurrence of relatively seriousand long-standing complications such aspersistent epithelial defects, suture infil-trates, and partial melts proved relatively fre-quent in these cases and difficult to manage.

∑ Alternative treatments. Several proceduresare currently available that may allow for avisual outcome similar to a deep anteriorlamellar keratoplasty, but that are more pa-tient friendly. For example, amniotic mem-brane transplantation with subsequent con-

66 Chapter 5 Deep Anterior Lamellar Keratoplasty

tact lens fitting often provides a useful visualacuity. Phototherapeutic keratectomy can beeffective in the treatment of epithelial dys-trophies. Intrastromal ring segments maygive fairly good results in corneal thinningdisorders. The femtosecond laser may alsobroaden the possibilities for replacement ofspecific corneal layers.


∑ Important considerations prior to lamellarkeratoplasty include assessment of the con-dition of the recipient endothelium, overallprognosis, patient age, presence of an atopicconstitution, and less invasive treatments

5.5Psychological Preparation of the Patient

Once the decision to perform a deep anteriorlamellar procedure has been taken, it is recom-mended that the expectations of the patient areappropriately modified. Since most patients arefamiliar with the outcome of cataract surgery,the surgeon may explain to the patient that cur-rent keratoplasty surgical techniques on averagedo not provide similar visual results.

Since a lamellar procedure always bears therisk of perforation and the need for conversionto a penetrating procedure, it should be ex-plained to the patient that a lamellar keratoplas-ty will be attempted but that in the surgery hasto be completed as a conventional penetratingprocedure. Although the risk of perforationwith some techniques may be as low as 5%, thepatient then is less likely to perceive the surgeryas ‘a failure’ in the event of a perforation, i.e.conversion to a penetrating procedure. The pa-tient should also be informed that secondarysurgery may be necessary shortly after the ker-atoplasty to position the donor properly. If theanterior chamber needs to be filled with air tomanage a perforation, the patient anticipates on surgical aftercare, whereas such a minorsecondary treatment may otherwise alarm thepatient as it is quickly perceived as emergencysurgery to save the eye.

In the Netherlands, keratoconus is the indi-cation in approximately half of the patients

eligible for deep anterior lamellar keratoplasty.A fairly large number of these patients may havea long history of uncomplicated contact lenswear up to the moment that the steepeningcorneal contour causes unacceptable contactlens discomfort. Since none of the currentlyavailable DALK procedures allows for a con-trolled restoration of the corneal contour, andpostoperative contact lens fitting often greatlyimproves the final visual outcome, it may be rec-ommended to inform the patient that the goalof the surgery is to flatten the cornea and enablecontact lens wear rather than to restore thecorneal surface contour.


∑ Patient preparation for lamellar kerato-plasty may include a downgrading ofexpectations and a definition of the intended goal of surgery

5.6Choice of DALK Surgical Technique

Given the indication, the presence of a con-traindication for penetrating keratoplasty, andthe anatomy of the individual patient, the sur-geon may first want to consider whether intra-operative perforation is unacceptable and whatdissection depth is desired. In most cases, a dis-section depth of >90% or a separation of De-scemet’s membrane from the recipient posteri-or stroma will give the best postoperative result.The surgeon may also want to consider whatperforation might occur and/or how it shouldbe dealt with during the surgery.∑ Air- or viscodissection of Descemet’s mem-

brane. Separation of Descemet’s membranefrom the recipient posterior stroma by usingair dissection (big bubble technique) or vis-codissection has the advantage that a nearanatomical donor to host interface can beobtained with a minimal risk of interfacehaze development. The dissection method isindirect (not manually controlled but de-pendent on the pressure built by air or vis-coelastic at the cleavage plane above De-scemet’s membrane) and, as a result, if aperforation occurs, the perforation tends to

5.6 Choice of DALK Surgical Technique 67

be paracentral and large, often requiringconversion to a penetrating graft. For thatreason, the surgeon may choose to havedonor tissue available with good quality en-dothelium.

∑ Manual corneal dissection at a visually con-trolled depth. To monitor the dissectiondepth during surgery, the anterior chambershould be filled with air to create an opticalreference plane. Using an optical reflex, thesurgeon can perform a dissection of up to90–95% depth, removing most of the dis-eased corneal stroma. Visualisation of thedepth of the dissection instruments greatlyreduces the risk of perforation to less than5%. If a perforation occurs, the perforation isusually small and located at the 12 o’clocksurgical position or in the far periphery. Ifnecessary, the perforation can be sealed bydissecting slightly shallower ‘over’ the perfo-ration site, so that the hole is closed by a self-sealing stromal flap. The number of cas-es requiring conversion to a penetrating pro-cedure may therefore be low, so that theavailability of a donor cornea with goodquality endothelium is not mandatory.

∑ Laser dissections. Several groups have evalu-ated the use of the excimer laser to create ahost bed for a deep lamellar graft. The cur-rent equipment and software allows for theinput of topography and pachymetry values,so that recipient corneas with an irregularsurface or thinned stroma can be managedwith a topo-linked and pachy-linked deepexcimer ablation. With the introduction ofthe femtosecond laser, the precision of theablation/dissection plane created may fur-ther improve, but not all software currentlyallows dissections over 400 mm in depth.


∑ The lamellar keratoplasty surgical tech-nique should be chosen taking into consid-eration the desired dissection depth, man-agement of inadvertent perforations andavailability of donor tissues

5.7Clinical Results (Figs. 5.1–5.3)

In conventional anterior lamellar keratoplasty,the recipient anterior corneal tissue was re-moved ‘layer by layer’. The most common tech-nique was that of lifting the tissue, stretchingthe fibres at the dissection plane and cutting thefibres with a crescent knife. The donor tissuewas usually dissected using a blunt spatula toavoid perforation. Although the approach ap-peared effective, the dissection methods mayhave been the major cause for the developmentof interface haze, the major drawback of alamellar procedure, since it affected the clinicaloutcome significantly.

Physically the normal cornea is milky white,but within the spectrum of visible light it ap-pears transparent through ‘constructive inter-ference’, i.e. the incoming light rays will bouncebetween the layered collagen fibres until theyhave crossed the cornea. As a result, any disor-ganisation of the layered fibrous structure maybe expected to degrade the ability of the light topass the cornea. The scattered light is clinicallyobserved as opaque areas at the donor-to-hostinterface, or interface haze.


Fig. 5.1. Slit-lamp photograph of a deep anteriorlamellar keratoplasty on the first postoperative day.Note that the organ cultured full-thickness donorbutton positioned onto the lamellar bed is stillswollen

From LASIK treatments we learned thatcorneal dissections do not per se induce inter-face haze. Since histologically interface haze canbe correlated with interface scarring, it seemsessential to obtain a smooth interface. With allthe approaches mentioned above a smooth hostbed can be obtained.With air or viscodissection

of Descemet’s membrane the anterior De-scemet’s will be exposed, excimer laser ablationwill provide a smooth host bed, and if the dis-section is made manually it may be recom-mended to use sharp instruments rather than ablunt dissection knife.When the host bed is suf-ficiently deep, a full-thickness donor button canbe positioned onto the host bed. Descemet’smembrane may be stripped off the donorcornea, which leaves a perfectly smooth surfaceof posterior corneal stroma.

As a result, current techniques for a deep an-terior lamellar keratoplasty are not associatedwith significant interface haze development,and multiple investigators have found the finalvisual performance to be similar in eyes with alamellar or penetrating keratoplasty. The finalastigmatic error may be lower with lamellarprocedures, especially when a large graft diam-eter is used. The visual outcome does vary withthe indication for surgery. For example, eyesgrafted for keratoconus on average achieve amuch better visual acuity than those grafted forrecurrent herpes simplex.

Thus, with the main drawback in performinga lamellar transplant eliminated, there may befew arguments to prefer a penetrating to alamellar procedure. First, because the endothe-lial cell density after lamellar keratoplasty mayshow a pattern of cell loss as in virgin corneas,the long-term survival of a lamellar graft maybe expected to be much better than after. Sec-ond, the risk of allograft rejection is minimised.Although stromal rejections with secondarykeratic precipitates onto the recipient endothe-lium do occur, such rejections are most oftenless severe and easily managed with a topicalsteroid pulse therapy. Third, a lamellar proce-dure may provide the patient with much moresocial freedom. The integrity of the eye is betterpreserved, sutures can be removed much soon-er, medication can be tapered quicker, and theoverall aftercare necessary to maintain a func-tional graft is greatly reduced. Since deep ante-rior lamellar keratoplasty indications are limit-ed to anterior corneal disorders, and thesedisorders are usually found in young people,long-term graft survival as well as less depend-ency on the health care system is important inthis patient population.


Fig. 5.2. Slit-lamp photograph of a deep anteriorlamellar keratoplasty performed with manual dissec-tion at an optically controlled corneal depth

Fig. 5.3. Slit-lamp photograph of a deep anteriorlamellar keratoplasty performed using viscodissec-tion of Descemet’s membrane. Note the remnant re-cipient stromal fibres at the level of Descemet’s mem-brane that are typical for the procedure


∑ The clinical outcome of a lamellar kerato-plasty with current techniques may be simi-lar to that after penetrating keratoplasty

∑ Especially in the patient population eligiblefor lamellar keratoplasty, the proceduremay have major advantages for both thesurgeon and the patient, such as longergraft survival, the need for less aftercareand a lower dependence on the health caresystem

References

1. Alió JL, Attia WH (1999) LASIK after previouscorneal surgery. In: Burato L, Brint S (eds) LASIK,2nd edn, chap 30. Slack Inc., Thorofare, NJ, pp519–535

2. Alió JL, Shah S, Barraquer C, Bilgihan K, AnwarM, Melles GR (2002) New techniques in lamellarkeratoplasty. Curr Opin Ophthalmol 13:224–229

3. Amm M, Nolle B (2002) Gore-tex patch in im-munologically conditioned corneal ulcer. KlinMonatsbl Augenheilkd 219:735–739

4. Anwar M (1972) Dissection technique in lamellarkeratoplasty. Br J Ophthalmol 56:711–713

5. Anwar M (1974) Technique in lamellar kerato-plasty. Trans Ophthalmol Soc UK 94:163–171

6. Anwar M, Teichmann KD (2002) Big-bubble tech-nique to bare Descemet’s membrane in anteriorlamellar keratoplasty. J Cataract Refract Surg28:398–403; comment in: J Cataract Refract Surg2002 Dec; 28(12):2067; author reply 2067–2068

7. Anwar M, Teichmann KD (2002) Deep lamellarkeratoplasty: surgical techniques for anteriorlamellar keratoplasty with and without baring ofDescemet’s membrane. Cornea 21(4):374–383

8. Archila EA (1984/85) Deep lamellar keratoplastydissection of host tissue with intrastromal air in-jection. Cornea 3:217–218

9. Balestrazzi E, Balestrazzi A, Mosca L, BalestrazziA (2002) Deep lamellar keratoplasty with trypanblue intrastromal staining. J Cataract RefractSurg 28:929–931

10. Barraquer JI (1972) Lamellar keratoplasty (spe-cial techniques). Ann Ophthalmol 437–469

11. Barraquer JI (1989) Cirugia refractiva de lacornea. Instituto Barraquer de América, Bogotá,p 310

12. Bhojwani RD, Noble B, Chakrabarty AK, StewartOG (2003) Sequestered viscoelastic after deeplamellar keratoplasty using viscodissection.Cornea 22:371–373

13. Bilgihan K, Ozdek SC, Sari A, Hasanreisoglu B(2003) Microkeratome-assisted lamellar kerato-plasty for keratoconus: stromal sandwich.J Cataract Refract Surg 29:1267–1272

14. Bodenmueller M, Goldblum D, Frueh BE (2003)Penetrating keratoplasty in Down’s syndrome.Klin Monatsbl Augenheilkd 220:99–102

15. Chau GK, Dilly DA, Sheard CE, Rostron CK (1992)Deep lamellar keratoplasty on air with lyo-philized tissue. Br J Ophthalmol 76:646–650

16. Coombes AG, Kirwan JF, Rostron CK (2001) Deeplamellar keratoplasty with lyophilised tissue inthe management of keratoconus. Br J Ophthalmol85:788–791

17. De Molfetta V,Brambilla M,De Casa N et al. (1979)Residual corneal astigmatism after perforatingkeratoplasty. Ophthalmologica 179:316–321

18. Duong MH, Thimel S, Xuan TH (2001) Lamellarkeratoplasty with air or viscoelastic substance in-jection. J Fr Ophtalmol 24:930–936

19. Edwards M, Clover GM, Brookes N, PendergrastD, Chaulk J, McGhee CN (2002) Indications forcorneal transplantation in New Zealand: 1991–1999. Cornea 21:152–155

20. Guell JL, Velasco F, Guerrero E, Gris O, CalatayudM (2003) Preliminary results with posteriorlamellar keratoplasty for endothelial failure. Br JOphthalmol 87:241–243

21. Gupta V, Dada T, Pangtey M, Vajpayee RB (2001)Indications for lamellar keratoplasty in India.Cornea 20:398–399

22. Haugen OH, Hovding G, Eide GE, Bertelsen T(2001) Corneal grafting for keratoconus in men-tally retarded patients. Acta Ophthalmol Scand79:609–615

23. Hirano K, Sugita J, Kobayashi M (2002) Separa-tion of corneal stroma and Descemet’s mem-brane during deep lamellar keratoplasty. Cornea21:196–199

24. Hirano K, Kojima T, Nakamura M, Hotta Y (2001)Triple anterior chamber after full-thicknesslamellar keratoplasty for lattice corneal dystro-phy. Cornea 20:530–533

25. Hope-Ross MW, McDonell PJ, Corridan PG et al.(1993) The management of post-keratoplastyastigmatism by post-operative adjustment of asingle continuous suture. Eye 7:625–628

26. Jain S, Azar DT (2001) New lamellar keratoplastytechniques: posterior keratoplasty and deeplamellar keratoplasty. Curr Opin Ophthalmol12:262–268


27. Jones DH, Kirkness CM (2001) A new surgicaltechnique for keratoglobus-tectonic lamellarkeratoplasty followed by secondary penetratingkeratoplasty. Cornea 20:885–887

28. Krumeich JH, Schoner P, Lubatschowski H,Gerten G, Kermani O (2002) Excimer laser treat-ment in deep lamellar keratoplasty 100 microme-ters over Descemet’s membrane. Ophthalmologe99:946–948

29. Kruse FE, Reinhard T (2001) Limbus transplanta-tion for reconstruction of the ocular surface.Ophthalmologe 98:818–831

30. Laibson PR (2002) Current concepts and tech-niques in corneal transplantation. Curr OpinOphthalmol 13:220–223

31. Maldonado-Bas A,Pulido R (1995–1996) Querato-plastia Laminar con Anestesia Tópica. ArchSAOO 24:341–345

32. Manche EE, Holland GN, Maloney RK (1999)Deep lamellar keratoplasty using viscoelastic dis-section. Arch Ophthalmol 117:1561–1565

33. Maurino V, Allan BD, Stevens JD, Tuft SJ (2002)Fixed dilated pupil (Urrets-Zavalia syndrome) af-ter air/gas injection after deep lamellar kerato-plasty for keratoconus.Am J Ophthalmol 133:266–268

34. Melles GR (2002) Posterior lamellar keratoplasty.Arch Soc Esp Oftalmol 77:175–176

35. Melles GRJ, Remeijer L, Geerards A, Beekhuis WH(1999) The future of lamellar keratoplasty. CurrOpin Ophthalmol 10:253–259

36. Melles GRJ, Rietveld FJR, Beekhuis WH, BinderPS (1999) A technique to visualize corneal inci-sion and lamellar dissection depth during sur-gery. Cornea 18:80–86

37. Melles GRJ, Lander F, Rietveld FJR, Remeijer L,Beekhuis WH, Binder PS (1999) A new surgicaltechnique for deep stromal, anterior lamellarkeratoplasty. Br J Ophthalmol 83:327–333

38. Melles GRJ, Remeijer L, Geerards AJM, BeekhuisWH (2000) A quick surgical technique for deeplamellar keratoplasty using visco-dissection.Cornea 19:427–432

39. Melles GRJ, Eggink FAGJ, Lander F, Pels E,Rietveld FJR, Beekhuis WH, Binder PS (1998) Asurgical technique for posterior lamellar kerato-plasty. Cornea 17:618–626

40. Melles GRJ, Lander F, Beekhuis WH, Remeijer L,Binder PS (1999) Posterior lamellar keratoplastyfor a case of pseudophakic bullous keratopathy.Am J Ophthalmol 127:340–341

41. Melles GRJ, Lander F, van Dooren BTH, Pels E,Beekhuis WH (2000) Preliminary clinical resultsof posterior lamellar keratoplasty. Ophthalmolo-gy 107:1850–1857

42. Melles GR, Lander F, Nieuwendaal C (2002) Su-tureless, posterior lamellar keratoplasty: a casereport of a modified technique. Cornea 21:325–327

43. Melles GRJ, Rietveld FJR (2002) Transplantationof Descemet’s membrane carrying viable endo-thelium through a small scleral incision. Cornea21:415–418

43. Melles GR, Kamminga N (2003) Techniques forposterior lamellar keratoplasty through a scleralincision. Ophthalmologe 100:689–695

44. Minasian M, Ayliffe W (2002) Fixed dilated pupilfollowing deep lamellar keratoplasty (Urrets-Zavalia syndrome). Br J Ophthalmol 86:115–116

45. Murta JN, Amaro L, Tavares C, Mira JB (1994)Astigmatism after penetrating keratoplasty. Roleof the suture technique. Doc Ophthalmol 87:331–336

46. Price FW Jr (1989) Air lamellar keratoplasty. J Re-fract Corneal Surg 5:240–243

47. Rumelt S, Cohen I, Skandarani P, Delarea Y, BenShaul Y, Rehany U (2001) Ultrastructure of thelamellar corneal wound after laser in situkeratomileusis in human eye. J Cataract RefractSurg 27:1323–1327

48. Saini JS, Jain AK, Sukhija J, Saroha V (2003) Indi-cations and outcome of optical partial thicknesslamellar keratoplasty. Cornea 22:111–113

49. Seitz B, Langenbucher A, Hofmann-Rummelt C,Schlotzer-Schrehardt U, Naumann GO (2003)Nonmechanical posterior lamellar keratoplastyusing the femtosecond laser (femto-plak) forcorneal endothelial decompensation. Am J Oph-thalmol 136:769–772

50. Shimazaki J, Shimmura S, Ishioka M, Tsubota K(2002) Randomized clinical trial of deep lamellarkeratoplasty vs penetrating keratoplasty. Am JOphthalmol 134:159–165

51. Trimarchi F, Poppi E, Klersy C (2002) Deep lamel-lar keratoplasty. J Fr Ophtalmol 25:718–721

52. Trimarchi F, Poppi E, Klersy C, Piacentini C (2001)Deep lamellar keratoplasty. Ophthalmologica215:389–393

53. Vajpayee RB, Bhartiya P, Sharma N (2002) Centrallamellar keratoplasty with peripheral intralamel-lar tuck: a new surgical technique for kerato-globus. Cornea 21:657–660

54. Xie L, Shi W, Liu Z, Li S (2002) Lamellar kerato-plasty for the treatment of fungal keratitis.Cornea 21:33–37

55. Yang YF, Matheson M, Dart JK, Cree IA (2001)Persistence of acanthamoeba antigen followingacanthamoeba keratitis. Br J Ophthalmol 85:277–280

References 71

6.1Introduction

Despite the relative immune privilege of thecornea as a transplant tissue and both the recip-ient corneal bed and anterior chamber being animmune privileged site [35, 49], the most com-mon cause of corneal graft failure in all reportsis allogeneic rejection. In first graft recipientswith no vascularisation of the recipient cornealbed, 2-year survival rates exceed 90%; this de-creases to 35–70% in recipients with high risk

factors for rejection [18, 53]. In one-third of allcorneal grafts that are deemed failed, signs of adestructive attack by the immune system havebeen observed [61].

A rejection episode results in loss of donorendothelial cells, critical for maintenance ofcorneal transparency. As human endothelialcells do not repair by mitosis to any meaningfulextent, the consequence is that donor cornealtransparency is lost if cell density falls below thethreshold necessary for prevention of stromalswelling. Endothelial decompensation resultseither (1) from the time of an irreversibleepisode of acute graft rejection or (2) at an in-terval following one or more episodes of rejec-tion which have been reversed by therapy. En-dothelial cells are thus the critical target in theallogeneic response.

While, on the one hand, reversal of acutegraft rejection episodes does not present suchchallenges in cornea as in other transplantedtissues, effective prophylaxis in corneal graft re-cipients identified at high risk of rejection ismuch less evidence-based. Thus the impact ofgraft rejection continues to justify high priorityin corneal research. Although the first success-ful penetrating corneal graft was reported in1906, it took another half a century before thefirst description of opacification of a previouslyclear corneal graft was published. Paufiquenamed this event “maladie du greffon” (diseaseof the graft) and suggested that this clinicalfinding was caused by sensitisation of the donorby the recipient [37]. This description followedthe experiments reported by Medawar a fewyears previously, in which differences were ob-served between rabbit skin grafts of donor andrecipient origin, giving rise to the term “histo-

Corneal Transplant Rejection

T.P.A.M. Slegers, M.K. Daly, D.F.P. Larkin

6

|

∑ Allograft rejection is the commonest singlecause of corneal graft failure

∑ Corneal transplant antigen recognition inmost cases is almost exclusively mediatedby recipient antigen-presenting cells.CD4+ T-lymphocytes have the central rolein the alloreactive cell population

∑ Endothelial rejection episodes can be reversed by intensive topical steroid inmost patients. Poor outcomes result fromdelay in presentation and/or initiation intreatment

∑ Patients with recipient corneal vascularisa-tion, a previously rejected ipsilateral trans-plant and inflammation at the time oftransplantation, are at highest risk of rejec-tion and have very poor graft survival

∑ Little information is available from ran-domised trials on prophylaxis by transplan-tation antigen matching or immunosup-pression in this patient group

Core Messages

compatibility” [32]. Maumenee subsequentlyconfirmed this suggestion in a rabbit model ofcorneal transplantation in which he showedthat donor corneas could induce an immune re-action [31]. The development of corneal trans-plantation models in rat [59] and mouse [46] fa-cilitated study of rejection in inbred donor andrecipient animals with a wide range of inves-tigative immunological reagents.

6.2Incidence

In reports from large cohorts of corneal graftrecipients, the proportion undergoing a rejec-tion episode at some stage post-transplantranges from 18% to 21% [12, 22, 60]. In thosegraft recipients in whom rejection occurs, re-ported rates of successful reversal of the rejec-tion episode range from 50% to 90% [21, 34].Allograft rejection occurs most commonly inthe second 6 months postgrafting, and it hasbeen reported that more than 10% of the ob-served reactions can take place as late as at least4 years after surgery [23, 34, 39]. This indicatesthat all corneal grafts need long-term surveil-lance and are at risk practically indefinitely.

6.3Factors Predisposing to Corneal Graft Rejection

Preoperative characteristics of the graft recipi-ent eye can be clearly identified in many pa-tients to indicate significantly high risk of graftfailure. Proposed graft recipient corneas (1)with two or more quadrants of deep vasculari-sation, (2) bearing a previously rejected graft(Fig. 6.1) and (3) that are inflamed at the time oftransplantation are at significantly higher riskof rejection [2, 6, 30, 57, 63, 62]. There is lessrobust evidence in the published literature thatgrafts in children, large diameter donor corneasand proximity of donor cornea to the recipientlimbus are at higher risk (Table 6.1) [30, 41, 54,56]. Clearly more than one of these factors maybe operational in one patient. There may also beassociation of one or more of the above factors

predisposing to failure due to rejection with ad-ditional clinical features that confer significantrisk of graft failure due to other complications,such as glaucoma or ocular surface disease [26,41]. These preoperative clinical features must beevaluated carefully in the decision whether toproceed with corneal transplantation.

Once transplantation is successfully com-pleted, care must be taken to prevent postoper-ative events which predispose to rejection, suchas vascularisation of recipient cornea (Fig. 6.2)or graft wound, suture loosening, or graft infec-

74 Chapter 6 Corneal Transplant Rejection

Fig. 6.1. Subepithelial infiltrates in a penetratingcorneal allograft. The appearance is similar to thatseen in adenovirus keratitis, involving the donorcornea only

Table 6.1. Risk factors for rejection

Preoperative Deep vascularisation of two or more quadrants

Previously rejected ipsilateral graft

Active corneal inflammation at time of graft

Paediatric graft recipient

Large diameter graft

Graft proximity to the limbus

Postoperative Loosening of sutures

Removal of sutures,wound dehiscence

Graft inflammation

HSV infection recurrence

Non-viral graft infection

tion by bacteria or recurrent herpes simplexvirus (HSV).

6.4Clinical Features

Epithelial rejection, diagnosed by a linear opac-ity which stains with fluorescein, comprised upto 10% of all rejection episodes in one seriesand occurs on average 3 months after grafting[1].Although dead donor epithelial cells are rap-idly replaced by recipient epithelial cells and noscarring occurs, the presence of this type ofrejection reflects that the recipient is now sensi-tized to the donor and can progress to stromaland/or endothelial rejection. Stromal rejectionis characterised by nummular subepithelialinfiltrates (Fig. 6.1), identical to those found inadenovirus keratitis. Patients with both epithe-lial and stromal types of rejection may beasymptomatic or have mild ocular discomfortonly. In contrast, patients with endothelial re-jection will usually present with visual distur-bance and iritis symptoms. If examined earlyafter rejection symptom onset, anterior cham-ber cell infiltration without flare or graft abnor-mality will be seen.At later times after symptomonset, the signs in succession are (1) aggregatedalloreactive cells adherent to graft endotheliumevident as keratic precipitates, (2) an endo-thelial line with precipitates and (3) localisedoedema corresponding to a rejection line ortotal graft oedema (Fig. 6.2). Visible graft pre-cipitates on slit-lamp biomicroscopy imply focaland variable but irreversible endothelial cellloss, compromising endothelial pump functionand resulting in stroma oedema in those graftswith severe inflammation or low endothelial celldensity prior to rejection onset. Pachymetry ishelpful in detecting an increase in oedema andalso deturgescence following the start of steroidtreatment. In one study it was found that next to the preoperative diagnosis, graft thicknessduring rejection, as objectively measured bypachymetry, is a prognostic sign for reversibili-ty of a rejection episode [34]. Risk factors forsignificant endothelial cell loss are delay in ini-tiating anti-rejection treatment more than 1 dayand recipient age greater than 60 years [13].

6.5Histopathology

Descriptions of the pathological features ofcorneal transplant rejection result from exami-nation of grafts replaced following irreversiblefailure. Therefore these specimens illustrate latechanges in end-stage corneal opacification,usually some months at least following treat-ment of rejection. Characteristic findings instroma are vascularisation with mononuclearcell infiltration and keratocyte loss; few if anyendothelial cells remain (Fig. 6.3) [28]. Severalstudies have shown increased numbers of HLAclass II positive cells infiltrating stroma in sec-tions of rejected grafts [38, 58].

6.5 Histopathology 75

Fig. 6.2. Endothelial rejection line, keratic precipi-tates and folds in Descemet’s membrane in rejection

Fig. 6.3. Almost total loss of endothelial cells incorneal graft specimen removed at graft replacement6 months following rejection onset

6.6Immunopathological Mechanisms

6.6.1Immune Privilege and Its Breakdown

Immune privilege is a dynamic phenomenon in which the destructive effect of a “normal”immune response to particular antigens iseither altered or absent in order to protect themicroanatomy of highly organised tissues in the eye. In corneal transplantation, both (1) therecipient corneal bed and anterior chamber and(2) the transplanted tissue have features ofimmune privilege.

Several features of the anterior chamber con-tribute to immune privilege. There are mechan-ical barriers that impair immune cell access tothe anterior chamber and transplanted cornea.One barrier is the lack of blood and lymphaticvessels in a normal cornea. While experimentaland clinical studies have clearly shown thattransplants are much more likely to be rejectedin vascularised corneas, the stimuli to vascular-isation are likely also to induce lymph vesselgrowth. Following transplantation, it is in lymphvessels that antigen-presenting cells (APC) mi-grate from the graft to lymphoid organs for pres-entation of graft antigens to T lymphocytes.An-other route for alloreactive cells to reach theanterior chamber and donor corneal endotheli-um is closed by the tight junction barrier formedbetween non-pigmented epithelial cells andnon-fenestrated iris vessels [10].

In the event that leukocytes enter the anteri-or chamber, mechanisms are available to eitherdeviate or blunt a potentially harmful immuneresponse. For example the aqueous humourcontains immunosuppressive molecules as trans-forming growth factor (TGF)-b, vasoactiveintestinal polypeptide (VIP), a-melanocytestimulating hormone (MSH), and calcitoningene related protein (CGRP), which contributeto induction by an allograft of deviated sys-temic delayed-type hypersensitivity [35, 52].

In addition to lack of blood and lymphaticvessels, cornea allografts have been shown inlaboratory studies to have additional featureswhich contribute to immune privilege. These

include the paucity of donor-derived major his-tocompatibility complex (MHC) class II+ APC,and corneal epithelial and endothelial expres-sion of Fas ligand [7, 16], interaction of whichwith Fas on alloreactive effector cells leads todeath of the infiltrating leukocyte.

Corneal grafts at high risk of rejection areidentified by several risk factors, most of whichreflect breakdown of facets of immune privi-lege. Prospective clinical outcome studies iden-tify the most significant of these to be recipientcorneal vascularisation, corneal inflammationat the time of transplantation, which inducesAPC infiltration in the recipient cornea prior to surgery, and a previously rejected ipsilateralgraft.

6.6.2Afferent Arm of the Allogeneic Response

In circumstances where the immune privilegedfeatures of the cornea are bypassed by the im-mune system, the first stage in rejection isrecognition of the presence of non-self tissue.There are two routes of allorecognition. By theindirect pathway, recipient APC enter the graftto capture and process donor antigens, migrat-ing to the lymphoid system to present the anti-gen in context with self MHC class II moleculesto T cells. Most experimental evidence points tothe neck lymph glands as the location for anti-gen presentation [40, 45, 65].

Recent identification in the central cornea of a population of dendritic cells, which can be-come MHC II+ and migrate to the draininglymph nodes [11, 17, 29], and MHC class II+

macrophages [9] indicates that direct allore-cognition of the corneal graft antigens is possi-ble. By this pathway, donor APC bearing allo-antigens migrate from the graft and activate Tlymphocytes via their own non-self MHC classII molecules. Direct allorecognition would bemore prominent in the occasional clinical cir-cumstance in which a donor cornea is trans-planted which has an increased population ofAPCs, such as after viral infection. However, inmost circumstances it is assumed that cornealallorecognition is predominantly by the indi-rect pathway.


Evidence from cell kinetic studies in murinecorneal grafts demonstrates that within severalhours of transplantation the graft is infiltratedby granulocytes and macrophages [27]. Frommacrophage depleting studies evidence hasbeen provided that these cells play a crucial rolein the afferent phase of graft rejection [47].

6.6.3Efferent Arm of the Allogeneic Response

When T-helper cells have identified the present-ed antigen as non-self, effector mechanisms aregenerated against donor tissue. Cytokines in-cluding particularly tumour necrosis factor [42]and interferon-g [25] have been clearly identi-fied in aqueous humour and the cornea prior toobserved endothelial rejection onset. Aftercorneal transplantation it has been shown thatalloantibody, cytotoxic T lymphocytes and de-layed type hypersensitivity responses are com-ponents of the effector response. Experimentalstudies, using CD4+ knockout mice and mono-clonal antibodies directed against CD4+ T cells,have pointed to the central role of this lympho-cyte subpopulation [3, 64]. The mechanism bywhich corneal endothelial cells are killed is notyet clear. At time of graft destruction increasinglevels of natural killer (NK) cells, known to beable to lyse corneal endothelial cells, are detect-ed in the aqueous humour of grafted rats [14].There is additional evidence that nitric oxidecould mediate in destruction of donor endo-thelial cells [8, 44, 51].

6.7Treatment of Rejection

The objective of treatment is to reverse the re-jection episode at the earliest possible time, inorder to minimise donor endothelial cell lossand preserve graft function. With the anatomi-cal advantage that corneal transplants aresuperficial, intensive administration of topicalcorticosteroid, such as dexamethasone 0.1%,treatment is successful in reversing most endo-thelial rejection episodes. In most cases inwhich topical steroid fails to reverse rejection, it

is likely to be due to delay in recognition andinitiation of treatment, with resulting signifi-cant donor endothelial cell loss [13]. In others,failure to reverse rejection may be due to failureof topical steroid to reverse effector compo-nents of the allogeneic response. In respect ofadditional systemic steroid, a single dose of in-travenous methylprednisolone was found to bemore effective than oral steroid in patients withendothelial rejection who presented within8 days of onset [20]. A second pulse of intra-venous methylprednisolone at 24 or 48 h gaveno benefit when compared to a single dose atinitial diagnosis [19]. However, a subsequentrandomised trial demonstrated no significantbenefit of intravenous methylprednisolone inaddition to topical steroid, in respect of graftsurvival or interval to a subsequent rejectionepisode within a 2-year follow-up period [21]. Inthe same study, endothelial rejection was re-versed in 33 of 36 patients treated, indicatingthat steroid-resistant rejection is uncommon.Other studies examining the efficacy of topicalor oral cyclosporin administered in combina-tion with intravenous steroid have reportedsimilar outcomes, with irreversible rejection ina small proportion of patients [66, 67].

6.8Prevention of Rejection

6.8.1Immunosuppression

In patients without risk factors for graft rejec-tion identified prior to surgery, typical postop-erative immunosuppression comprises steroiddrops such as dexamethasone 0.1% four timesdaily for the first 2–3 months, reducing gradual-ly to zero by 6 months post-transplant. Regimesvary from centre to centre. There is much lessconsensus on which additional measures to takeas prophylaxis in patients at high risk of rejec-tion (Table 6.1), in whom topical steroid alone is insufficient to prevent rejection. The result of a continuing shortage of large comparativeprospective studies is that immunosuppressionprotocols in current use result from individualclinical experience, with some influence from

6.8 Prevention of Rejection 77

experimental evidence and small uncontrolledand/or retrospective clinical studies. However,ophthalmologists are cautious about adminis-tering potentially toxic systemic immunosup-pressive agents, even in those patients in whoma surviving graft would allow vision in the onlyeye. The subject of immunosuppression in pre-vention of corneal graft rejection is discussed inanother chapter in this text.

6.8.2HLA Matching

In vascularised organ allotransplantation thereis robust evidence supporting HLA matching ofdonor and recipient, with the data of Opelz andothers demonstrating stratification of the riskof rejection according the number of class I andespecially class II mismatches [36]. HLA match-ing is in routine use internationally in cadaver-ic renal and other organ transplantation. Incorneal transplantation by contrast, for recipi-ents at high risk of rejection HLA class I andclass II -DR matching is routinely done in somecountries, whereas in other countries no match-ing takes place at all. Roelen suggested a benefitfor HLA-A and -B matching in high-risk cornealallograft recipients based on his findings thatprimed, donor-specific cytotoxic T cells werepresent in rejected corneas but absent in donorswith good graft function [43]. However, the ben-efit of histocompatibility matching in cornealtransplantation has been disputed and is cer-tainly less clear than for solid organ grafts, evenin corneal recipients at perceived high risk ofgraft rejection. Two large prospective studies onHLA-A, -B, or HLA-DR antigen matching high-risk recipients have reported divergent findings.The Collaborative Corneal Transplant StudiesResearch Group reported that matching of theseantigens did not decrease the risk of cornealgraft failure secondary to rejection [53]. In con-trast the Corneal Transplant Follow-up Studyfound there was increased risk of graft rejectionwith mismatch of HLA class I antigens (relativerisk 1.27 per mismatch), but decreasing risk ofrejection with -DR mismatches (relative risk0.58 per mismatch) in high risk patients [55].This study therefore supported matching at

HLA-A, and -B but not HLA-DR. The possiblebenefit of planned -DR mismatching in a settingof known class I histocompatibility is at presentbeing investigated in an ongoing prospectivetrial, the outcome of which is awaited with in-terest. In 1996, a randomised although retro-spective study reported a beneficial effect ofDRB1 matching in recipients at high risk on ac-count of vascularisation and/or retransplanta-tion [4]. Subsequently a beneficial effect ofHLA-DPB1 matching in high-risk corneal trans-plantation with a significantly higher rate of1-year rejection-free graft survival compared tothose without matching was shown [33].

In corneal transplantation therefore the ef-fect of HLA matching is less than clear and thedata are most ambiguous for class II matching.Resolution of this clinically important issue isnot simple. In contrast to solid organs, results ofmatching for cornea are likely to be influencedby the facts that: (1) allorecognition is predomi-nantly by the indirect pathway in most patients[5], and (2) minor transplantation antigens,shown to have a significant effect on graft sur-vival in untreated rodent recipients [24, 48, 50]and presented by the indirect pathway, remainunmatched in HLA-matched recipients. It isalso worth noting here that the effects of HLAmatching on corneal graft outcome have not yetbeen investigated in the setting of systemic im-munosuppression prophylaxis. Studies in solidorgan transplantation have shown that moreeffective rejection prophylaxis can override anHLA matching effect in unsensitised recipients.

6.9Future Prospects

Reducing the impact of allograft rejection is amajor challenge in corneal disease. It can beexpected that following developments in tech-niques of lamellar keratoplasty,wider use of thistype of surgical procedure, particularly for stro-mal corneal pathology, will reduce the impact ofendothelial rejection. However, the presence ofa group of patients with no alternative to pene-trating keratoplasty, a high risk of rejection andno alternative prophylactic intervention justi-fies clinical trials of novel treatment strategies


[15]. These developments are likely to derivefrom a better understanding of the immunemechanisms of rejection and from importationof refinements in systemic immunosuppressionstrategies validated in solid organ transplanta-tion. There remains one opportunity in cornealgrafting which is unique in transplantation: thepossibility of storage ex vivo of donor corneafor a period of weeks prior to elective surgery.Modification of donor tissue by pharmacologi-cal or gene-based approaches may be of addi-tional benefit in attenuating donor injury by theallogeneic response.

References

1. Alldredge OC, Krachmer JH (1981) Clinical typesof corneal transplant rejection: their manifesta-tions, frequency, preoperative correlates, andtreatment. Arch Ophthalmol 99:599–604

2. Arentsen JJ (1983) Corneal transplant allograftrejection: possible predisposing factors. TransAm Ophthalmol Soc 81:361–402

3. Ayliffe W, Alam Y, Bell EB, McLeod D, HutchinsonIV (1992) Prolongation of rat corneal graft sur-vival by treatment with anti-CD4 monoclonalantibody. Br J Ophthalmol 76:602–606

4. Baggesen K, Lamm LU, Ehlers N (1996) Signifi-cant effect of high-resolution HLA-DRB1 match-ing in high-risk corneal transplantation. Trans-plantation 62:1273–1277

5. Batchelor JR (1995) The laws of transplantation –a modern perspective. Eye 9:152–154

6. Boisjoly H, Tourigny R, Bazin R, Laughrea PA,Dubé I, Chamberland G, Bernier J, Roy R (1993)Risk factors of corneal graft failure. Ophthalmol-ogy 100:1728–1735

7. Bora NS, Gobleman CL, Atkinson JP, Pepose JS,Kaplan HJ (1993) Differential expression of thecomplement regulatory proteins in the humaneye. Invest Ophthalmol Vis Sci 34:3579–3584

8. Bourges J-L, Valamanesh F, Torriglia A, Jeanny J-C, Savoldelli M, Renard G, BenEzra D, de KozakY,Behar-Cohen F (2004) Cornea graft endothelialcells undergo apoptosis by way of an alternate(caspase-independent) pathway. Transplantation78:316–323

9. Brissette-Storkus CS, Reynolds SM, Lepisto AJ,Hendricks RL (2002) Identification of a novelmacrophage population in the normal mousecorneal stroma. Invest Ophthalmol Vis Sci 43:2264–2271

10. Caprioli J (1992) The ciliary epithelia and aque-ous humor. In: Hart WM Jr (ed) Adler’s physiolo-gy of the eye. Mosby, St. Louis, pp 228–247

11. Cella M, Engering A, Pinet V, Pieters J, Lanzavec-chia A (1997) Inflammatory stimuli induce accu-mulation of MHC class II complexes on dendriticcells. Nature 388:782–787

12. Chan CM, Wong TY, Yeong SM, Lim TH, Tan DT(1997) Penetrating keratoplasty in the SingaporeNational Eye Centre and donor cornea acquisi-tion in the Singapore Eye Bank. Ann Acad MedSingapore 26:395–400

13. Claerhout I, Beele H, De Bacquer D, Kestelyn P(2003) Factors influencing the decline in endo-thelial cell density after corneal allograft rejec-tion. Invest Ophthalmol Vis Sci 44:4747–4752

14. Claerhout I, Kestelyn P, Debacker V, Beele H,Leclercq G (2004) Role of natural killer cells inthe rejection process of corneal allografts in rats.Transplantation 77:676–682

15. George AJT, Larkin DFP (2004) Corneal trans-plantation: The forgotten graft. Am J Transplant4:678–685

16. Griffith TS, Brunner T, Fletcher SM, Green DR,Ferguson TA (1995) Fas ligand-induced apoptosisas a mechanism of immune privilege. Science270:1189–1192

17. Hamrah P, Zhang Q, Liu Y, Dana MR (2002) Novel characterization of MHC class II-negativepopulation of resident corneal Langerhans cell-type dendritic cells. Invest Ophthalmol Vis Sci43:639–646

18. Hill JC (1994) Systemic cyclosporine in high-riskkeratoplasty: short- versus long-term therapy.Ophthalmology 101:128–133

19. Hill JC, Ivey A (1994) Corticosteroids in cornealgraft-rejection – Double versus single-pulse ther-apy. Cornea 13:383–388

20. Hill JC, Maske R, Watson P (1991) Corticosteroidsin corneal graft rejection. Oral versus single pulsetherapy. Ophthalmology 98:329–333

21. Hudde T, Minassian DC, Larkin DF (1999) Ran-domised controlled trial of corticosteroid regi-mens in endothelial corneal allograft rejection.Br J Ophthalmol 83:1348–1352

22. Ing JJ, Ing HH, Nelson LR, Hodge DO, BourneWM (1998) Ten-year postoperative results of pen-etrating keratoplasty. Ophthalmology 105:1855–1865

23. Jonas JB, Rank RM, Budde WM (2002) Immuno-logic graft reactions after allogenic penetratingkeratoplasty. Am J Ophthalmol 133:437–443

24. Katami M (1991) Corneal transplantation-im-munologically privileged status. Eye 5:528–548

References 79

25. King WJ, Comer RM, Hudde T, Larkin DFP,George AJT (2000) Cytokine and chemokine ex-pression kinetics after corneal transplantation.Transplantation 70:1225–1233

26. Kuchle M, Cursiefen C, Nguyen NX, Langenbuch-er A, Seitz B, Wenkel H, Martus P, Naumann GO(2002) Risk factors for corneal allograft rejection:intermediate results of a prospective normal-riskkeratoplasty study. Graefes Arch Clin Exp Oph-thalmol 240:580–584

27. Kuffova L, Lumsden L, Vesela V, Taylor JA, FilipecM, Holan V, Dick AD, Forrester JV (2001) Kineticsof leukocyte and myeloid cell traffic in themurine corneal allograft response. Transplanta-tion 72:1292–1298

28. Larkin DF,Alexander RA, Cree IA (1997) Infiltrat-ing inflammatory cell phenotypes and apoptosisin rejected human corneal allografts. Eye 11:68–74

29. Liu Y, Hamrah P, Zhang Q, Taylor AW, Dana MR(2002) Draining lymph nodes of corneal trans-plant hosts exhibit evidence for donor major his-tocompatibility complex (MHC) class II-positivedendritic cells derived from class II-negativegrafts. J Exp Med 195:259–268

30. Maguire MG, Stark WJ, Gottsch JD, Stulting RD,Sugar A, Fink NE, Schwartz A (1994) Risk factorsfor corneal graft failure and rejection in thecollaborative corneal transplantation studies.Collaborative Corneal Transplantation StudiesResearch Group. Ophthalmology 101:1536–1547

31. Maumenee AE (1951) The influence of donor-recipient sensitization on corneal grafts. Am JOphthalmol 34:142–152

32. Medawar PB (1944) The behaviour and fate ofskin autografts and skin homografts in rabbits.J Anat 78:176–199

33. Munkhbat B, Hagihara M, Sato T, Tsuchida F, SatoK, Shimazaki J, Tsubota K, Tsuji K (1997) Associa-tion between HLA-DPB1 matching and 1-year re-jection-free graft survival in high-risk cornealtransplantation. Transplantation 63:1011–1016

34. Naacke HG, Borderie VM, Bourcier T, Touzeau O,Moldovan M, Laroche L (2001) Outcome ofcorneal transplantation rejection. Cornea 20:350–353

35. Niederkorn JY (1999) The immune privilege ofcorneal allografts. Transplantation 67:1503–1508

36. Opelz G, Schwarz V, Engelmann A, Back D, WilkM, Keppel E (1991) Long-term impact of HLAmatching on kidney graft-survival in cyclo-sporine-treated recipients. Transplant Proc 23:373–375

37. Paufique L, Sourdille GP, Offret G (1948) Les Greffes de la Corneé. Masson, Paris

38. Pepose JS, Gardner KM, Nestor MS, Foos RY, Pet-tit TH (1985) Detection of HLA class-I and class-II antigens in rejected human corneal allografts.Ophthalmology 92:1480–1484

39. Pleyer U, Steuhl KP, Weidle EG, Lisch W, Thiel HJ(1992) Corneal graft rejection: incidence, mani-festation, and interaction of clinical subtypes.Transplant Proc 24:2034–2037

40. Plsková J, Duncan L, Holáñ V, Filipec M, Kraal G,Forrester JV (2002) The immune response tocorneal allograft requires a site-specific draininglymph node. Transplantation 73:210–215

41. Price MO, Thompson RW Jr, Price FW (2003) Riskfactors for various causes of failure in initialcorneal grafts. Arch Ophthalmol 121:1087–1092

42. Rayner SA, King WJ, Comer RM, Isaacs JD, HaleG, George AJT, Larkin DFP (2000) Local bioactivetumour necrosis factor (TNF) in corneal allo-transplantation. Clin Exp Immunol 122:109–116

43. Roelen DL, van Beelen E, van Bree SPMJ, vanRood JJ, Völker-Dieben HJ, Claas FHJ (1995) Thepresence of activated donor HLA class I-reactiveT lymphocytes is associated with rejection ofcorneal grafts. Transplantation 59:1039–1042

44. Sagoo P, Chan GL, Larkin DFP, George AJT (2004)Inflammatory cytokines induce apoptosis ofcorneal endothelium through nitric oxide. InvestOphthalmol Vis Sci 45:3964–3973

45. Schulte F, Zhang E-P, Franke J, Ignatius R, Hoff-mann F (2004) Ipsilateral submandibular lym-phadenectomy does not prolong orthotopiccorneal graft survival in mice. Graefes Arch ClinExp Ophthalmol 242:152–157

46. She SC, Steahly LP, Moticka EJ (1990) A methodfor performing full-thickness, orthotopic, pene-trating keratoplasty in the mouse. OphthalmicSurg 21:781–785

47. Slegers TPAM, Broersma L,Van Rooijen N, Hooy-mans JMM, Van Rij G, Van Der Gaag R (2004)Macrophages play a role in the early phase ofcorneal allograft rejection in rats. Transplanta-tion 77:1641–1646

48. Sonoda Y, Streilein JW (1992) Orthotopic cornealtransplantation in mice – evidence that the im-munogenetic rules of rejection do not apply.Transplantation 54:694–704

49. Streilein JW (2003) New thoughts on the im-munology of corneal transplantation. Eye 17:943–948

50. Streilein JW, Arancibia-Caracamo C, Osawa H(2003) The role of minor histocompatibilityalloantigens in penetrating keratoplasty. DevOphthalmol 36:74–88

51. Strestikova P, Plsková J, Filipec M (2003) FK506and aminoguanidine suppress iNOS induction inorthotopic corneal allografts and prolong graftsurvival in mice. Nitric Oxide 9:111–117

52. Taylor AW (1999) Ocular immunosuppressive mi-croenvironment. In: Streilein JW (ed) Immuneresponses and the eye. Karger, Basel, pp 72–89


53. The Collaborative Corneal Transplantation Stud-ies (CCTS) (1992) Effectiveness of histocompati-bility matching in high-risk corneal transplanta-tion. Arch Ophthalmol 110:1392–1403

54. Vail A, Gore SM, Bradley BA, Easty DL, Rogers CA(1994) Corneal graft survival and visual outcome.A multicenter study. Ophthalmology 101:120–127

55. Vail A,Gore SM,Bradley BA,Easty DL,Rogers CA,Armitage WJ (1997) Conclusions of the cornealtransplant follow up study. Br J Ophthalmol81:631–636

56. Vail A,Gore SM,Bradley BA,Easty DL,Rogers CA,Armitage WJ, on behalf of collaborating surgeons(1997) Conclusions of the corneal transplant fol-low up study. Br J Ophthalmol 81:631–636

57. Völker-Dieben HJ, D’Amoro J, Kok-van AlphenCC (1987) Hierarchy of prognostic factors forcorneal allograft survival. Aust NZ J Ophthalmol15:11–18

58. Williams KA, Ash JK, Coster DJ (1985) Histocom-patibility antigen and passenger cell content ofnormal and diseased human cornea.Transplanta-tion 39:265–269

59. Williams KA, Coster DJ (1985) Penetratingcorneal transplantation in the inbred rat: a newmodel. Invest Ophthalmol Vis Sci 26:23–30

60. Williams KA, Muehlberg SM, Lewis RF, Coster DJ(1995) How successful is corneal transplantation– a report from the Australian Corneal GraftRegister. Eye 9:219–227

61. Williams KA, Muehlberg SM, Lewis RF, Coster DJ,on behalf of all contributors to the AustralianCorneal Graft Registry (1997) Long-term out-come in corneal allotransplantation. TransplantProc 29:983

62. Williams KA, Roder D, Esterman A, MuehlbergSM, Coster DJ, on behalf of all contributing sur-geons (1992) Factors predictive of corneal graftsurvival: report from the Australian CornealGraft Registry. Ophthalmology 99:403–414

63. Williams KA, White MA, Ash JK, Coster DJ (1989)Leukocytes in the graft bed associated withcorneal graft failure. Analysis by immunohistol-ogy and actuarial graft survival. Ophthalmology96:38–44

64. Yamada J, Kurimoto I, Streilein JW (1999) Role ofCD4+ T cells in immunobiology of orthotopiccorneal transplants in mice. Invest OphthalmolVis Sci 40:2614–2621

65. Yamagami S, Dana MR (2001) The critical role oflymph nodes in corneal alloimmunization andgraft rejection. Invest Ophthalmol Vis Sci 42:1293–1298

66. Young AL, Rao SK, Cheng LL, Wong AKK, LeungATS, Lam DSC (2002) Combined intravenouspulse methylprednisolone and oral cyclosporineA in the treatment of corneal graft rejection:5-year experience. Eye 16:304–308

67. Zhao JC, Jin XY (1995) Local therapy of corneal allograft-rejection with cyclosporine. Am J Oph-thalmol 119:189–194

References 81

7.1Introduction

Corneal avascularity is of paramount impor-tance in maintaining corneal transparency, thelatter being essential for good visual acuity.Therefore, in all higher animals depending ongood vision, the cornea normally is devoid ofblood and lymphatic vessels [12, 13, 21, 30]. Nev-ertheless, several diseases and surgical manipu-lations can lead to corneal (hem)angiogenesis(i.e., ingrowths of blood vessels from the limbalvascular arcade into the cornea) and lymphan-giogenesis (i.e., ingrowths of lymphatic vesselsfrom the limbal vascular arcade into the cornea[12, 13, 30]). Corneal hem- and lymphangiogen-esis can cause a significant reduction in visualacuity and blindness as well as render thesecorneas high risk in the case of a subsequentpenetrating keratoplasty [12, 13, 30]. In fact,corneal angiogenesis is associated with themost common cause of corneal blindnessworldwide (trachoma) as well as the most com-mon form of infectious blindness in westerncountries (herpetic keratitis [12, 13, 30]). Where-as the animal cornea has been used as in vivomodel to study the mechanisms of angiogenesisfor decades, the molecular pathways responsiblefor maintaining normal avascularity of the hu-man cornea (“angiogenic privilege”) have onlystarted to evolve in recent years [15]. The same istrue for the role of lymphatic vessels growinginto the cornea in inflammatory corneal dis-eases (lymphangiogenesis [3, 4, 11, 16]). Corneallymphangiogenesis has recently been shown tobe of essential importance in the induction ofimmune responses after corneal transplanta-tion, so that novel antihem- and antilymphan-

New Aspects of Angiogenesis in the Cornea

Claus Cursiefen, Friedrich E. Kruse

7

|

∑ Corneal angiogenesis is associated with themost common forms of corneal blindnessboth worldwide as well as in industrializedcountries

∑ Corneal angiogenesis is primarily caused by inflammatory diseases of the cornea(most commonly keratitis), corneal hypoxia(contact lens wear) and limbal antiangio-genic barrier defects (most commonlyaniridia, chemical burns)

∑ In corneal inflammation, (hem)angiogene-sis (i.e., outgrowth of pathologic blood vessels into the cornea) is usually accompa-nied by lymphangiogenesis (outgrowth oflymphatic vessels)

∑ Pathologic corneal lymphatic vessels are invisible at the slit-lamp, but can be visual-ized using specific immunohistochemicalmarkers in explanted vascularized corneas

∑ Preexisting blood and lymphatic vessels are strong risk factors for immune rejec-tions after keratoplasty

∑ In addition, about 50% of patients under-going low-risk keratoplasty also developcorneal angiogenesis postoperatively. In animal models of corneal transplantation,this postoperative mild combined hem-and lymphangiogenesis significantlyincreases the risk for immune rejections

∑ Novel antiangiogenic and antilymphangio-genic therapies can improve graft survivalafter both low-risk and high-risk keratoplas-ty by reducing the incidence of immunerejections (novel therapeutic concept)

∑ Novel, directly antiangiogenic drugs for application against corneal angiogenesiswill be available medium term as a spin-offof antiangiogenic cancer treatments

Core Messages

giogenic therapies are starting to emerge as newtools to improve graft survival in both the low-risk and the high-risk setting of corneal trans-plantation [3, 17].

7.2“Angiogenic Privilege of the Cornea”or “How Does the Normal Cornea MaintainIts Avascularity?”

Although the cornea – due to its anatomicallyexposed position – is in constant contact withnumerous minor inflammatory and thereby an-giogenic stimuli, the normal cornea remainsavascular [12, 13, 15]. Even after more severetrauma – such as refractive surgery – the corneain contrast to other tissues does not respond

with angiogenesis. This active maintenance ofcorneal avascularity has been termed “cornealangiogenic privilege” [12, 13, 15]. Corneal angio-genic privilege is not only essential for good vi-sual acuity but is also responsible for the excel-lent survival of corneal grafts placed intoavascular low-risk recipient beds, since in thesecases the graft is physically separated from boththe afferent (lymphatic) and the efferent (bloodvascular) arm of a so-called immune-reflex arc,leading to immune rejection after keratoplasty(Fig. 7.1A [3, 13, 17, 30]). Whereas the cornea hasserved as the in vivo model system for the studyof the mechanisms of angiogenesis for decades[in fact, as early as the 1940s Michaelson hy-pothesized the existence of a soluble angiogenicfactor mediating corneal angiogenesis (whichlater turned out to be primarily the angiogenic

84 Chapter 7 New Aspects of Angiogenesis in the Cornea

Fig. 7.1. A Schematic diagramof the so-called “immune reflexarc” leading to immune rejec-tions after keratoplasty. Theafferent arm consists of lym-phatic vessels allowing exit ofantigen presenting cells fromthe donor cornea, the centralprocessing unit is the regionalcervical lymph nodes [initiationof production of immune effec-tor cells], and the efferent armconsists of blood vessels allow-ing entry of immune effectorcells to the graft [with kindpermission from Streilein JW(1999) Immune responses andthe eye. Karger, New York, p 17].B Immune reflex arc in a vascu-larized high-risk cornea (withkind permission from [13]). Thedonor cornea has direct accessboth to the afferent lymphaticarm (1, 2) and thereby to the re-gional lymph node (3) as well asto the efferent blood vasculararm (4) of an immune-reflexarc. This explains the muchhigher rate of immune rejec-tions occurring in vascularizedhigh-risk eyes

B

A

growth factor VEGF)], the strategies used by thecornea for the normal maintenance of its avas-cularity are only partly understood [15, 19]. Ingeneral, angiogenesis or inhibition of angiogen-esis depend on a balance between proangio-genic factors (such as the VEGF family growthfactors VEGF-A, -C and -D) and antiangiogenicfactors (such as the thrombospondins; Table 7.1[1, 28]). If the balance tips more to proangio-genic factors, angiogenesis starts [1, 28]. In thecornea, normally the balance is shifted towardsantiangiogenic factors to maintain avascularity.Indeed, several antiangiogenic factors (such asPEDF, thrombospondins 1 and 2, antiangiogenicmatrix cleavage products such as angiostatinand endostatin, IL1RA) have already been iden-tified in the cornea [2, 12, 13]. It seems that theseantiangiogenic factors are strategically locatedat the inner and outer linings of the cornea (Descemet’s membrane and epithelial basementmembrane) to counteract angiogenic stimuliboth from inside (e.g., high concentrations ofangiogenic growth factors in the aqueous hu-mor during proliferative diabetic retinopathy)and from outside (e.g., against angiogenicgrowth factors from the tear film [2, 12, 13]).Animal experiments using mice deficient in one or more antiangiogenic factors (such asthrombospondins 1 and 2) have shown that thecorneal angiogenic privilege is redundantly or-ganized so that the absence of one or two factorsdoes not cause spontaneous ingrowths of limbalblood vessels [15]. This is in contrast to otherintraocular tissues such as the iris, where theabsence of these factors causes increased vascularity [15]. This demonstrates that evolu-tionarily the cornea has acquired a robust and redundant antiangiogenic system normallymaintaining avascularity unless it is overrun byoverwhelmingly strong (usually inflammatory/

infectious) stimuli for angiogenesis whichthreaten the integrity of the whole eye or eventhe whole body [15].


∑ Cornea and cartilage are the only avasculartissues of the human body

∑ Corneal avascularity is actively maintained,e.g., after refractive surgery (“cornealangiogenic privilege”)

∑ Corneal angiogenesis is associated with andpotentially causative of the most commoncauses of corneal blindness worldwide (trachoma) and the most common form ofinfectious corneal blindness in industrial-ized countries (herpetic keratitis)

7.3Corneal (Hem)angiogenesis

7.3.1General Mechanisms of Corneal (Hem)angiogenesis

According to Folkman, a balance between an-giogenic and antiangiogenic factors in each tis-sue and situation determines whether angio-genesis occurs or not. If the balance is tippedtowards angiogenic growth factors, vessel out-growth starts (“angiogenic switch”), whereas ifinhibitors prevail, angiogenesis is prohibited.Several angiogenic growth factors [primarilygrowth factors from the VEGF family (VEGF-A,VEGF-C, VEGF-D), FGF, IL-1, etc.] as well as in-hibitors of angiogenesis have been identified inrecent years (PEDF, thrombospondins, angio-statin, endostatin, etc.; see Table 7.1 [1, 28]).Pathologic angiogenesis (to clearly separate thisprocess from lymphangiogenesis, we will subse-

7.3 Corneal (Hem)angiogenesis 85

Table 7.1. Angiogenic and antiangiogenic factors (selection)

Angiogenic growth factors Antiangiogenic factors

VEGF (VEGF-A, VEGF-C, VEGF-D) Thrombospondins (TSP 1 and TSP 2)

FGF (bFGF, aFGF) PEDF

Interleukin-1 Angiostatin

TGF (alpha, beta) Endostatin

quently refer to it as “hemangiogenesis”) andlymphangiogenesis into the cornea mainly oc-cur in settings of an inflammatory “insult” tothe cornea, corneal hypoxia or limbal barrierdefects, all overriding the angiogenic privilegeof the cornea, which is actively maintained [2,12, 13, 15]. Clinical conditions most commonlyassociated with corneal neovascularization in-clude keratitis (herpetic and bacterial in na-ture), contact lens wear as well as inherited oracquired limbal deficiency states (primarilychemical burns [2, 12, 13, 15]). Growth factors ofthe VEGF family have been identified as keyplayers in both inflammation-driven hem- andlymphangiogenesis into the normally avascularcornea [7, 12, 13].

Release of angiogenic growth factors gener-ally is induced primarily by two factors: (1)inflammation and inflammatory cytokines (atthe cornea, e.g., keratitis) and (2) hypoxia (atthe cornea, e.g., contact-lens induced). The gen-eral process of sprouting angiogenesis followsthe following steps: (1) vasodilatation, (2) degra-dation of extracellular matrix, (3) mitotic acti-vation of endothelial cells and (4) chemotacticmigration of endothelial cells out of the pre-existing vessels towards an angiogenic stimulus.

The precise mechanisms whereby the sharplimbal border between hem- and lymphvascu-larized conjunctiva and avascular cornea ismaintained are unclear. It is possible that limbalstem cells located at niches in that area con-tribute to the antiangiogenic barrier of the lim-bus. Clinically the fact that destruction of limbalstem cells, e.g., by cautery, causes destruction

also of the limbal antiangiogenic barrier sup-ports this concept.

7.3.2Common Causes of Corneal (Hem)angiogenesis

The most common causes for corneal angiogen-esis are enlisted in Table 7.2 [6, 24]. According to the mechanisms leading to angiogenesis outlined above, these corneal diseases fall intothree general categories: (1) diseases leading tostrong inflammation within the cornea (auto-immune or infectious; most commonly herpet-ic keratitis); (2) diseases leading to hypoxiawithin the cornea (most commonly contactlenses with low Dk values) and (3) diseases withinherited (e.g., aniridia) or acquired (e.g., afterchemical cautery) defects of the limbal “anti-angiogenic” barrier [2]. In addition, “second-ary” corneal angiogenesis can occur after surgical manipulations at the cornea, which pri-marily involve placement of corneal sutures(e.g., after corneal wound repair, after cornealtransplantation, after block excision [9]).

7.3.3Clinical Consequences of Corneal Hemangiogenesis

Corneal angiogenesis can lead to reduced visu-al acuity not only by the physical presence ofblood vessels itself, but also due to leakage of


Table 7.2. Common causes of corneal angiogenesis

Pathomechanism Diseases/conditions

1. Corneal inflammation/infection Keratitis (most commonly herpetic keratitis,but also bacterial and fungal)Graft rejectionAutoimmune diseases affecting the cornea/sclera

2. Corneal hypoxia Contact lenses with low Dk values (especially extended wear)

3. Defects of the limbal antiangiogenic barrier Inherited defects (e.g., aniridia)Acquired limbal defects (e.g., after chemical cautery)

4. Secondary/iatrogenic After keratoplastyAfter corneal wound repair

products from immature corneal blood vessels(Fig. 7.2). This includes corneal edema due towater leakage, corneal lipid keratopathy due tolipid leakage and intrastromal or subepithelialhemorrhage (e.g., in contact lens patients). Fur-

thermore, as outlined in Sects. 7.3.4 and 7.4.2,corneal angiogenesis impairs the prognosis ofcorneal grafts placed into vascularized high-risk corneas. In fact, the Collaborative CornealTransplantation Study [24] (and numerousother clinical and experimental studies [30]) revealed preexisting corneal blood vessels asthe strong(est) risk factor for subsequent im-mune rejections.


∑ Corneal angiogenesis starts when the balance between proangiogenic and anti-angiogenic factors in the cornea is shiftedtowards angiogenic growth factors

∑ The most common clinical conditions associated with corneal angiogenesis arecorneal inflammation (keratitis), hypoxia(contact lens) and limbal barrier defects(chemical burns)

∑ Corneal angiogenesis leads to reduced visual acuity by the physical presence ofvessels itself, but also by leakage of water,lipids and erythrocytes

∑ Preexisting corneal blood vessels are astrong risk factor for subsequent immunerejections after keratoplasty

7.3.4Corneal Hemangiogenesis After Keratoplasty

Preexisting corneal blood vessels – as men-tioned above – have long been identified asstrong risk factors for immune rejection afterkeratoplasty [24]. But, until very recently therole of the mild angiogenesis (and – as we willdiscuss in Sect. 7.4.2 – parallel lymphangiogen-esis) occurring after keratoplasty in preopera-tively avascular recipient beds was unclear[8–10].

7.3.4.1Corneal Hemangiogenesis After Low-Risk Keratoplasty

The exact pathomechanism for postkeratoplas-ty neovascularization is unknown. The interac-tion of suture material with corneal epithelium/


Fig. 7.2 A–C. Complications of corneal angiogenesisleading to reduced visual acuity: A lipid keratopathy(arrows); B intrastromal hemorrhage; C secondarystromal edema due to leakage from immature bloodvessels. In addition, corneal angiogenesis makes sucha cornea a high-risk recipient bed in the case of sub-sequent keratoplasty (with kind permission from [12])

A

B

C

stroma as well as wound healing processes inthe interface seem to be important. Indeed,comparing corneal neovascularization withinthe first postoperative year between patientshaving undergone mechanical (more intensewound healing) versus nonmechanical (ex-cimer laser: less wound healing) keratoplastyshowed that the incidence of corneal angiogen-esis was lower in the nonmechanical (48%)compared to the mechanical trephination group(75%; p<0.01 [10]). This indicates that in thelow-risk setting, development of postoperativecorneal neovascularization seems to be affectedby the trephination technique and subsequentwound-healing response. Support for this con-cept comes from experimental data where post-operative corneal hem- and lymphangiogenesiswithin the first week do not differ between allo-geneic and syngeneic grafts in the mouse mod-el of corneal transplantation [17]. Since in the syngeneic model by definition there is noimmune response possible because donor andhost are immunologically identical, postopera-tive hem- and lymphangiogenesis really seem to be triggered by surgery and wound healingresponses [17]. This establishes surgery itselfand the degree of wound healing after kerato-plasty as novel risk factors for the induction ofimmune responses after keratoplasty. Conven-tional steroid therapy is not able to stop orprevent this postoperative angiogenesis suffi-

ciently, further underlining the need for morespecific antiangiogenic therapies (see Sect. 7.5.2[9, 10]).

A retrospective semiquantitative analysis of136 patients having undergone low-risk kerato-plasty (primarily patients with keratokonus andFuchs’ dystrophy) revealed that more than 50%of patients after low-risk keratoplasty (withpreoperatively avascular corneas) postopera-tively develop corneal angiogenesis within thefirst year [8, 9]. New vessels are primarily locat-ed in the 6 o’clock and 12 o’clock positions andtend to grow towards the outer suture turningpoints (Fig. 7.3). Thereafter, capillaries usuallyfollow the suture track towards the interface [9].In about 10% of patients these new vessels actu-ally reach donor tissue. Risk factors for postop-erative neovascularization include: embeddingof the suture knots in the host stroma, activeblepharitis, and a large recipient bed. Experi-ments in the mouse model of low-risk kerato-plasty have recently shown that these capillariesare always accompanied by biomicroscopicallyinvisible lymph vessels (lymphangiogenesis;Fig. 7.4; see Sect. 7.4.2 [16, 17]). Therefore, even ifit clinically appears only mild, this combinedangiogenesis and lymphangiogenesis after low-risk keratoplasty provides access for both arms(afferent lymphatic as well as efferent blood vas-cular) of an immune reflex arc towards the graft(Fig. 7.1). Indeed, experiments in the mouse


Fig. 7.3 A, B. Secondarycorneal angiogenesis after low-risk keratoplasty (A withkind permission from [12]).New capillaries develop in everysecond patient, are usually oriented towards the outer suture turning point and thengrow centripetally along thesuture track (arrows). In 10% ofpatients they reach the donor-host junction or grow furtherinto donor tissue. The mostcommon location of postkerato-plasty angiogenesis is at the6 o’clock and 12 o’clock posi-tionsA B

model of low-risk keratoplasty recently identi-fied postkeratoplasty neovascularization as arisk factor for subsequent immune rejections[16, 17]. An antihem- and antilymphangiogenictherapy significantly improved graft survivalafter low-risk keratoplasty (Fig. 7.5). Studies are under way to evaluate whether this alsoholds true for the human low-risk keratoplastysetting.

7.3.4.2Corneal Hemangiogenesis After High-Risk Keratoplasty

Even after high-risk keratoplasty, preexistingblood vessels tend to increase (Cursiefen et al.,unpublished observation, 2004). Only afterkeratoplasty for herpetic keratitis does anecdot-al evidence suggest that removal of the angio-


Fig. 7.4 A–I. The mouse model of low-risk kerato-plasty demonstrates early and parallel outgrowths ofboth blood and lymphatic vessels after low-risk ker-atoplasty (left column slit-lamp aspect,middle columncorneal whole mounts, right column detail showinglimbus at left and interface at right; blood vesselsstained in green, lymphatic vessels stained in red;

with kind permission from [17]). Whereas there areno blood or lymphatic vessels immediately postoper-atively in the low-risk recipient corneal bed, after3 days both vessel types clearly grow out from the lim-bal arcade and at day 7 in this model reach the host-graft interface

A B C

D E F

G H I

genic stimulus leads to a reduction in cornealangiogenesis [33]. Animal experiments recentlyclearly demonstrated that even after high-riskkeratoplasty there is a significant further in-crease in both hem- and lymphangiogenesis. Inaddition, inhibition of these processes evenafter high-risk keratoplasty (in the mouse mod-el) could improve subsequent graft survival(Cursiefen et al., submitted).


∑ Corneal angiogenesis postoperatively occurs in about 50 % of patients afterlow-risk keratoplasty in preoperativelyavascular recipient beds

∑ Postoperative angiogenesis reaches donortissue in more than 10 % of patients

∑ Animal experiments suggest that angiogen-esis after keratoplasty is accompanied byclinically invisible lymphangiogenesis

∑ Postoperative hem- and lymphangiogenesishave been identified as risk factors forimmune rejection after keratoplasty(mouse model)

∑ Inhibition of postkeratoplasty angiogenesisand lymphangiogenesis seem to improvegraft survival both in the low-risk andhigh-risk setting (mouse experiments)

∑ Surgery itself and the degree of woundhealing after keratoplasty are novel risk factors for the induction of immune responses after keratoplasty

7.3.5Corneal Angiogenesis Due to Contact Lens Wear

Prevalence of contact lens-associated cornealangiogenesis varies widely in the literature, butis generally estimated to be within a range of11–23% of contact lens wearers. The intensityalso can vary from some small capillaries usual-ly at the 6 o’clock and 12 o’clock positions todeep stromal mature blood vessels with second-ary scar formation. The cause seems to be a re-duced oxygenation of corneal epithelium andstroma, not primarily due to reduced diffusionthrough the contact lens, but due to a reducedexchange of the sub-lens tear film which leadsto reduced oxygenation of this tear film by lidcapillaries and thereby corneal hypoxia. This inturn leads to upregulation of proinflammatorycytokines, VEGF-A and then angiogenesis[5–7]. Since even the mild hem- and lymphan-giogenesis occurring after low-risk keratoplastyhave been identified as risk factors for immunerejections after keratoplasty – in analogy – care-ful attention should be given to contact lens-in-duced corneal angiogenesis occurring in kera-toconus patients, since these patients might goon to keratoplasty and one might create a high-er-risk scenario in the case of later keratoplasty[16, 17].


Fig. 7.5. Inhibition of heman-giogenesis and lymphangio-genesis after low-risk kerato-plasty using a VEGF-A specificcytokine trap (VEGF Trap) inthe mouse model of cornealtransplantation significantlyreduces immunological graftrejections (p<0.05; with kindpermission from [17])

7.3.6Angiogenesis as a Cause of Disease Progression, not a Sequel(Herpetic Keratitis)

Corneal angiogenesis not only can follow in-flammatory and infectious diseases of thecornea, but may also be pathogenetically rele-vant for the induction of certain corneal dis-eases. Recent work by Rouse and coworkersdemonstrated that inhibition of angiogenesis inanimal models of herpetic keratitis could pre-vent or diminish the intensity of herpetic stro-mal keratitis [33]. This suggests that efferentblood vessels may be essential in the pathogen-esis of stromal herpetic keratitis by providingCD4+ lymphocytes an entry site into thecorneal stroma. Novel emerging antiangiogenictherapies (see Sect. 7.5.2) may become part ofthe pharmacologic armamentarium to treat orprevent herpetic stromal keratitis [33].

7.3.7Surgery in Vascularized Corneas

Surgery in vascularized corneas necessitatesspecial approaches. Whereas refractive surgeryin heavily vascularized eyes cannot be recom-mended since intraoperative bleeding causeschanges in the ablation profile, LASIK, e.g., canbe performed in eyes with minor peripheralblood vessels in the cornea. Care should be tak-en not to cause bleeding and if so to carefullystop bleeding and keep the ablation zone free oferythrocytes.

For penetrating keratoplasties in heavilyvascularized corneas it may be advisable to preoperatively – functionally – occlude largervessels, e.g., using fine-needle diathermy [27].Alternative approaches are discussed below inSect. 1.5.2.


∑ Care should be taken with contact-lens in-duced corneal angiogenesis in keratoconuspatients, since that may compromise thesuccess of a subsequent keratoplasty due toincreased risk of immune rejections

∑ Angiogenesis seems to play a pathogenicrole in herpetic stromal keratitis; anti-angiogenic therapy should be helpful inthese patients

∑ Fine-needle diathermy is an easy to per-form, quick and cheap approach for thetemporary occlusion of larger cornealvessels prior to corneal surgery

7.4Corneal Lymphangiogenesis

Lymphangiogenesis, i.e., the development ofnew lymph vessels, has recently gained wide interest for its important role in tumor metasta-sis and induction of alloimmunity after or-gan transplantation [26]. Antilymphangiogenicstrategies have improved survival in animaltumor models by reducing tumor metastasis.Furthermore, antihem- and antilymphangio-genic strategies have improved graft survival af-ter organ transplantation in the mouse model ofcorneal transplantation (see below in Sect. 7.4.2[3, 17]). On the other hand, pro-lymphangio-genic treatment is desirable for patients withcongenital or acquired lymphedema.

7.4.1Mechanisms of CornealLymphangiogenesis

Whereas it has been known for more than100 years that the normally avascular corneacan be invaded by blood vessels (hemangiogen-esis), it was unclear until very recently whetherthe normally alymphatic human cornea couldbe invaded by lymphatic vessels from the lym-phatic arcade at the limbus (lymphangiogenesis[2, 12, 13]). The main reasons for that unclaritywere: (1) the fact that lymph vessels – in contrastto erythrocyte-filled blood vessels – are not de-tectable biomicroscopically using the normalslit-lamp magnification and (2) the lack of spe-cific markers for lymphatic endothelium. Thelatter has changed in the last 5–10 years with theadvent of several specific markers of lymphaticendothelium (such as LYVE-1, podoplanin andVEGF receptor 3 [26]). These novel markers

7.4 Corneal Lymphangiogenesis 91

enabled for the first time the precise identifica-tion of lymphatic vessels in vascularized humancorneas (Fig. 7.6 [11]). Lymphatic vessels weresignificantly more common in corneas with ashort history of corneal inflammation (usuallykeratitis or trauma) and also were significantlymore common in heavily vascularized corneas[11]. Therefore the chance of having both patho-logical blood and clinically invisible lymphaticvessels present is strongly correlated with thedegree of corneal angiogenesis, which can bejudged by slit-lamp evaluation. Furthermore,recent work suggests that it is possible todemonstrate lymphatic vessels in vivo in thecornea using confocal microscopy (HRT II us-ing the Rostock module). Since lymphatic ves-

sels are invisible at slit-lamp magnifications,they might not be as detrimental for cornealtransparency as blood vessels are. In fact, ani-mal experiments suggest that the “antilymph-angiogenic privilege” of the cornea is notredundantly organized.

Although previous studies especially byCollin and coworkers from the 1970s suggestedthe existence of lymphatic vessels in vascular-ized animal corneas [4], only very recently has itagain become possible unequivocally to identi-fy corneal lymphangiogenesis, e.g., in themouse model of corneal transplantation usingthe above-mentioned markers (Fig. 7.7 [3, 16,17]). Using the mouse model of corneal neovas-cularization, we were recently able to demon-


Fig. 7.6 A, B. Lymphatic vessels in vascularized hu-man corneas. A Immunohistochemistry with a novelmarker specific for lymphatic endothelium (LYVE-1)clearly separates blood vessels (stained here in green) from non-erythrocyte-filled lymphatic vessels(stained in red). B Electron microscopy reveals thelarge, non-erythrocyte-filled lumen of a thin-walledlymphatic vessel in a vascularized human cornea (toppanels). In contrast, erythrocyte-filled blood vesselshave a thick, multilayered basement membrane (lower panels; with kind permission from [11]; Lu lumen, EN endothelial cell, Pe pericyte, ECM extracel-lular matrix)

A

B

strate that after an inflammatory stimulus to thecornea, there is usually parallel and very early(within 48 h) outgrowth of both blood and lym-phatic vessels. Both originate from the limbalvascular arcade (Fig. 7.7 [16]). The cornea there-fore is also an excellent model system withwhich to study the mechanisms not only of an-giogenesis but also lymphangiogenesis and testpharmacologic compounds for the relative inhi-bition of both processes in the animal model[19]. Compared to blood vessels, lymphatic ves-sel tend to regress much more quickly and morecompletely after an inflammatory challenge tothe cornea [18]. For example, after a short,2-week-long inflammatory stimulus (cornealsutures), all lymphatic vessels in the mousecornea are completely regressed after 6 months,whereas blood vessels persist (partly as non-perfused ghost vessels) indefinitely. As outlined

below, this supports the clinical practice of notperforming penetrating keratoplasties in fresh-ly inflamed eyes, but of waiting until inflamma-tion has calmed down to improve graft survival[18]. Lymphangiogenesis is mediated by theVEGF family growth factors VEGF-A, -C and -Das well as by FGF and PDGF [26]. Stimuli for therelease of the main lymphangiogenic growthfactor VEGF-C are primarily inflammatory innature, explaining the clinical observation thathuman corneal lymphangiogenesis is morecommon shortly after keratitis [11, 26].


∑ During corneal inflammation there are parallel outgrowths of both blood and lymphatic vessels from the limbus into thecornea (combined hemangiogenesis andlymphangiogenesis)

7.4 Corneal Lymphangiogenesis 93

Fig. 7.7 A–D. Pathologic corneal blood vessels(stained in green with CD31) and lymphatic vessels(stained in red with LYVE-1) originate from the lim-bal arcade (A). Time course of parallel outgrowth of

blood and lymphatic vessels after an inflammatorystimulus (suture) in segments from a corneal wholemount (limbus at bottom, central cornea with sutureat the top; B blood vessel, L lymphatic vessel)

A B

DC

∑ Novel immunohistochemical studies pro-vide unequivocal evidence for lymphangio-genesis in vascularized human corneas,although lymphatic vessels are not visibleusing slit-lamp magnifications in vivo

∑ Lymphatic vessels are more common inheavily vascularized human corneas andare more common shortly after a cornealinflammation (keratoplasty, keratitis,immune rejection, etc.)

∑ In vivo confocal microscopy seems to be atechnique for the visualization of lymphaticvessels in vivo in the cornea

7.4.2Importance of Lymphangiogenesis for Induction of Alloimmunity After Keratoplasty

Normal corneas lack both blood and lymphaticvessels. This corneal avascularity is not only es-sential for corneal transparency but also con-tributes to the enhanced prognosis of low-riskkeratoplasty compared to other solid organtransplantation by suppressing both “arms” of apotential “immune reflex” that could lead totransplant rejection after keratoplasty [29, 30].However, both secondary to a variety of diseasesand after surgical manipulations, the corneacan be invaded by new blood and lymphaticvessels outgrowing from limbal blood vessels, asoutlined above. This implies that, e.g., in vascu-larized high-risk corneas after keratoplasty thegraft has direct contact both to the blood and tothe lymphatic system (Fig. 7.1B). Whereas theblood vessels provide a route of entry for im-mune effector cells (CD4+ alloreactive T-lym-phocytes, macrophages, etc.), corneal lymphaticvessels provide a drainage pathway for bothantigenic material (cells, cellular debris) andantigen-presenting cells (APCs) from the graftto the regional lymph node. In addition, im-munomodulatory cytokines, present in high-risk beds, or induced after surgical manipula-tion of the graft, could travel to the lymph nodeas could memory T cells and hyaluronic acid(HA) breakdown products that are known toactivate dendritic cells [12, 13, 29]. Besides en-abling the transport itself, lymphatic vessels

also enhance speed and amount of antigenicmaterial or APCs traveling to the regionallymph node. This induces alloimmunization atthe lymph node and production of alloreactiveeffector cells, which then travel via the efferentblood vascular arm to the donor cornea andinduce graft rejection (Fig. 7.1B).

The relative importance of lymphatic vessels(representing an exit route for APCs) versusblood vessels (representing an entry route foreffector cells) in vascularized corneas in rela-tion to graft rejections is not fully understood.But since clinically detectable corneal bloodvessels are neither necessary nor sufficient forimmune rejection of experimental cornealgrafts, an important role of the afferent lym-phatic pathway mediated by lymphatic vessels islikely [20, 25, 30–32]. The relatively higher im-portance of the afferent lymphatic arm of theimmune response in dictating the outcome ofcorneal graft survival has recently been demon-strated by several elegant studies: Indefinitesurvival of both fully mismatched orthotopicnon-high-risk grafts [31] and 90% survival offully mismatched high-risk corneas [32] inBALB/c mice was achieved by removal of cervi-cal lymph nodes by cervical lymphadenectomy.Furthermore, pharmacologic strategies inhi-biting (angiogenesis and) lymphangiogenesisafter low-risk keratoplasty (see Sect. 7.5.2) andeven after high-risk keratoplasty can signifi-cantly improve corneal graft survival. In sum-mary, interference with both the afferent lym-phatic and the efferent blood vascular arm of animmune reflex arc after both low-risk and high-risk keratoplasty is a novel and interesting strat-egy to improve graft survival. All this supportsthe novel concept that antiangiogenic therapiescan modulate immune responses after kerato-plasty and thereby improve graft survival.

7.4.3Non-immunological Effects of Corneal Lymphangiogenesis

Whereas the normal human cornea is devoid of HA, upregulation of HA expression in allcorneal layers can consistently be observed ininflammatory corneal diseases, after trauma


and keratoplasty. Increased amounts of HA inthe cornea, e.g., after refractive procedures, areassociated with reduced corneal transparency(“haze”). This might suggest that HA causeslocal shifts in water content in the cornealwound and thereby also local shifts in trans-parency due to interference with the fine-tunedspacing of corneal collagen fibrils. In extraocu-lar tissues, most of the inflammation-associatedHA deposited is transported to and metabo-lized in regional lymph nodes and the liver.LYVE-1, one of the specific lymphendothelialmarkers, is an HA receptor, and is thought tomediate HA uptake into lymphatic vessels andto facilitate transport to regional lymph nodes.Since LYVE-1 is expressed on corneal lymphvessels, these could be involved in transport ofpathologic corneal HA from the cornea to re-gional lymph nodes. Lymphangiogenesis intothe cornea in this setting might be beneficial forremoval of surplus HA, which would otherwiseinterfere with corneal transparency [12].


∑ Since lymphatic vessels are the essentialand required afferent arm of immune rejections after keratoplasty, antilymph-angiogenic pharmacologic strategies canimprove graft survival

∑ New concept: Antiangiogenic therapymodulates immune responses

7.5Antiangiogenic Therapy at the Cornea

Antiangiogenic therapeutic approaches at thecornea can be broadly divided into three cate-gories [21–23]:1. Angiostatic/antiangiogenic, i.e., to stop the

outgrowths of new vessels (classical antian-giogenic approach)

2. Angioregressive (meaning regression of al-ready established pathologic vessels, which isespecially important, e.g., in prevascularizedhigh-risk eyes)

3. Angio-occlusive [meaning the (temporary)functional occlusion of blood vessels, usual-ly prior to corneal surgery]

All three approaches need to be modified ac-cording to whether only blood, only lymphaticor both vessel types are to be targeted. So far, nospecific antiangiogenic therapy for applicationto the cornea is available. But a lot of specific an-tiangiogenic drugs have already entered phaseII and III clinical trials in cancer and, e.g., AMDtreatment, so that there is a realistic chance thatas a “spin-off” of antitumor treatment, specificantihem- and antilymphangiogenic agents willbe available (preferably in topical formulations)for use at the cornea medium-term.

7.5.1Establishedand Novel Antiangiogenic Therapies

The drugs available so far to inhibit angiogene-sis in the cornea only have indirect antiangio-genic effects. Topical steroids and cyclosporin Asuppress the inflammatory component induc-ing angiogenesis. Since they have no strong di-rect antiangiogenic mechanism, their effect islimited [9]. But since there is so far no alterna-tive, they are still the mainstay of topical antian-giogenic treatment at the cornea. Amnioticmembrane transplantation and even amnioticmembrane supernatant have been shown toexhibit antiangiogenic effects.Whether this alsocovers an antilymphangiogenic effect is un-known. Much more attractive for inhibitingcorneal angiogenesis and lymphangiogenesisare direct antiangiogenic agents. Numerous ofthese have already entered phase II and IIIclinical trials for anticancer indications (seewww.cancer.gov/clinicaltrials/developments/anti-angio-table.html). The only antiangiogenicagent with FDA approval so far is Avastin(Genentech). Other candidates in trials includethe cytokine trap VEGF Trap (Regeneron),VEGF aptamers (Macugen, Eyetech), antiangio-genic steroids (anecortave acetate, Alcon) andmany more. We recently demonstrated a verypotent antiangiogenic effect of the VEGF-Acytokine trap (Regeneron) on both inflamma-tion-induced corneal hemangiogenesis andsurprisingly also on lymphangiogenesis(Fig. 7.8 [16]). This nearly complete inhibitionshows the much higher potency of direct

7.5 Antiangiogenic Therapy at the Cornea 95


Fig. 7.8 A–G. Profound inhibitory effect of theVEGF-A cytokine trap (VEGF Trap) on both in-flammation-induced corneal hemangiogenesis andlymphangiogenesis in the mouse model of suture-in-duced corneal neovascularization (with kind permis-

sion from [16]). Left pictures are controls, right pic-tures are VEGF trap-treated animals (blood vessels:green; lymphatic vessels: red). Note the nearly com-plete inhibition of both hem- and lymphangiogenesis

A B

C D

E F

antiangiogenic agents compared for,e.g.,steroids(although no formal comparison has been per-formed so far). Since all of these compounds arein trial for systemic applications in tumor andposterior eye diseases, some time will evolveuntil topical formulations for treatment at thecornea become available. Nevertheless, the dra-matic developments in the antiangiogenic can-cer field will definitively provide useful spin-offproducts for the anterior eye segment.

Angioregressive therapies would allow the re-gression of preformed pathologic corneal bloodvessels. Whereas the regression of novel, newlyoutgrown blood vessels (in the so-called “prun-ing phase” [14]) can be achieved by removal ofangiogenic agents such as VEGF,older and moremature and pericyte-covered vessels no longerdepend on angiogenic signaling [14]. Inductionof regression of these mature vessels is morecomplex, and would have to involve anti-VEGFstrategies combined with agonists of the vascu-lar endothelial TIE2 receptor (angiopoietin 2).The regression phase for immature vesselsagain is very short, so that, e.g., removal of aloose suture or a hypoxia-inducing contact lensneeds to be performed very early after the onsetof vessel outgrowths to cause regression of thenew vessels. Since the mechanisms responsiblefor the maintenance of lymphatic vessels arepoorly understood, so far no approach for the regression of lymphatic corneal vessels isknown. Fortunately, lymphatic vessels in thecornea seem to regress spontaneously after the(inflammatory) stimulus subsides [18].

Angio-occlusive approaches are useful, e.g.,to prevent intraoperative bleeding during ker-

atoplasty in vascularized high-risk eyes or tostop leakage into the cornea out of these bloodvessels. Besides the more experimental ap-proach of corneal photodynamic therapy, fine-needle diathermy is a reliable, cost-effective andquick treatment option. Corneal vessels areeither directly cautered or a suture needle isplaced into/next to the vessel and the needle tipis then cautered (Fig. 7.9 [27]).

7.5 Antiangiogenic Therapy at the Cornea 97

Fig. 7.8 G. (continued)

G

Fig. 7.9 A, B. Fine-needle diathermy is an easy andquick method for the functional occlusion of largercorneal stromal blood vessels, e.g., prior to kerato-plasty (with kind permission from [27]). Either thevessel is coagulated directly or a needle is passedthrough/adjacent to the vessel and the cautery ap-plied to the needle

A

B


∑ Antiangiogenic treatments fall into threecategories: angiostatic, angioregressive and angio-occlusive

∑ Mainstays so far are topical steroids andcyclosporin A eyedrops

∑ Novel (topical) antiangiogenic drugs willdramatically improve the potential toinhibit corneal angiogenesis effectively

7.5.2Novel Antihemangiogenic and Antilymphangiogenic Therapies to Improve Graft Survival After Keratoplasty

Using one of the novel direct antiangiogenicagents [the VEGF-A specific cytokine trap(VEGF Trap) from Regeneron], it was shown re-cently that inhibition of hemangiogenesis andlymphangiogenesis after low-risk keratoplastyimproves corneal graft survival significantly(Fig. 7.5 [17]). Furthermore, pharmacologicstrategies targeting the VEGF receptor 3, medi-ating primarily lymphangiogenesis and partlyhemangiogenesis, were also able significantly toimprove graft survival postkeratoplasty [3].This establishes postkeratoplasty neovascular-ization in the low-risk bed as a novel risk factorfor subsequent immune rejections and supportsthe novel concept of modulating immune re-sponses after corneal grafting by antihem- andantilymphangiogenic therapies.

References

1. Carmeliet P (2003) Angiogenesis in health anddisease. Nat Med 9:653–660

2. Chang JH, Gabison EE, Kato T et al. (2001)Corneal neovascularization. Curr Opin Ophthal-mol 12:242–249

3. Chen L, Hamrah P, Cursiefen C et al. (2004) Vas-cular endothelial growth factor receptor-3 (VEG-FR-3) mediates dendritic cell migration to lymphnodes and induction of immunity to cornealtransplants. Nature Med 10:813–815

4. Collin HB (1966) Endothelial cell lined lymphat-ics in the vascularized rabbit cornea. InvestOphthalmol 5:337–354

5. Conners MS, Stoltz RA, Davis KL et al. (1995) Aclosed eye contact lens model of corneal inflam-mation. Part 2: Inhibition of cytochrome P450arachidonic acid metabolism alleviates inflam-matory sequelae. Invest Ophthalmol Vis Sci 36:841–850

6. Cursiefen C, Küchle M, Naumann GOH (1998)Angiogenesis in corneal diseases: Histopathologyof 254 human corneal buttons with neovascular-ization. Cornea 17:611–613

7. Cursiefen C, Rummelt C, Küchle M (2000) Im-munohistochemical localization of VEGF, TGFaand TGFb1 in human corneas with neovascular-ization. Cornea 19: 526–533

8. Cursiefen C, Wenkel H, Langenbucher A et al.(2001) Standardisiertes Beurteilungsschema zursemiquantitativen Analyse der kornealen Neo-vaskularisation mittels projizierter Hornhaut-photographien: Pilotstudie zur Analyse derkornealen Neovaskularisation nach Nicht-Hoch-risiko-Keratoplastik vor anschließender Im-munreaktion. Klin Monatsbl Augenheilkd 218:484–491

9. Cursiefen C, Wenkel H, Martus P et al. (2001) Pe-ripheral corneal neovascularization after non-high-risk keratoplasty: influence of short- versuslongtime topical steroids. Graefes Arch Clin ExpOphthalmol 239:514–521

10. Cursiefen C, Martus P, Nguyen NX et al. (2002)Corneal neovascularization after nonmechanicalversus mechanical corneal trephination for non-high-risk keratoplasty. Cornea 21:648–652

11. Cursiefen C, Schlötzer-Schrehardt U, Küchle M etal. (2002) Lymphatic vessels in vascularized hu-man corneas: immunohistochemical investiga-tion using LYVE-1 and Podoplanin. Invest Oph-thalmol Vis Sci 43:2127–2135

12. Cursiefen C, Seitz B, Dana MR et al. (2003) Angio-genesis and lymphangiogenesis in the cornea:pathogenesis, clinic and treatment. Ophthal-mologe 100:292–299

13. Cursiefen C, Chen L, Dana MR et al. (2003)Corneal lymphangiogenesis: Evidence, mecha-nisms and implications for transplant immunol-ogy. Cornea 22:273–281

14. Cursiefen C, Rummelt C, Küchle M et al. (2003)Pericyte recruitment in human corneal angio-genesis. Br J Ophthalmol 87:101–106

15. Cursiefen C, Masli S, Ng TF et al. (2004) Roles ofthrombospondin 1 and 2 in regulating sponta-neous and induced angiogenesis in the corneaand iris. Invest Ophthalmol Vis Sci 45:1117–1124

16. Cursiefen C, Chen L, Borges L et al. (2004) Viabone marrow-derived macrophages,VEGF A me-diates lymph- and hemangiogenesis in inflam-matory neovascularization. J Clin Invest 113:1040–1050


17. Cursiefen C, Maruyama K, Liu Y et al. (2004) In-hibition of hemangiogenesis and lymphangio-genesis after normal-risk corneal transplantationby neutralizing VEGF promotes graft survival. In-vest Ophthalmol Vis Sci 45:2666–2673

18. Cursiefen C, Maruyama K, Jackson DG et al.(2005) Time-course of angiogenesis and lymph-angiogenesis after corneal inflammation. Cornea(in press)

19. Cursiefen C, Ikeda S, Nishina P et al. (2005) Spon-taneous corneal hem- and lymphangiogenesis inmice with destrin-mutation depend on VEGFR3-signaling. Am J Pathol 166:1366–1377

20. Küchle M, Cursiefen C, Nguyen NX et al. (2002)Risk factors for corneal allograft rejection: inter-mediate results of a prospective normal-risk ker-atoplasty study. Graefes Arch Clin Exp Ophthal-mol 240:580–584

21. Kruse FE, Volcker HE (1997) Stem cells, woundhealing, growth factors, and angiogenesis in thecornea. Curr Opin Ophthalmol 8:46–54

22. Kruse FE, Joussen AM, Rohrschneider K, BeckerMD, Volcker HE (1998) Thalidomide inhibitscorneal angiogenesis induced by vascular en-dothelial growth factor. Graefes Arch Clin ExpOphthalmol 236:461–466

23. Kruse FE, Cursiefen C, Seitz B et al. (2003) Klassi-fikation von Erkrankungen der Augenober-flaeche, part I. Ophthalmologe 100:899–915

24. Maguire MG, Stark WJ, Gottsch JD et al. (1994)Risk factors for corneal graft failure and rejectionin the collaborative corneal transplantation stud-ies. Collaborative Corneal Transplantation Stud-ies Research Group. Ophthalmology 101:1536–1547

25. Nguyen NX, Seitz B, Langenbucher A et al. (2004)Clinical aspects and treatment of immunologicalendothelial graft rejection following penetratingnormal-risk keratoplasty. Klin Monatsbl Augen-heilkd 221:467–472

26. Pepper MS, Tille JC, Nisato R et al. (2003) Lym-phangiogenesis and tumor metastasis. Cell Tis-sue Res 314:167–177

27. Pillai CT, Dua HS, Hossain P (2000) Fine needlediathermy occlusion of corneal vessels. InvestOphthalmol Vis Sci 41:2148–2153

28. Saharinen P, Tammela T, Karkkainen MJ et al.(2004) Lymphatic vasculature: development, mo-lecular regulation and role in tumor metastasisand inflammation. Trends Immunol 25:387–395

29. Streilein JW, Yamada J, Dana MR et al. (1999) An-terior chamber-associated immune deviation,ocular immune privilege, and orthotopic cornealallografts. Transplant Proc 31:1472–1475

30. Streilein JW (2003) Ocular immune privilege:therapeutic opportunities from an experiment ofnature. Nat Rev Immunol 3:879–889

31. Yamagami S, Dana MR (2001) The critical role oflymph nodes in corneal alloimmunization andgraft rejection. Invest Ophthalmol Vis Sci 42:1293–1298

32. Yamagami S, Dana MR, Tsuru T (2002) Draininglymph nodes play an essential role in alloimmu-nity generated in response to high-risk cornealtransplantation. Cornea 21:405–409

33. Zheng M, Schwarz MA, Lee S et al. (2001) Controlof stromal keratitis by inhibition of neovascular-ization. Am J Pathol 159:1021–1029

References 99

8.1Introduction

In this chapter, the recent evolvements in histo-compatibility matching for penetrating kerato-plasty are presented and a strategy for clinicalpractice is recommended.

8.1.1Immune Reactions Constantly ThreatenGraft Survival

Despite the anterior eye chamber immune priv-ilege, graft rejections are a major complicationof penetrating keratoplasty as they facilitatesubsequent graft failure. Application of topicalcorticosteroids for several months has com-monly been thought to be sufficiently protec-tive. On average, 18% of patients with normal-risk keratoplasty [13] and up to 75% in high-riskcases [15] nevertheless experience immunereactions. Life span of affected grafts is signifi-cantly reduced from an endothelial graft reac-tion. Immune reactions thus increase the inci-dence of re-keratoplasties due to failed grafts inthe long run.

Graft reactions can be reduced by means ofintensified and prolonged prophylaxis with top-ical corticosteroids and with systemic immuno-suppression [14]. The protective effect, however,ceases upon discontinuation of the immuno-suppressive regimen. Any long-term depend-ence on immunosuppression is associated with additional costs and potentially seriousside effects.

Antigen matching, that is avoiding graftsbearing antigens that are foreign to the recipi-ent upon allocation, improves histocompatibili-ty. Matching of human leukocyte antigens(“HLA Matching,” Sect. 8.1.2) and more recentlyof further antigen systems (“Minor Matching,”Sect. 8.1.3) has turned out to be a potent adjunctto immunosuppression in the fields of allogene-ic transplantation, i.e., in penetrating kerato-plasty.

Histocompatibility Matching in Penetrating Keratoplasty

Daniel Böhringer, Rainer Sundmacher, Thomas Reinhard

8

|

∑ HLA matching reduces graft rejections innormal- as well as in high-risk keratoplasty

∑ Most patients can be served with an HLAcompatible graft within well below a year,even on a monocenter waiting list

∑ Waiting time for a histocompatible graftcan be predicted and discussed with eachpatient in advance

∑ The HLAMatchmaker algorithm can balancewaiting time and histocompatibility for patients with rare HLA phenotype

∑ The HLA-A1/H-Y minor antigen equals immunogenicity of HLA mismatches:allocating male HLA-A1 positive donors for female recipients should be avoided

∑ Long-term graft survival will improve uponroutinely matching major and selected minor histocompatibility antigens.This strategy will outweigh the cost of HLA typing in the long run

Core Messages


∑ Unlike immunosuppression, HLA matchingcan permanently reduce graft rejections.

8.1.2Major Transplantation Antigens (HLA)

Experimental transfer of tumors from onemouse strain to others led to the discovery ofthe major histocompatibility complex (MHC).In humans, these antigens were termed humanleukocyte antigens (HLAs). Antibodies againstHLAs were first discovered after a transfusionreaction of a multiparous woman despite bloodgroup compatibility.

The HLA antigens are subdivided into threeclasses according to their expression patternand their association with other molecules.Only HLA genes of class I and II are relevant totransplantation medicine as they encode forpolymorphic membrane bound molecules.∑ Class I molecules are expressed on almost all

nucleated body cells. They comprise a heavychain, a light chain (b2-microglobulin) and asmall peptide of nine amino acid residues.Three gene loci of major importance havebeen identified for this system: HLA-A, HLA-B and HLA-C.

∑ Class II molecules are homodimers. Thesemolecules are exclusively expressed on cellfamilies that share the ability to present ex-tracellular antigens to the immune system(e.g., monocyte and lymphocyte derivedstrains). HLA class II loci of major immuno-logic importance are HLA-DR and HLA-DQ.

8.1.2.1Methods for Typing HLA Antigens

HLA antigens were discovered by means ofserological assays. In the beginning, only themacroagglutination assay was available. For thisassay, patient serum was mixed with test cellsbearing only the respective antigen.Any macro-scopically visible agglutination demonstratedthe presence of antibodies directed against therespective test cells. Depending on the availabil-ity of appropriate test cells, only a subset of all

HLA antibodies could be detected at that time.Today, the serologic HLA typing assay is per-formed using the complement-dependent cyto-toxicity assay (CDC). For HLA typing, cells areincubated against a battery of standardized an-tibodies as defined by the International Histo-compatibility Workshop (IHW). These IHW testsera define the HLA antigens that can be detect-ed from the assay. The assay is further incubat-ed with rabbit complement. Lysis is induced bycomplement activation in cells that were re-cognized by a specific antibody. For detection oflysis, a dye is eventually added.

The first generation of CDC test sera was un-able to detect differences of only a few aminoacid residues between certain HLA alleles. Uponavailability of monoclonal antibody techniques,the International Histocompatibility Workshopreleased a new generation set of test sera thateventually was able to split most alleles knownat that time (“broads”) into further HLA alleles(“splits”).

As the HLA nomenclature was already estab-lished at that time, the newly discovered alleleswere termed splits of the established broad anti-gens and sequentially assigned higher numbersthan the established broad antigens.

Further improvements in HLA typing wereachieved by means of molecular techniques asdiscovery of new alleles with these molecularmethods no longer depends on isolation ofviable cells. Using the polymerase chain reac-tion (PCR), small parts of the genome can bedetected by means of specific primers and anamplification reaction. This is achieved eitherby sequence specific oligonucleotide primingwhere alleles are identified by characteristic“fingerprint” sequences or by the more expen-sive sequence typing of the whole allele. Thesemolecular techniques led to further subdiffer-entiation of most split alleles.

For example, the broad antigen HLA DR3 is dividable into the split antigens DR17 andDR18, which in turn can be subdivided intoDRB1*0302 and DRB1*0303 at the molecularlevel.

In corneal transplantation, blood for HLAtyping is commonly collected up to 72 h afterthe donor’s death. From autolytic changes, a suf-ficient amount of viable cells is often unavail-

102 Chapter 8 Histocompatibility Matching in Penetrating Keratoplasty

able for serologic analysis. In this situation,molecular methods can still detect the HLAphenotype with excellent precision.


∑ Various HLA molecules are bound to thesurface of all nucleated cells. These antigensare potentially targeted by the immune system unless they are tolerated as of birth.One or two different versions (alleles) ofeach HLA locus can be produced by eachcell. Over 20 alleles have been identified foreach of the HLA loci.

∑ In penetrating keratoplasty, the alleles ofdonor and recipients should be typed withimmunogenetic techniques due to higheraccuracy and precision, especially in bloodsamples collected up to 72 h after clinicaldeath of the donor.

8.1.2.2HLA Matching in Penetrating Keratoplasty

In penetrating keratoplasty, contrary to otherfields of transplantation, the potentials of HLAmatching are currently mostly unexploited asthe beneficial effect was demonstrated only re-cently. This calls for explanation as the humancornea has been known for longer to expressHLA antigens and these antigens are known tar-gets of cytotoxic T cells in the process of graftrejection [12].

8.1.2.2.1Methodical Problems with Older Investigations

Numerous studies performed in the past2 decades came to contradictory results as to theusefulness of HLA matching [17]. Three short-comings of study design in these older investi-gations, however, compromised the power todemonstrate any matching effect:1. The major problem with almost all older

studies is the poor quality of HLA typing atthat time. The importance of highly accurateHLA typing for successful HLA matchingwas recently recognized. Even 5% of faultyHLA DR typing obscures the beneficial effectas demonstrated in a recent simulation

analysis [18]. The CCTS for example wasbased on typing data that differed by 55%from retyping with modern techniques, in-validating all conclusions drawn from thatparticular investigation.

2. Additionally, most studies suffered frompoor statistical power due to heterogeneityof the study groups. Multiple centers, lack ofstandardization regarding surgical experi-ence and keratoplasty procedure as well asdifferences in immunosuppression regimensmost likely also influenced outcome and thusconfounded or obscured the HLA effect.

3. Finally, most studies were performed onhigh-risk patients, who not only are at in-creased risk of immune reactions but arealso at risk of graft failure from events otherthan graft rejection. When correlating totalgraft failures with HLA matching, any statis-tical association might be obscured fromnon-immunological graft failures such asprotracted elevated intraocular pressure.

8.1.2.2.2Current Evidence

On the basis of modern and reliable HLA typingtechniques, recently four monocenter studiesfrom Europe were published. These well-de-signed investigations uniformly confirmed abeneficial effect of HLA compatibility in pene-trating keratoplasty [1, 13, 15, 18]. Each studystresses additional aspects with respect to HLAmatching as follows:∑ In 1,681 consecutive transplantations from

only one center, a benefit was found frommatching the class II locus HLA-DR addi-tionally to the class I loci A and B [18].

∑ A beneficial effect from class I matchingalone was only observed when matching is based on split rather than broad (seeSect. 8.1.2.1) HLA alleles [1].

∑ A beneficial effect was demonstrated evenfor normal risk (first keratoplasty for bullouskeratopathy, Fuchs’ endothelial dystrophy,keratoconus with centrally sutured graft oravascular corneal scars) keratoplasty alonewhen matching HLA A, B and DR broadalleles (Fig. 8.1) [13].

8.1 Introduction 103


∑ A mounting body of evidence supports the beneficial effects of HLA matching innormal- as well as in high-risk keratoplasty.

∑ Accuracy and precision of HLA typing arecrucial to HLA matching.

8.1.2.2.3Variable Immunogenicity of IndividualHLA Mismatches

HLA mismatches differ in strength of immuno-genicity and thus in deterioration of graft sur-vival. This has been observed for longer in kid-ney transplantation [6] and in keratoplasty [5].For HLA class I loci, the structural basis of thisphenomenon has recently been established [7, 8,9, 10]. This paved the way for predicting “accept-able” mismatches on an individual basis: theHLAMatchmaker algorithm defines nearly 50omnipresent epitopes within the molecularstructure of all HLA class I alleles. These epi-topes are thought to be particularly exposed tothe immune system and partitioned intotriplets of amino acid residues to account forthermodynamic characteristics of the antibodyrecognition reaction. All triplets are formally

concatenated to form the triplet-string for aparticular allele. The association of the triplet-string with a particular allele is exclusively de-fined for alleles at molecular typing resolution.

Degree of matching is assessed by countingall triplets of the donor’s triplet-strings that arenot identical to any of the correspondingtriplets of the recipient’s four triplet-strings(two for HLA-A and -B each).

HLA mismatches with zero to few mis-matched triplets are supposed to be fully histo-compatible with regard to the antibody epi-topes. Additionally, they are known not to causea deterioration of graft survival in kidney trans-plantation [10]. In penetrating keratoplasty,recently a beneficial effect of this algorithm wasdemonstrated (Fig. 8.2) [3].


∑ The HLAMatchmaker algorithm can help inreducing graft rejections for penetratingkeratoplasty to a similar extent as conven-tional HLA matching

∑ This algorithm is crucial for providing his-tocompatible grafts to patients with rareHLA phenotypes within a reasonable time.Alternatively, waiting time can be traded fora better match grade


Fig. 8.1. Beneficial effect fornormal risk keratoplasty alonewhen matching HLA A, B andDR broad alleles [13]

8.1.3Minor Transplantation Antigens

Graft rejections are observed even in HLA iden-tical allogeneic transplantations. These effectsare ascribed to disparities in minor histocom-patibility (H) antigens. A single minor H mis-match even exceeds the immunogenicity of asingle MHC mismatch in a mouse-model forhigh-risk keratoplasty.

H antigens are peptides derived from poly-morphic proteins. Their immunogenicity arisesas a result of their presentation on the plasmamembrane in the context of HLA class I or II,where they are recognized by alloreactive HLArestricted T cells. In animal models,antigen pre-senting cells (APCs) such as limbal Langerhanscells have been demonstrated to migrate fromthe graft to the host spleen via the camero-splenic axis. The spleen might thus be thesource of a cytotoxic specific immune responsedirected against foreign graft H antigens pre-sented by graft APCs.

8.1.3.1Selected Minor Antigens

8.1.3.1.1H-Y

Male grafts can be subject to alloimmune reac-tivity in female recipients, as antigens of the H-Y group are only expressed in male individu-als and not in females. H-Y antigens are sup-posed to occur in all tissues including thehuman cornea.

Epitopes of the H-Y antigen family are ex-pressed either in the context of HLA-A1 or HLA-A2. The HLA-A1/H-Y antigen is located in the Y-chromosome-encoded DFFRY protein, TheHLA-A*0201-restricted HLA-A2/H-Y antigencontains a post-translationally modified cys-teine that significantly affects T-cell recognition.

As to matching the HLA-A1/H-Y epitope, a20% reduction of graft rejections was recentlydemonstrated on 252 keratoplasties (manu-script in preparation), whereas matching of theHLA-A2/H-Y epitope did not affect graft sur-vival.

From this observation, male HLA-A1 posi-tive donors should not be allocated to femalerecipients. The prevalence of this setting is ashigh as 13% in the German keratoplasty popu-lation.

8.1.3.1.2HA-3

Another HLA-A1 restricted H antigen expressedin all corneal layers is the HA-3 epitope. Thisepitope is derived from the lymphoid blast cri-sis (Lbc) oncoprotein. Two alleles,VTEPGTAQY(HA-3T) and VMEPGTAQY (HA-3M), have been demonstrated. T-cell immune reactivityhas only been observed in the direction ofHA-3T. The prevalence of the immunogenic setting is as low as 3% in the German kerato-plasty population. HLA-A1/HA-3 matching canthus be thought of as being of slight importanceand HA-typing is not recommended for routineuse.


Fig. 8.2. Beneficial effect of the HLAMatchmakeralgorithm in penetrating keratoplasty [3]

8.1.3.1.3Blood Group Antigens

Blood group antigens are expressed on thecornea. In high risk situations, a significant re-duction of graft reactions after penetrating ker-atoplasty was observed in retrospective investi-gations [4, 11].

Future research holds the prospect ofdemonstrating an additional HLA restricted in-fluence of blood group antigens in normal risksituations as well.


∑ HLA-A1 positive grafts from male donorsshould not be allocated to female recipients

∑ Matching blood group antigens reducesgraft rejections in high-risk keratoplasty

∑ The rapidly evolving field of minor anti-gens is subject to ongoing and future research in penetrating keratoplasty

8.2Time on the Waiting List Associated with Histocompatibility Matching

8.2.1Waiting Time Variance Has Been a Barrier to Histocompatibility Matching

Histocompatibility matching is associated withadditional time on the waiting list from refusingall newly available grafts while waiting for the first that satisfies the histocompatibility re-quirements. Due to the social and individualcosts of blindness, waiting periods exceeding1 year are hardly reasonable.

This waiting period is highly variable, de-pending on the recipient’s histocompatibilityantigens: individuals with common HLA phe-notypes can be routinely provided a matchinggraft within a few months as prevalence of com-patible phenotypes is common in the donorpopulation as well. On the other hand, individu-als with a rare HLA phenotype commonly re-main on the waiting list for years without beingallocated a compatible graft. These patients,waiting in vain for an HLA match, contributed

to the long-standing reluctance towards HLAmatching in penetrating keratoplasty.

When these patients are identified in advance,a randomly assigned graft with appropriate im-munosuppression can be opted for a priori or thestringency with respect to histocompatibility canbe reduced,e.g.,using the HLAMatchmaker algo-rithm (Sect. 8.1.2.2.3). An algorithm for predict-ing the waiting period is thus vital for informedconsent on histocompatibility, which has to bediscussed with each patient individually. Thisproblem was solved recently with an algorithmthat can predict the expected time on the waitinglist on an individual basis.

8.2.2Algorithm for Predicting the Time on the Waiting List

An algorithm that is based on the HLA pheno-type, a database of the most common haplotypefrequencies in the donor population and lastbut not least parameters of the local corneabank can robustly predict the estimated waitingtime for an HLA compatible graft. This algo-rithm has been retrospectively validated againstan historical waiting list of almost 1,400 HLAtyped patients (Table 8.1) [2]. The assumptionsof this algorithm are summarized in the follow-ing two sections.

8.2.2.1Percentage of HLA Compatible Grafts

Twenty-seven different HLA phenotypes matchany recipient who is a heterozygote for the HLAloci A, B and DR. Only one phenotype, however,matches an individual who is completely homo-zygous at these loci. The total percentage of thedonor population matching the HLA phenotypeof any recipient is well approximated by the sumof all population frequencies (donor popula-tion) of the compatible HLA phenotypes. Thefrequency of any HLA phenotype is the productof both haplotype frequencies. Haplotype fre-quencies for the HLA loci A, B and DR can beretrieved from a common database comprisingthe respective donor population [16].


8.2.2.2Actual Estimation of the Waiting Time

The daily rate of new HLA compatible graftsequals the product of the daily rate of new HLAtyped grafts and the percentage of HLA com-patible donors as described in the previous sec-tion. Assuming a Poissonian distribution ofdonors, expected waiting period is reciprocal tothe daily rate of new HLA compatible grafts.The algorithm is summarized in Eq. 8.1.

(8.1)

where t [years] is expected waiting period, GF istotal share of compatible HLA phenotypes andGR is local daily rate of new donors.

A certain percentage of grafts are unsuitablefor transplantation due to quality control. Thiscan be adjusted using a local empiric constantfor each cornea bank.


∑ Most patients can be served with an HLAcompatible graft within well below a year,even on a monocenter waiting list

∑ Patients that waited in vain for an HLAmatch for a long time contribute to the reluctance towards HLA matching in penetrating keratoplasty

∑ Waiting time for a histocompatible graftcan be predicted from the HLA phenotypeand discussed with the patient in advance

tGR GF

=∑

1365

2

8.3Recommended Clinical Practice

According to current knowledge, HLA matchingshould be performed at least for the HLA loci A,B and DR.

All corneal grafts should be typed for theseHLA loci in order to increase the pool availablefor histocompatibility matching. Molecular typ-ing should be preferred over serologic methodsas molecular methods can still detect the HLAphenotype with excellent precision when bloodfor HLA typing is collected after up to 72 h post-mortem. An additional benefit of moleculartyping is the applicability of the HLAMatch-maker algorithm (Sect. 8.1.2.2.3).

As for the recipient, all patients should betyped upon the keratoplasty being indicated,again preferably with immunogenetic tech-niques for the HLAMatchmaker algorithm.

All patients awaiting normal risk keratoplas-ties should be told their expected time on thewaiting list (Sect. 8.2). Improved prognosis fromHLA compatibility should be weighed againstthe expected waiting time. This strategy needsto be discussed with each patient individually.

Patients awaiting high-risk keratoplastyshould be provided a histocompatible graft inalmost all cases. The HLAMatchmaker algo-rithm can be applied to balance time on thewaiting list with the degree of histocompatibili-ty that is realistically achievable in patients withrare HLA phenotype.

8.3 Recommended Clinical Practice 107

Table 8.1. Validation of algorithm against a historical waiting list of almost 1400 HLA typed patients [2]

Zero mismatches One mismatch Two mismatches

Predicted period (only of recipients for which a match was found below) 17±159 7±49 1±6

Simulated period 15±14 5±9 1±3

(29%) (71%) (83%)

R=0.28; p<0.001 R=0.36; p<0.001 R=0.45; p<0.001

Matching of HLA and additional antigen sys-tems (e.g., H-Y/HLA-A1 and blood group anti-gens), the HLAMatchmaker algorithm and wait-ing time prediction are only feasible with anintegrated highly specialized software packagefrom professionalized high-volume institutionsresponsible for allocation.

In summary, long-term graft survival willimprove upon routinely matching major andcertain minor histocompatibility antigens in all keratoplasties. This policy will outweigh thecosts of HLA typing in the long run.

References

1. Beekhuis WH, Bartels M, Doxiadis II, van Rij G(2003) Degree of compatibility for HLA-A and -Baffects outcome in high-risk corneal transplanta-tion. Dev Ophthalmol 36:12–21

2. Böhringer D, Reinhard T, Böhringer S, EnczmannJ, Godehard E, Sundmacher R (2002) Predictingtime on the waiting list for HLA matched cornealgrafts. Tissue Antigens 59:407–411

3. Böhringer D, Reinhard T, Duquesnoy RJ, Böh-ringer S, Enczmann J, Lange P et al. (2004) Bene-ficial effect of matching at the HLA-A and -Bamino-acid triplet level on rejection-free cleargraft survival in penetrating keratoplasty. Trans-plantation 77:417–421

4. Borderie VM, Lopez M, Vedie F, Laroche L (1997)ABO antigen blood-group compatibility in cornealtransplantation. Cornea 16:1–6

5. Creemers PC, Kahn D, Hill JC (1999) HLA-A and -B alleles in cornea donors as risk factors for graftrejection. Transpl Immunol 7:15–18

6. Doxiadis II, Smits JM, Schreuder GM, Persijn GG,van Houwelingen HC, van Rood JJ et al. (1996)Association between specific HLA combinationsand probability of kidney allograft loss: the tabooconcept. Lancet 348:850–853

7. Duquesnoy RJ (2002) HLAMatchmaker: a molec-ularly based algorithm for histocompatibility de-termination. I. Description of the algorithm. HumImmunol 63:339–352

8. Duquesnoy RJ, Howe J, Takemoto S (2003) HLA-matchmaker: a molecularly based algorithm forhistocompatibility determination. IV. An alterna-tive strategy to increase the number of compati-ble donors for highly sensitized patients. Trans-plantation 75:889–897

9. Duquesnoy RJ, Marrari M (2002) HLAMatch-maker: a molecularly based algorithm for histo-compatibility determination. II. Verification ofthe algorithm and determination of the relativeimmunogenicity of amino acid triplet-definedepitopes. Hum Immunol 63:353–363

10. Duquesnoy RJ, Takemoto S, de Lange P, DoxiadisII, Schreuder GM, Persijn GG et al. (2003) HLA-matchmaker: a molecularly based algorithm forhistocompatibility determination. III. Effect ofmatching at the HLA-A,B amino acid triplet levelon kidney transplant survival. Transplantation75:884–889

11. Inoue K, Tsuru T (1999) ABO antigen blood-group compatibility and allograft rejection incorneal transplantation. Acta Ophthalmol Scand77:495–499

12. Niederkorn JY (2001) Mechanisms of cornealgraft rejection: the sixth annual Thygeson Lec-ture, presented at the Ocular Microbiology andImmunology Group meeting, October 21, 2000.Cornea 20:675–679

13. Reinhard T, Böhringer D, Enczmann J, Kogler G,Mayweg S, Wernet P et al. (2004) Improvement ofgraft prognosis in penetrating normal-risk ker-atoplasty by HLA class I and II matching. Eye18:269–277

14. Reinhard T, Reis A, Kutkuhn B, Voiculescu A,Sundmacher R (1999) Mycophenolatmofetil nach perforierender Hochrisiko-Keratoplastik.Eine Pilotstudie. Klin Monatsbl Augenheilkd 215:201–202

15. Reinhard T, Spelsberg H, Henke L, Kontopoulos T,Enczmann J, Wernet P et al. (2004) Long-term re-sults of allogeneic penetrating limbo-keratoplas-ty in total limbal stem cell deficiency. Ophthal-mology 111:775–782

16. Schipper RF, Oudshoorn M, D’Amaro J, van derZanden HG, de Lange P, Bakker JT et al. (1996)Validation of large data sets, an essential pre-requisite for data analysis: an analytical survey ofthe Bone Marrow Donors Worldwide. TissueAntigens 47:169–178

17. Vail A,Gore SM,Bradley BA,Easty DL,Rogers CA,Armitage WJ (1997) Conclusions of the cornealtransplant follow up study. Collaborating Sur-geons. Br J Ophthalmol 81:631–636

18. Völker-Dieben HJ, Claas FH, Schreuder GM,Schipper RF, Pels E, Persijn GG et al. (2000) Beneficial effect of HLA-DR matching on thesurvival of corneal allografts. Transplantation 70:640–648


9.1Introduction

Immunologic graft rejection is the single mostimportant reason for graft failure followingcorneal transplantation. If corneal transplanta-tion is performed in a high-risk situation with-out the use of systemic immunosuppression,corneal graft failure can be expected in over50% of cases within the first postoperative year[15, 33].

Despite the advantage that the transplantedorgan can directly (and not via the vascular sys-tem) be reached with topical steroids in ex-tremely high concentrations, thereby interfer-ing with the host’s immune system right at the“battlefield” of graft rejection, this strategy isonly sufficient in a normal-risk situation.

The clonal expansion of alloreactive T cellsoccurs in lymphoid organs (i.e. lymph nodesand spleen): after the recognition of the foreigntissues by T cells, these specific T cells start toproliferate and generate an immunologicalarmy against the graft. It is therefore crucial notonly to work with topical steroids but to employimmunosuppressive substances systemically. Ina high-risk situation you have to fight the inhos-pitable host in its hinterland to achieve graft sur-vival in the long run.

To understand the possible targets of im-munosuppression and immunomodulation weneed to take a look at the underlying immunol-ogy.

9.2Immunology

Immunological responses against the trans-planted cornea remain the major cause of allo-graft injury and loss. The innate and adaptiveimmune systems are variously involved in rejec-tion. Several factors determine the strength andnature of the immune response: (1) the nature ofthe grafted cornea, i.e. whether it is a clearcorneal button or a limbocorneal transplant;and (2) the nature of the recipient’s graft bed(i.e. whether it is clear, vascularised or has alimbal stem cell insufficiency). Additionally,

Current Systemic Immunosuppressive Strategies in Penetrating Keratoplasty

Alexander Reis, Thomas Reinhard

9

|

∑ Immunologic rejection is the main cause of corneal graft failure

∑ Acute rejection is mainly mediated by Tcells and can be prevented with steroids,IL-2 inhibitors (cyclosporine, tacrolimus),mycophenolate mofetil and TOR inhibitors(everolimus, rapamycin)

∑ Based on their risk of immunologic rejec-tion, corneal transplants are rated as eithernormal-risk or high-risk transplants

∑ In a normal-risk situation the postoperativeapplication of topical steroids is sufficientto prevent acute graft rejection in most cases

∑ In high-risk keratoplasty systemic immuno-suppression with cyclosporine, myco-phenolate mofetil or tacrolimus has to be used to maintain clear graft survival

∑ As corneal transplantation is not a life-saving procedure, the side-effect profile is a central issue when choosing immuno-suppressive medication

Core Messages

inflammatory responses (and graft rejection is aform of inflammatory response) are physiolog-ically suppressed in the anterior chamber: onthe one hand, antigens injected intraocularlyelicit deviant systemic immune responses thatare devoid of immunogenic inflammation (aphenomenon called anterior chamber associat-ed immune deviation, ACAID). On the otherhand, the ocular microenvironment (aqueoushumor, secreted by cells that surround thischamber) suppresses intraocular expression ofimmunogenic inflammation [51].

These special anatomical features are re-sponsible for the excellent results in normal-risk corneal transplantation when compared tosolid organ transplantation or high-risk cornealtransplantation.

The nature of the host’s immune responsecan be determined by its histopathology andtime course as acute or chronic rejection.

9.2.1Acute Rejection

Acute rejection which may occur weeks to yearsafter transplantation involves both humoraland cell-mediated immune reactions. T cellsplay a central role in acute rejection by respond-ing to alloantigens, predominantly major histo-compatibility complex (MHC) molecules, pre-sented on endothelial, epithelial, or stromalcells. Both CD4+ and CD8+ T cells contribute toacute rejection. CD4+ T cells mediate acute re-jection by secreting cytokines and inducing de-layed-type hypersensitivity-like reactions in thegraft. Recognition and lyses of foreign cells bycytotoxic CD8+ T cells are an important mech-anism of acute rejection. T cells may be activat-ed by two distinct mechanisms: the direct andthe indirect pathway.

Based on the target, immune reactionsagainst the transplanted cornea may be dividedinto endothelial, stromal or epithelial rejection.The most frequent and most severe form of im-mune response is against the endothelium. Thereason is that the immunogenic epithelial cellsare replaced within approximately 1 year by thehost’s epithelium and the stroma mostly consistsof intracellular substance and only a small num-

ber of cells. Therefore the endothelium showsthe highest immunogenicity of a corneal graft.

9.2.2Major Histocompatibility Complex

9.2.2.1Direct Pathway of Allorecognition

Direct recognition of foreign MHC antigens byT cells is the primary cause of acute rejection:recipient T cells recognise donor MHC class Iand class II molecules, resulting in the genera-tion and clonal proliferation of helper and cyto-toxic T cells.

9.2.2.2Indirect Pathway of Allorecognition

This occurs when the MHC molecules of thedonor tissues are taken up and processed byantigen-presenting cells, which present the for-eign peptides to T cells. Since MHC moleculesare highly polymorphic in nature, they aremainly responsible for allograft rejection.Transplantation between individuals with iden-tical MHC molecules may also fail in the latephase because at this time the so-called minorhistocompatibility antigens come into play.

9.2.3Chronic Rejection

The pathogenesis of chronic rejection is notclear [3, 30]. It is likely that most of the adaptiveand innate immune systems are involved in thisprocess. Chronic rejection cannot be preventedwith current immunosuppressive drugs (whichmainly work through their interference with Tcells), so the present strategy is to limit thenumber of acute rejection episodes. The bestprospects for overcoming late graft loss due tochronic rejection may reside in a new genera-tion of immunosuppressive agents [28]. Manyrisk factors may increase the incidence ofchronic rejection: MHC incompatibility, thenumber and severity of acute rejection episodesand the recurrence of herpetic ocular disease.

110 Chapter 9 Current Systemic Immunosuppressive Strategies in Penetrating Keratoplasty

9.3Normal-Risk Versus High-RiskTransplantation

Based on their risk of graft rejection, cornealtransplants can be divided into normal-risk orhigh-risk transplants.

9.3.1Normal-Risk Transplantation

In a normal-risk situation (e.g. first transplantin keratokonus or Fuchs’s endothelial dystro-phy), a 5-month course of topical steroids (e.g.prednisolone acetate 1%, Inflanefran forte®) 5 times a day, reduced by one drop every month) accompanied by systemic steroids(prednisolone 1 mg/kg tapered within 3 weeks)is sufficient to maintain a 5-year clear graft sur-vival of up to 90%. Up to 20% of normal-riskcorneal transplants experience an acute rejec-tion episode which can be converted in about50% of cases with topical and systemic steroids.

9.3.2High-Risk Transplantation

Postoperative systemic immunosuppression iswidely accepted as the treatment of choice in immunologic high-risk groups. High-riskcorneal transplantation can be defined as fol-lows:∑ History of previous graft rejections∑ Deep vascularisation of the recipient cornea

in more than three quadrants∑ Limbal stem cell deficiency, which requires a

corneolimbal graft∑ Severe atopic dermatitis

In addition to topical and systemic applicationof steroids as mentioned previously, systemicimmunosuppression should be applied for atleast 6 months following transplantation.

9.3.3Rationale for Systemic Immunosuppression

The first goal of a timely limited systemic im-munomodulation is the prevention of acute re-jection episodes. The second goal is the interfer-ence with the initial graft-host interaction in away that graft-protective cells and cytokines arepromoted, hence enabling a clear graft survivalwithout any further medication. We havealready shown clinically that we can reach thefirst goal in most patients when using cyclo-sporine or mycophenolate mofetil systemically.Unfortunately, we still do not have convincingresults with our therapeutic strategies whenlooking at long-term graft survival.

9.3.4Why Is Immunomodulation with Topical Steroids Not Sufficient To Prevent Immunologic Graft Rejection in High-Risk Patients?

The cornea is a privileged place for transplanta-tion, for both its anatomical features (see above)and the possibility of bringing medication di-rectly to the transplanted organ, thereby reduc-ing systemic side effects. In a high-risk situationthe immunological privilege is diminished andthe risk of graft loss within 1 year lies over 50%without the use of systemic immunosup-pression [15, 33]. Why are topical steroids notenough?

The activation of the recipient’s immune sys-tem against the transplanted cornea, i.e. thepriming of naïve T cells, occurs in lymphoid tis-sues. This hypothesis is supported by experi-ments in which T-cell activation and thereforegraft rejection did not occur when secondarylymphoid organs were absent [18]. These exper-imental data indicate that leukocytes partici-pate in host T-cell priming by migrating fromthe graft to the host’s lymph node and/orspleen, where they activate alloreactive host Tcells in the direct and indirect pathway. Suchprimed T cells circulate and target MHC mole-cules expressed by cells of the graft.

9.3 Normal-Risk Versus High-Risk Transplantation 111

As topical steroids do not reach the second-ary lymphoid organs, and even systemicsteroids do not interfere sufficiently with theclonal expansion of activated T cells, it is essen-tial to administer systemic immunosuppres-sives in order to achieve clear graft survival.


∑ If corneal transplantation is performed in a high-risk situation without the use ofsystemic immunosuppression, corneal graftfailure can be expected in over 50 % of caseswithin the first postoperative year

∑ Definition of high-risk corneal transplan-tation:– History of previous graft rejections– Deep vascularisation of the recipient

cornea in more than three quadrants– Limbal stem cell deficiency which

requires a corneolimbal graft– Severe atopic dermatitis

∑ In a high-risk situation it is crucial not onlyto work with topical or systemic steroidsbut to employ immunosuppressive sub-stances systemically

∑ T cells play a central role in rejection byresponding to alloantigens, predominantlyMHC molecules, presented on endothelial,epithelial, or stromal cells

∑ The activation of the recipient’s immunesystem against the transplanted cornea,i.e. the priming of naïve T cells, occurs in lymphoid tissues

∑ As topical steroids do not reach the second-ary lymphoid organs, and even systemicsteroids do not interfere sufficiently withthe clonal expansion of activated T cells,it is essential to administer systemic immunosuppressives in order to achieveclear graft survival

∑ The first goal of a timely limited systemicimmunomodulation is the prevention of acute rejection episodes

∑ The second goal is the interference with theinitial graft-host interaction in a way thatgraft-protective cells and cytokines are pro-moted, hence enabling a clear graft survivalwithout any further medication

9.4Immunosuppressive Agents

9.4.1History

Along with the increase in the number of solidorgan transplants, our therapeutic armamen-tarium and knowledge of immunosuppressivedrugs in corneal transplantation has been im-proved.

In the 1950s the selection of immunosup-pressive drugs was limited to corticosteroidsand azathioprine. In the 1960s polyclonal anti-lymphocyte (ALG) and antithymocyte (ATG)globulins supplemented the repertoire. In thelate 1970s cyclosporine A led to a real break-through in clinical solid organ transplantation(Table 9.1). Motivated by the encouraging re-sults in graft survival, the research in this im-munological field then led us to a wide range of


Table 9.1. Immunosuppressives: history

1949 Cortisone was shown to alleviate rheumatoid arthritis

1959 Cyclophosphamide was demonstrated to suppress the formation of antibodiesand was used for bone marrowtransplantation

1960s Azathioprine was found to delay organgraft rejection

1969 Methotrexate was shown to inhibitantibody formation and the developmentof delayed hypersensitivity in guinea pigs

1976 T-cell-inhibiting properties ofcyclosporine were demonstrated

1982 Development of mycophenolate mofetil

1987 Tacrolimus (FK506) was shown to inhibit IL-2 production and lymphocyte proliferation

1980s Interest in the antibiotic sirolimus was renewed when it was shown to prevent allograft rejection

1978 Leflunomide

1981 Mizoribine, deoxyspergualin, brequinar

1990s IL-2 antagonists (daclizumab,basiliximab)

potent immunosuppressive agents with highlyspecific sites of action (Fig. 9.1).

According to their mode of action these newdrugs can be divided up into agents that selec-tively inhibit cytokine gene transcription/expression (cyclosporine, tacrolimus), anti-proliferative agents (mycophenolate mofetil,azathioprine) and agents that interfere withintracellular signal transduction (rapamycin,everolimus). Immunosuppressives might alsobe classified as biologics which are defined asnaturally occurring or genetically engineeredmammalian proteins (thymoglobulin, basilix-imab and daclizumab) or xenobiotics (drugs

produced from microorganisms, e.g. cyclo-sporine, tacrolimus) (Table 9.2).

Despite the tremendous breadth of the disci-pline of immunosuppressive molecules, only asmall number of drugs have made it as far as be-ing used for experimental or clinical cornealtransplantation. We have decided to focus onthe following agents:∑ Corticosteroids∑ Cyclosporine (Sandimmun, Neoral, CSA)∑ Tacrolimus (Prograf, FK506)∑ Mycophenolate mofetil

(CellCept, Myfortic, MMF)∑ RAD (Everolimus, Certican)

9.4 Immunosuppressive Agents 113

Table 9.2. Immunosuppressives: mode of action

Mode of action Substances

Regulators of gene expression Glucocorticoids, vitamin D analogs, deoxyspergualin

Alkylating agents Cyclophosphamide

Kinases and phosphatase inhibitors Cyclosporine, tacrolimus, everolimus, rapamycin

Inhibitors of de novo purine synthesis First generation: mercaptopurine, azathioprine,methotrexate

Second generation: mizoribine and MMF

Inhibitors of de novo pyrimidine synthesis Brequinar, leflunomide, malononitrilamides

Fig. 9.1. Immunosuppressives:sites of action

∑ Rapamycin (Sirolimus, Rapamune)∑ FTY720∑ Biological agents (basiliximab, daclizumab)

9.4.2Corticosteroids

Corticosteroids prevent interleukin (IL)-1 andIL-6 production by macrophages and inhibit allstages of T-cell activation. Adverse effects ofsystemic steroids include Cushing’s disease,bone disease (e.g. osteoporosis, avascularnecrosis), cataract, glucose intolerance, infec-tions, hyperlipidaemia, and growth retardation.Adverse effects of topically applied steroidsinclude cataract, glaucoma and in the case ofepithelial defects – steroid ulcers.

9.4.3Cyclosporine A (CSA, Sandimmun,Sandimmun Optoral, Sandimmun Neoral)

The fermentation product from the fungiTolypocladium inflatum Gams was first isolatedin 1970 by Thiele and Kis. Its immunosuppres-sive properties were discovered in 1972 by Borel.Sandimmun Neoral is a special galenic formula-tion based on microemulsion technology.

Cyclosporine A binds to the intracellular im-munophilin cyclophilin (immunophilins areproteins which bind to immunosuppressivedrugs). The CSA-cyclophilin complex blockscalcineurin-calmodulin-induced phosphoryla-tion of NFAT (nuclear factor of activated Tcells), transcription factor for IL-2 and otherearly T-cell specific genes (Fig. 9.1) and hence ishighly T cell specific.

Clinical efficacy and safety data have most-ly been acquired in solid organ transplanta-tion, and it is still the gold standard in all forms of solid organ transplantation (except liver transplantation) mainly in combinationwith steroids, azathioprine or mycophenolatemofetil.

CSA in Corneal Transplantation. The first doc-umented clinical experiences in corneal trans-plantation date back to the mid 1980s with the

exceptional efforts undertaken by Hill and col-leagues in South Africa [14, 15]. These initialpositive clinical experiences with systemic CSAto prevent corneal allograft rejection in highrisk keratoplasty have been confirmed by others[32, 33, 35].

Despite the significant improvement of out-come in high-risk keratoplasty, the use of CSA islimited due to its considerable toxicity and theneed for costly drug monitoring. The toxicity is mostly caused by the CSA-cyclophilin-cal-cineurin-calmodulin complex, which interfereswith tubular and endothelial cell functions:nephro- and hepatotoxicity, and alterations inglucose metabolism, hypertension, and gingivalhyperplasia. To avoid the systemic toxicity,attempts have been undertaken to apply CSA intopical formulations including the use of colla-gen carriers [6, 16]. The encouraging results ofthese mostly experimental studies in preventingcorneal graft rejection did not hold true clini-cally [29]. However, we have shown that topicalCSA is efficient in the treatment of distinctimmunological disorders of the cornea (e.g.Thygeson’s keratitis, persistent nummular infil-trates following adenovirus infections) [34, 36].

9.4.4Tacrolimus (FK506, Prograf)

Tacrolimus has been proven clinically superiorto CSA following solid organ transplantation [5,54]. Tacrolimus, like CSA, is a macrolide antibi-otic (structurally related to erythromycin andrapamycin) derived from a fungus, Strepto-myces tsukubaensis [17]. Its immunosuppressiveproperties were discovered by Ochai in 1985.In vitro studies have shown that, even in con-centrations 40–200 times lower than CSA,tacrolimus possesses extremely powerful im-munosuppressive effectiveness [45, 55]. Al-though the final step in modulating the immunesystem is the same for CSA and tacrolimus, i.e.interfering with the intracytoplasmatic cal-cineurin system and hence the interleukin IL-2production, both drugs manage this in a differ-ent manner. Tacrolimus binds to the intra-cellular FKBP-12 (FK-binding protein-12). Thetacrolimus-FKBP-12 complex blocks calcineurin-


calmodulin-induced phosphorylation of thecytoplasmic component of NFAT transcriptionfactor for IL-2 and other “early”genes. Like CSA,tacrolimus is a highly specific inhibitor of lym-phocyte activation. Its toxicities are similar toCSA (probably due to its calcineurin-mediatedinterference with tubular and endothelial cells),i.e. nephro-, neurotoxicity, arterial hyperten-sion, diabetogenicity.

Tacrolimus in Corneal Transplantation. Up tonow there have only been limited clinical dataavailable about the efficacy of tacrolimus incorneal transplantation [49]. This might par-tially be explained by its relatively narrow safe-ty margins. Whereas CSA might be given in abody weight adjusted dose (a suboptimal thera-peutic approach which is practised in some cen-tres in the United States), tacrolimus has to beclosely monitored because the risk of overim-munosuppression is great. This is also thereason for the initially rather unjustified poorreputation of this drug: initially blood levels of20–30 ng/ml were targeted (with correspondingadverse events), whereas today blood levels of5–10 ng/ml are considered to be the optimalrange.

The potency of this drug to inhibit experi-mental corneal allograft rejection after systemicadministration has been proven [2, 42, 43]. Aswith CSA, much hope is pinned on finding anefficient topical administration to prevent sys-temic side effects. Experimentally the efficacy oftopical tacrolimus has yet been proven [9, 13, 22,23], and clinical studies of topical tacrolimus inatopic conjunctivitis are under way.

9.4.5Mycophenolate Mofetil (MMF, CellCept, Myfortic)

Mycophenolate mofetil (MMF) is the bio-availability-enhanced morpholinoethylester ofmycophenolic acid (MPA), which was originallyisolated from Penicillium spp. MMF is rapidlyconverted to MPA, its active compound. Its safe-ty and effectiveness in combination with CSAfollowing kidney transplantation have beenproven in several clinical studies [8, 11, 48, 52].

Unlike CSA or tacrolimus, MMF does not inter-fere with IL-2 pathways. Mycophenolic acidreversibly inhibits the de novo formation ofguanosine nucleotides [1] by inhibiting the en-zyme inosine monophosphate dehydrogenase(with high affinity to the isoform II, which isexpressed in activated lymphocytes).As T and Bcells are predominantly dependent on the denovo synthesis of guanosine nucleotides, thepurine biosynthesis of these cells is selectivelyinhibited [24].

As MMF is not an antimetabolite and doesnot lead to genetic miscoding, it is not carcino-genic.

Mycophenolate Mofetil in Corneal Transplanta-tion. We have been able to prove the potency ofthis drug and its synergistic effect on CSA andFK506 in delaying corneal allograft rejection inthe rat keratoplasty model [41]. Following theseinitial positive experiences we conducted aprospective clinical trial with MMF and CSA inhigh-risk keratoplasty patients. The data fromthis study show a similar efficacy of MMF andCSA in preventing allograft rejection [44]. Butdue to the high therapeutic margin andfavourable safety profile of MMF, costly drugmonitoring is not indicated. Additionally weused this substance in immunological disordersof the eye, again with favourable results [40].

9.4.6Rapamycin (Sirolimus, Rapamune)

Sirolimus (Rapamune) is an immunosuppres-sive agent previously known as rapamycin. Itwas under development for more than 20 yearsbefore it gained FDA approval in 1999. Sirolimusis a macrocyclic lactone produced by Strepto-myces hygroscopicus found in the soil of EasterIsland. Structurally, sirolimus resembles tacro-limus and binds to the same intracellular bind-ing protein or immunophilin known as FKBP-12. However, sirolimus has a novel mechanismof action. Whereas tacrolimus and cyclosporineblock lymphokine (e.g. IL-2) gene transcrip-tion, sirolimus acts later to block IL2-dependentT-lymphocyte proliferation and the stimulationcaused by cross-linkage of CD28, possibly by

9.4 Immunosuppressive Agents 115

blocking activation of a kinase referred to as themammalian target of rapamycin or “mTOR”, aserine-threonine kinase that is important forcell cycle progression. Therefore, sirolimus isbelieved to act in synergy with cyclosporine (ortacrolimus) in suppressing the immune system.

Rapamycin has been shown to be highly effi-cient in preventing experimental solid organ[25, 53] and clinical renal transplantation [4, 26].It is noteworthy that rapamycin is not nephro-toxic, which makes this drug especially interest-ing for renal transplant recipients.

Sirolimus in Corneal Transplantation. A cou-ple of experimental studies have shown the effi-cacy of sirolimus in inhibiting murine cornealallograft rejection [27, 52]. We have conducted asmall clinical study with sirolimus in high-riskcorneal transplantation. We started Rapamuneon the day of transplantation at a dose of2 mg/day. The dose was adjusted to reach plas-ma levels of 4–10 ng/ml on subsequent days, try-ing to keep plasma levels close to 4 ng/ml. Wehave seen that the efficacy of Rapamune in pre-venting corneal allograft rejection is compara-ble to that of cyclosporine and MMF. But it isworth mentioning that we have seen a high inci-dence of side effects in this small group ofpatients.

9.4.7RAD (Everolimus, Certican)

Everolimus is an oral rapamycin derivative pro-duced by Novartis Pharma. It is chemically de-rived from rapamycin which has been obtainedby fermentation of an Actinomycetes strain. Ithas been found that everolimus (40–0-[2-hy-droxyethyl])-RPM is stable in oral formulationsand that its efficacy after oral dosing is at leastequivalent to that of rapamycin [7, 47]. Themode of action is equivalent to that of ra-pamycin, i.e. binding to FKBP, inhibiting TOR1and 2 and hence inhibiting cell-cycle progres-sion of activated T cells.

Everolimus and sirolimus are also called pro-liferation signal inhibitors (PSI), because theyprevent proliferation of T cells.

Everolimus may have a special role in solidorgan transplantation as it has been shown toreduce chronic allograft vasculopathy in suchtransplants [10].

Everolimus in Corneal Transplantation. Wehave tested this new compound in the rat mod-el of corneal transplantation both as a singletherapy and in combination with CSA andMMF. It appears that the potency of everolimusto prevent corneal allograft rejection is compa-rable to CSA. Additionally we have found a syn-ergistic effect of everolimus in a double-drugregimen with CSA as well as with MMF [36, 38,39].

There are to date no clinical data on the effi-cacy and safety of everolimus in corneal trans-plantation.

9.4.8FTY 720

The chemical 2-amino-2[2-(4-octylphenyl)ethyl]-1, 3,propane diol is one of a class of small-molecule immunosuppressive agents. Thiscompound was chemically synthesised in aneffort to minimise the toxic in vivo properties ofa structurally related and highly potent im-munosuppressive agent, myriocin. The mecha-nism of action of FTY720, although not fullycharacterised, appears to be unique among im-munosuppressants. In vivo, FTY720 induces asignificant reduction in the number of circulat-ing lymphocytes. It is thought to act by alteringlymphocyte trafficking/homing patterns throughmodulation of cell surface adhesion receptors.Although much research has yet to be done tounravel the nature of the mechanism of actionof FTY720, its efficacy has been sufficientlyproven in numerous animal models, especiallywhen administered in combination with cyclo-sporine. It has been shown that FTY720 is effica-cious in a variety of transplant and autoimmunemodels without inducing a generalised im-munosuppressed state and is effective in humankidney transplantation.


FTY720 in Corneal Transplantation. We havebeen able to show the efficacy of FTY720 in in-hibiting murine corneal allograft rejection [20].There are to date no data of FTY720 in clinicalcorneal transplantation.

9.4.9Biologic Agents

Reports about the use of biological agents incorneal transplantation are very rare [46]. Tocomplete this overview their mode of action isbriefly outlined.

9.4.9.1Polyclonal Antibodies (e.g. Antithymocyte Globulins)

These agents are derived by injecting animalswith human lymphoid cells, then harvestingand purifying the resultant antibody. Polyclonalantibodies induce the complement lysis oflymphocytes and uptake of lymphocytes by thereticuloendothelial system and mask the lym-phoid cell-surface receptors. Preparations in-clude horse antithymocyte globulin (Atgam)and rabbit antithymocyte globulin (thymoglob-ulin). These agents are used for induction ther-apy and for the treatment of acute rejection insolid organ transplantation.

Adverse effects include fever, chills, throm-bocytopenia, leukopenia, haemolysis, respirato-ry distress, serum sickness, and anaphylaxis.

9.4.9.2Muromonab-CD3

Muromonab-CD3 is a murine monoclonal anti-body of immunoglobulin 2A clones to the CD3portion of the T-cell receptor. It blocks T-cellfunction and has limited reactions with othertissues or cells. This agent is used for inductionand for the therapy of acute rejection (primarytreatment or steroid resistant).

Adverse effects include cytokine release syn-drome (i.e. fever, dyspnoea, wheezes, headache,hypotension) and pulmonary oedema.

9.4.9.3Basiliximab and Daclizumab

Basiliximab (Simulect) and daclizumab (Zena-pax) are humanised monoclonal antibodies thattarget the IL-2 receptor. Clinically, both agentsare very similar, and both are used for inductiontherapy in solid organ transplantation. Theseagents have a very low prevalence of adverseeffects, although hypersensitivity reactionshave been reported with basiliximab (Simulect),albeit rarely.

Perioperative basiliximab has been tested incombination with cyclosporine postoperativelyin a small clinical study with favourable results[46]. In 2004 we started a prospectively ran-domised clinical trial of basiliximab as mono-therapy compared to cyclosporine.


∑ To date the efficacy in preventing cornealgraft rejection has only been proven for cyclosporine, mycophenolate mofetil andrapamycin in prospective clinical trials

∑ The efficacy and safety of tacrolimus inhigh-risk corneal transplantation has beendescribed in a retrospective manner

∑ Especially in high-risk corneal transplanta-tion as it is not a life-saving procedure it isimportant to weigh the pros and cons ofany immunosuppressive regimen

∑ With respect to the profile of side effects, weprefer cyclosporine and mycophenolatemofetil over rapamycin and tacrolimus

9.5Guidelines for Practitioners

9.5.1Preoperative Evaluation

As systemic immunosuppression might pro-mote tumour growth or reactivation of a chron-ic infection, patients need to be checked by theirinternists to rule out neoplasms and infections(blood chemistry, abdominal ultrasound, chestX-ray) before immunosuppression is started.Additionally due to possible drug-specific sideeffects of immunosuppressive agents, renal and

9.5 Guidelines for Practitioners 117

hepatic functions should be controlled. The pa-tients should undergo these examinations at thetime they are put on the waiting list for trans-plantation. If any contraindications against sys-temic immunosuppression are found, theseconditions need to be cleared before transplan-tation. If the conditions cannot be cleared, theindication for high-risk corneal transplantationshould be reconsidered. In this situation the useof an optimally matched graft might be an in-teresting alternative to systemic immunosup-pression. In the case of drug-specific con-traindications, alternative drugs should be used(e.g. in the case of renal impairment, mycophe-nolate mofetil should be used instead of CSA).

9.5.2How To Use Cyclosporine in High-Risk Corneal Transplantation

In addition to perioperative topical and sys-temic steroids, CSA is started on the day of op-eration in a dosage of 100 mg twice daily. With-in the first postoperative week, full blood troughlevels of CSA (12 h after administration) need tobe checked daily, and the dose adjusted to reachserum levels of 120–150 ng/ml. We adjust thedose by using increasing or decreasing steps of25 mg. If serum levels appear to be stable, we re-duce drug-monitoring to once a week in the firstmonths and afterwards to once a month. Addi-tionally we check for liver and kidney functions.Depending on the risk situation we continuetherapy for at least 6 or 12 months and tapertherapy by reducing CSA in 25-mg steps daily.

CSA is especially helpful in high-risk pa-tients who also suffer from atopic dermatitis.CSA should only be used with great caution inpatients with renal impairment, diabetes melli-tus and arterial hypertension.

9.5.3How To Use MMF in High-Risk Corneal Transplantation

The application of MMF following high-riskcorneal transplantation is easier than the use ofCSA. In addition to perioperative topical and

systemic steroids, MMF is started on the day ofoperation in a dosage of 1 g twice daily. Drug-monitoring is not mandatory due to the drug’sbroad safety margins. We perform blood chem-istry once a month as MMF might be myelosup-pressive and might lead to a rise in liver en-zymes. If side effects occur, we reduce MMF to0.5 g twice daily. In the case of drug-specific sideeffects or graft rejection drug-monitoring isindicated to rule out inadequate dosing.

MMF is especially valuable in herpetic oculardisease when combined with acyclovir due to itssynergistic antiherpetic effect [20].

In cases of:∑ Arterial hypertension∑ Diabetes mellitus∑ Chronic renal disease∑ Low compliance

MMF is favoured above CSA or tacrolimus be-cause of its safety margins and the profile ofpossible side effects.After at least 6 or 12 monthsfollowing transplantation, MMF is tapered anddiscontinued within 1 week. In the case of drug-specific side effects we monitor blood levels.

9.5.4How To Use Rapamycin in High-Risk Corneal Transplantation

In our pilot study we have seen that Rapamuneeffectively prevents acute allograft rejection.But due to the rather broad range of side effects,we do not recommend Rapamune in high-riskkeratoplasty at this time point. Rapamuneshould especially not be given to patients withmetabolic disorders (i.e. hypercholesterolaemiaand hypertriglyceridaemia) as it aggravatesthese conditions in more than 50% of patients.

9.5.5How To Use Tacrolimus in High-Risk Corneal Transplantation

In addition to perioperative topical and sys-temic steroids, tacrolimus is started on the dayof transplantation. As with CSA, blood levelsshould be controlled daily in the 1st week. Plas-


ma levels of 5 ng/ml should be aimed for. If plas-ma levels appear to be stable, drug monitoringcan be reduced to once a week in the firstmonths and once a month thereafter. Therapyshould be applied for at least 6–12 months de-pending on the clinical course and tapered outwithin 2 weeks. Probably due to its calcineurinmediated interference with tubular and en-dothelial cells, the profile of side effects is simi-lar to that of CSA. Tacrolimus should not beused in patients with diabetes mellitus, arterialhypertension or renal impairment. Due to thedrug’s narrow safety margins, special care mustbe taken in cases where compliance is not guar-anteed. MMF should be used in these patientsinstead.

9.5.6Combination Therapies

In special situations, i.e. high-risk keratoplastyin an oculus ultimus patient or in limbal stemcell deficiency, the use of a highly effective im-munosuppressive therapy might be indicated.Immunosuppressive potency might be en-hanced by combining two immunosuppres-sants, thereby minimising a drug-specific toxicside effect. We have demonstrated the efficacyand safety of a double-drug regimen with cyclo-sporine and MMF in patients following limbok-eratoplasty [31].

9.6Conclusion

It is now 60 years since we learned about im-munologic graft failure thanks to the pioneer-ing work of Medawar and Maumenee in the1940s. Since that time impressive developmentshave been undertaken in the field of pharma-ceutical immunosuppression. CSA has for manyyears been the gold standard in organ and high-risk corneal transplantation, but it is now ac-companied by mycophenolate mofetil andtacrolimus.

The new compounds now give us the possi-bility not just to maximise immunosuppressivepotency but to apply a more patient-oriented

immunosuppressive regimen. This means first-ly that the immunosuppressants may be chosenaccording to the underlying diseases of the pa-tient (i.e. CSA or tacrolimus in keratoplasty in apatient with atopic dermatitis, MMF in kerato-plasty in patients with herpetic eye disease orpatients with impaired renal function). Second-ly the efficacy of immunosuppression may nowbe adjusted to the clinical situation by addinganother immunosuppressant to a baseline med-ication, thereby minimising a drug-specific tox-ic side effect. Especially in high risk cornealtransplantation since it is not a life-saving pro-cedure, it is important to weigh the pros andcons of any immunosuppressive regimen.

References

1. Allison A, Hovi R, Watts A, Webster A (1977) Therole of de novo purine synthesis in lymphocytetransformation. CIBA Found Symp 48:207

2. Benelli U, Lepri A, Del-Tacca M, Nardi M (1996)FK-506 delays corneal graft rejection in a modelof corneal xenotransplantation. J Ocul PharmacolTher 12(4):425–431

3. Birnbaum F, Reinhard T, Bohringer D, Sundmach-er R (2005) Endothelial cell loss after autologousrotational keratoplasty. Graefes Arch Clin ExpOphthalmol 243(1):57–59

4. Brattstrom C, Tyden G, Sawe J, Herlenius G, Claes-son K, Groth CG (1996) A randomized, double-blind, placebo-controlled study to determinesafety, tolerance, and preliminary pharmacoki-netics of ascending single doses of orally admin-istered sirolimus (rapamycin) in stable renaltransplant recipients. Transplant Proc 28(2):985–986

5. Bussutil RW, Holt CD (1997) Tacrolimus (FK506)is superior to cyclosporine in liver transplanta-tion. Transplant Proc 29:534–538

6. Chen YF, Gebhardt BM, Reidy JJ, Kaufman HE(1990) Cyclosporine-containing collagen shieldssuppress corneal allograft rejection. Am J Oph-thalmol 109(2):132–137

7. Crowe A,Lemaire M (1998) In vitro and in situ ab-sorption of SDZ-RAD using a human intestinalcell line (Caco-2) and a single pass perfusionmodel in rats: comparison with rapamycin.Pharm Res 15(11):1666–1672

8. Deierhoi M, Sollinger H, Diethelm A et al. (1993)One-year follow-up results of a phase 1 trial ofmycophenolate mofetil (RS-61443) in cadavericrenal transplantation. Transplant Proc 25:693f

References 119

9. Dickey JB, Cassidy EM, Bouchard CS (1993) Peri-ocular FK-506 delays allograft rejection in ratpenetrating keratoplasty. Cornea 12(3):204–207

10. Eisen HJ, Tuzcu EM, Dorent R et al. (2003)Everolimus for the prevention of allograft rejec-tion and vasculopathy in cardiac-transplant re-cipients. N Engl J Med 349:847–858

11. European Mycophenolate Mofetil CooperativeStudy Group (1995) Placebo controlled study ofmycophenolate mofetil combined with cyclo-sporin and corticosteroids for the prevention ofacute rejection. Lancet 345:1321–1325

12. Mayer K, Birnbaum F, Reinhard T, Reis A, Braun-stein S, Claas F, Sundmacher R (2004) FTY720prolongs clear corneal allograft survival with adifferential effect on different lymphocyte popu-lations. Br J Ophthalmol 88(7):915–919

13. Hikita N, Lopez JS, Chan C, Mochizuki M,Nussenblatt RB, de Smet MD (1997) Use of topicalFK506 in a corneal graft rejection model in Lewisrats. Invest Ophthalmol Vis Sci 38:901–909

14. Hill JC, Maske R (1988) An animal model forcorneal graft rejection in high-risk keratoplasty.Transplantation 46(1):26–30

15. Hill JC (1995) Systemic cyclosporine in high-riskkeratoplasty: long-term results. Eye 9(4):422–428

16. Hoffmann F, Wiederholt M (1986) Topical cyclo-sporin A in the treatment of corneal graft reac-tion [editorial]. Cornea 5(3):129

17. Kino T, Hatayama H, Hashimoto M et al. (1987)FK-506, a novel immunosuppressant isolatedfrom Streptomyces: 1. Fermentation, isolation, andphysio-chemical and biological characteristics.J Antibiot 40:1249–1255

18. Lakkis FG, Arakelov A, Koniecny BT, Inoue Y(2000) Immunologic “ignorance” of vascularizedorgan transplants in the absence of secondarylymphoid tissue. Nat Med 6:686–688

19. Mayer K, Reinhard T, Reis A, Voiculescu A, Sund-macher R (2003) Synergistic antiherpetic effect ofacyclovir and mycophenolate mofetil followingkeratoplasty in patients with herpetic eye disease:first results of a randomised pilot study. GraefesArch Clin Exp Ophthalmol 241(12):1051–1054

20. Mayer K, Reinhard T, Reis A, Voiculescu A, Sund-macher R (2003) Synergistic antiherpetic effect ofacyclovir and mycophenolate mofetil followingkeratoplasty in patients with herpetic eye disease:first results of a randomised pilot study. GraefesArch Clin Exp Ophthalmol 241(12):1051–1054

21. Mayer K, Birnbaum F, Reinhard T, Reis A, Braun-stein S, Claas F, Sundmacher R (2004) FTY720prolongs clear corneal allograft survival with adifferential effect on different lymphocyte popu-lations. Br J Ophthalmol 88(7):915–919

22. Mills RA, Jones DB,Winkler CR,Wallace GW,Wil-helmus KR (1995) Topical FK-506 prevents exper-imental corneal allograft rejection. Cornea 14(2):157–160

23. Minamoto A, Sakata H, Okada K, Fujihara M(1995) Suppression of corneal graft rejection bysubconjunctival injection of FK-506 in a rat mod-el of penetrating keratoplasty. Jpn J Ophthalmol39(1):12–19

24. Morris R, Hoyt E, Murphy P (1990) Mycophenolicacid morpholinoethylester (RS-61443) is a newimmunosuppressant that prevents and haltsheart allograft rejection by selective inhibition ofT- and B-cell purine synthesis. Transplant Proc22:1659

25. Morris RE, Huang X, Gregory CR, Billingham ME,Rowan R, Shorthouse R, Berry GJ (1995) Studiesin experimental models of chronic rejection: useof rapamycin (sirolimus) and isoxazole deriva-tives (leflunomide and its analogue) for the sup-pression of graft vascular disease and obliterativebronchiolitis. Transplant Proc 27(3):2068–2069

26. Murgia MG, Jordan S, Kahan BD (1996) The sideeffect profile of sirolimus: a phase I study in qui-escent cyclosporine-prednisone-treated renaltransplant patients. Kidney Int 49(1):209–216

27. Olsen TW, Benegas NM, Joplin AC, Evangelista T,Mindrup EA, Holland EJ (1994) Rapamycin in-hibits corneal allograft rejection and neovascu-larization. Arch Ophthalmol 112(11):1471–1475

28. Otto Ch, Ulrichs K (2004) The immunology ofallograft rejection. Tx Med 16:158–171

29. Perry HD,Donnenfeld ED,Acheampong A,Kanel-lopoulos AJ, Sforza PD, D’Aversa G,Wallerstein A,Stern M (1998) Topical cyclosporine A in the management of postkeratoplasty glaucoma and corticosteroid-induced ocular hypertension(CIOH) and the penetration of topical 0.5% cy-closporine A into the cornea and anterior cham-ber. CLAO J 24(3):159–165

30. Reinhard T, Bohringer D, Enzmann J, Kogler G,Wernet P, Bohringer S, Sundmacher R (2004)HLA class I/II matching and chronic endothelialcell loss in penetrating normal risk keratoplasty.Acta Ophthalmol Scand 82(1):13–18

31. Reinhard T, Kontopoulos T, Wernet P, Enzmann J,Sundmacher R (2004) Long-term results of ho-mologous penetrating limbokeratoplasty in totallimbal stem cell insufficiency after chemical/thermal burns. Ophthalmologe 101(7);682–687

32. Reinhard T, Moller M, Sundmacher R (1999) Pen-etrating keratoplasty in patients with atopic der-matitis with and without systemic cyclosporin A.Cornea 18(6):645–651


33. Reinhard T,Sundmacher R,Godehardt E,HeeringP (1997) [Preventive systemic cyclosporin A afterkeratoplasty at increased risk for immune reac-tions as the only elevated risk factor.] Systemis-che Cyclosporin-A-Prophylaxe nach Keratoplas-tiken mit erhohtem Risiko für Immunreaktionenals einzigem erhohten Risikofaktor. Ophthal-mologe 94(7):496–500

34. Reinhard T, Sundmacher R (1996) Local cyclo-sporin A therapy in Thygeson superficial punc-tate keratitis – a pilot study. Klin Monatsbl Au-genheilkd 209(4):224–227

35. Reinhard T, Sundmacher R (1992) Perforatingkeratoplasty in endogenous eczema. An indica-tion for systemic cyclosporin A – a retrospectivestudy of 18 patients. Klin Monatsbl Augenheilkd201(3):159–163

36. Reinhard T, Sundmacher R (1999) Topical cy-closporin A in Thygeson’s superficial punctatekeratitis. Graefes Arch Clin Exp Ophthalmol237(2):109–112

37. Reis A, Megahed M, Reinhard T, Braunstein C,Godehardt E, Sundmacher R (2001) RAD, a newimmunosuppressive macrolide in murine cornealtransplantation. Graefes Arch Clin Exp Ophthal-mol 239(9):689–692

38. Reis A, Megahed M, Reinhard T, Godehardt E,Braunstein C, Sundmacher R (2002) Synergism ofRAD and cyclosporin A in prevention of acute ratcorneal allograft rejection. Cornea 21(1):81–84

39. Reis A, Megahed M, Reinhard T, Godehardt E,Spelsberg H, Braunstein C, Sundmacher R (2000)Coadministration of the new macrolide immuno-suppressant RAD and mycophenolate mofetil inexperimental corneal transplantation. Transplan-tation 70(9):1397–1401

40. Reis A, Reinhard T, Sundmacher R, Althaus C,Voiculescu A, Kutkuhn B (1998) [Mycophenolatemofetil in ocular immunological disorders. Asurvey of the literature with 3 case reports].Mycophenolatmofetil (CellCept) bei okularenimmunologischen Störungen. Literaturübersichtmit 3 Kasuistiken. Klin Monatsbl Augenheilkd213(5):257–261

41. Reis A, Reinhard T, Sundmacher R, Braunstein C,Godehardt E (1998) Effect of mycophenolatemofetil, cyclosporin A, and both in combinationin a murine corneal graft rejection model. Br JOphthalmol 82:700–703

42. Reis A, Reinhard T, Sundmacher R, Godehard E,Braunstein C (1998) A comparative investigationof the effect of FK506 (Prograf®) and CyclosporinA (Sandimmun®, CSA) in murine corneal trans-plantation. Graefes Arch Clin Exp Ophthalmol236:785–789

43. Reis A, Reinhard T, Sundmacher R, Godehard E,Braunstein C (1998) FK506 and mycophenolatemofetil: two novel immunosuppressants in murinecorneal transplantation. Transpl Proc 30(8)

44. Reis A, Reinhard T, Voiculescu A, Kutkuhn B,Godehardt E, Spelsberg H, Althaus C, Sundmach-er R (1999) Mycophenolatemofetil versus cyclo-sporin A in high-risk keratoplasty patients: aprospectively randomized clinical trial. Br J Oph-thalmol 83:1268–1271

45. Sawada S, Suzuki G, Kawase Y, Takaku F (1987)Novel immunosuppressive agent, FK-506: in vitroeffects on the cloned T cell activation. J Immunol40:1249–1255

46. Schmitz K, Hitzer S, Behrens-Baumann W (2002)Immune suppression by combination therapywith basiliximab and cyclosporin in high riskkeratoplasty. A pilot study. Ophthalmologe 99(1):38–45

47. Schuurman HJ, Schuler W, Ringers J, Jonker M(1998) The macrolide SDZ RAD is efficacious in anonhuman primate model of allotransplantation.Transplant Proc 30(5):2198–2199

48. Simmons WD, Rayhill SC, Sollinger HW (1997)Preliminary risk-benefit assessment of mycoph-onlate mofetil in transplant rejection. Drug Safe-ty 17(2):75–92

49. Sloper CM, Powell RJ, Dua HS (2001) Tacrolimus(FK506) in the management of high-risk cornealand limbal grafts; Ophthalmology 108(10):1838–1844

50. Streilein JW, Okamoto S, Sano Y, Taylor AW(2000) Neural control of ocular immune privi-lege. Ann NY Acad Sci 917:297–306

51. The Tricontinental Mycophenolate Mofetil RenalTransplantation Study Group (1996) A blinded,randomized clinical trial of mycophonlatemofetil for the prevention of acute rejection incadaveric renal transplantation. Transplantation61(7):1029–1037

52. Thompson P, Xu D, Brunette I, Chen H (1998)Combined effect of rapamycin and cyclosporinein the prevention of rat corneal allograft rejec-tion. Transplant Proc 30(4):1033–1035

53. Vu MD, Qi S, Xu D, Wu J, Peng J, Daloze P, SehgalS, Leduc B, Chen H (1998) Synergistic effects ofmycophenolate mofetil and sirolimus in preven-tion of acute heart, pancreas, and kidney allograftrejection and in reversal of ongoing heart allo-graft rejection in the rat. Transplantation 66(12):1575–1580

54. Woodle ES, Thistlethwaite JR, Gordon JH for theFK506 Kidney Transplant Rescue Study Group(1996) A multicenter trial of FK506 (tacrolimus)therapy in refractory acute renal allograft rejec-tion. Transplantation 62:594–599

References 121

55. Yoshimura N, Matsui S, Hamashima T, Oka T(1989) Effects of a new immunosuppressiveagent,FK-506,on human lymphocyte responsive-ness in vitro: 1: Inhibition of expression of al-

loantigen-activated suppressor cells, as well as in-duction of alloreactivity. Transplantation 47:351–356


10.1Introduction

Zirm in 1905 was the first surgeon to perform asuccessful homologous penetrating keratoplas-ty (PKP) in a human patient [84]. The operationbecame more successful with the developmentof more delicate instruments, use of the operat-ing microscope, and the availability of antibi-otics, antivirals and corticosteroids. Today, stillunsolved problems include: (1) high/irregu-lar astigmatism, (2) trephination of unstablecornea, (3) surface pathologies, (4) immunolog-ic graft rejection, (5) secondary glaucomas, (6)chronic endothelial cell loss of the transplant,(7) recurrences of the disease, and (8) a lack ofdonor tissue.

With the improved understanding and man-agement of immunologic problems during pastfew decades, the microsurgeon’s main attentionin corneal transplantation has shifted from pre-serving a “clear graft” towards achieving a goodrefractive outcome. Thus, PKP today is nolonger just a “curative” but has also become asort of “refractive” procedure. Today, a crystalclear corneal graft after PKP with high and/orirregular astigmatism – especially if in associa-tion with high anisometropia – can no longer beconsidered “successful” in normal-risk kerato-plasties. Deluded by advertisements of refrac-tive surgery, patients expect an optimal visualacuity preferably without spectacles. Many pa-tients consider the necessity of wearing contactlens as representing a partial failure of the inter-vention. Especially older PKP patients cannotcope with contact lenses manually and/or men-tally. Additional “dysfunctional tear syndrome”and blepharitis further promote contact lens

Trephination in Penetrating Keratoplasty

Berthold Seitz, Achim Langenbucher, Gottfried O.H. Naumann

10

|

∑ Donor and host trephination should beperformed with the same system from the epithelial side

∑ A horizontal position of the limbal plane is essential

∑ The graft size should be adjusted individu-ally (“as large as possible, and as small asnecessary”)

∑ Limbal centration is to be preferred overpupil centration (especially in keratoconus!)

∑ Avoid excessive graft over- or undersize∑ Intraoperative adjustment is required of

double running suture∑ Nonmechanical excimer laser trephination

results in:– Lower astigmatism– Higher regularity of topography– Better visual acuity – especially

in young patients with keratoconus∑ In unstable corneas (e.g., after RK, iatro-

genic keratectasia after LASIK, descemeto-cele, perforated ulcer), laser applicationmakes trephination feasible

∑ New nut-and-bolt type variants for potentially self-sealing donor/host appositions are on the horizon (“no-stitch keratoplasty”)

∑ Femtosecond laser application may be the “excitement of tomorrow” in microsurgeryof the cornea

Core Messages

intolerance in this age group. Persisting cornealhypesthesia after PKP for many years can delayrecognition of contact lens induced damage tothe cornea.

It has been debated whether cutting or sutur-ing is more important for the regularity of thetransplant curvature. We have always stressedthat: (1) early postoperative astigmatism withsutures in place should be differentiated from(2) late persisting postoperative astigmatismwithout sutures [59].


Two major types of post-PKP astigmatismneed to be distinguished:1. Early postoperatively with sutures in place

predominantly depending on:– Symmetry of suture positions– Depth of suture track in graft

and recipient– Homogeneity of suture tension– Microsurgeon’s “hand writing”

2. Late postoperatively persisting without sutures predominantly depending on:– Cut quality– Wound configuration

(horizontal/vertical)– Symmetry of graft placement– Wound healing

10.2Astigmatism and Keratoplasty

10.2.1Definition of Post-keratoplastyAstigmatism

The cornea contributes about two-thirds of therefractive power of a human eye.Surgical proce-dures on the cornea may therefore influence thestate of refraction considerably. Corneal astig-matism is an optical aberration, resulting fromunequal refraction of entering light in differentmeridians of the corneal surface. Astigmatismafter PKP is often irregular, i.e., two or moremeridians are separated from each other by an angle not equal to 90°. Two or more steephemimeridians are not located opposite to eachother. The same may be true for the flat

hemimeridians. In addition, the refractive pow-er of corresponding hemimeridians may differ.Especially with sutures in place, patients acceptmuch less subjective cylinder than indicated byobjective measures such as keratometry or to-pography analysis [20]. In cases of highly irreg-ular astigmatism, good visual acuity can only beachieved by hard contact lenses (Table 10.1).

124 Chapter 10 Trephination in Penetrating Keratoplasty

Table 10.1. Assessment of astigmatism and visualacuity after keratoplasty (SRI, surface regularityindex; SAI, surface asymmetry index; PVA, potentialvisual acuity)

1. Uncorrected visual acuity

2. Keratometrya) Absolute valuesb) Angle of steep and flat meridian

separately (Dπ90°)c) Classification of irregularity [59, 62]

3. Topography analysisa) Meridiansb) Hemimeridiansc) Irregularity (SRI, SAI)d) Semiquantitative classification [29]

4. Objective refractometry/retinoscopy

5. Subjective refractometry and spectacle-corrected visual acuity

6. Pinhole

7. Diagnostic contact lens

Fig. 10.1. Semiquantitative classification of regular-ity of keratometry mires (ophthalmometer, type H,190071, Zeiss, Jena, Germany) (0, regular; 1, mildlyirregular; 2, severely irregular; 3, not measurable) [59,62, 83]

After PKP we recommend documenting thekeratometric refractive power separately in thesteep and in the flat meridian with individualaxis notation and assessment of the degree of“keratometric irregularity” (Fig. 10.1). Insteadof “42.0+4.5/0°,” we suggest writing “42.0/0°(irreg. 1); 46.5/70° (irreg. 2)” [62].

Besides keratometry, topography analysis isindispensable for mapping the corneal powerover the entire graft. Refractive powers and in-dividual axes of the four hemimeridians arecomplemented by system specific indices, e.g.,SRI (surface regularity index) and SAI (surfaceasymmetry index) of the TMS-1 topographysystem. In addition, we suggest a semiquantita-tive classification of post-keratoplasty topogra-phy in seven groups (Fig. 10.2).


Studies intending to compare the corneal curvatures after different trephination or suturing techniques for PKP should includethe following:∑ Subjective cylinder and keratometric/

topographic astigmatism∑ Portion of irregular/not measurable

astigmatism

∑ Astigmatism with “all-sutures-out”and vector-corrected astigmatism

10.2.2Reasons for Astigmatism After Keratoplasty(Table 10.2)

Each of the multiple steps from donor selection,intraoperative trephination and suturing tech-nique to type and quality of postoperative carecan determine not only the clarity of the graftbut also its final refractive result.

Besides intrinsic factors of donor and recipi-ent, the short-term astigmatism with sutures inplace seems to depend more on the symmetry ofthe sutures including methods of intra- andpostoperative suture adjustments. After sutureremoval corneal curvature typically becomesmore regular [35, 62], but the amount of netastigmatism may increase considerably [36, 38].

Thus, it has been concluded that factors di-rectly or indirectly related to the quality of thewound geometry have a predominant influenceon the long-term residual astigmatism aftersuture removal [59].

10.2 Astigmatism and Keratoplasty 125

Fig. 10.2. Semiquantitativeclassification of corneal topo-graphy after PKP [29]: 1, ortho-gonal symmetric (i.e., differenceof maximal powers of opposinghemimeridians is less than2 diopters and deviation of axisof opposing hemimeridians isless than 20°); 2, orthogonalnon-symmetric; 3, non-ortho-gonal symmetric; 4, non-orthogonal non-symmetric;5, keratoconus-like (a steep sector is opposing a flat sector at the apex, difference betweensteep and flat hemimeridian at least 2 diopters); 6, polyaxi-gonal (at least three steep/flatsectors can be recognized, atleast 2 diopters of power differ-ence between steep and flathemimeridians); 7, irregular


Table 10.2. Potential causative factors of high and/or irregular astigmatism after keratoplasty [59]

1. Preoperative factorsa) Age of donor (infant!)b) Size of recipient cornea

i) Keratoconus >Fuchs’ dystrophy [60]ii) Microcornea

c) Topography of donord) Topography of recipiente) Disharmony between donor and recipient topographyf) Pathologic properties of recipient

i) Peripheral thinning or ectasiaii) Focal edema/focal scariii) Defects in Bowman’s layeriv) Vascularizationv) Preceding keratoplasty (especially decentered)

g) Aphakia

2. Intraoperative factorsa) Decentration of donor excision and/or recipient bedb) “Vertical tilt” due to discrepancies of wound configuration [42]

i) Application of different trephine systems for donor and recipientii) Trephine tilt (i.e., not parallel to optical axis)iii) Limbal plane not horizontaliv) “Shifting” of trephine during cuttingv) Too high/low intraocular pressure

c) “Horizontal torsion” [42]i) Asymmetric placement of second cardinal suture (Dπ180°)ii) Mismatch of donor and recipient due to form incongruenceiii) Focal overlap or dehiscence of donor button in recipient bed

d) Excessive over-/undersize of donore) Distortion and squeezing of cornea (e.g., due to dull trephine)f) Traumatizing the cornea with instrumentsg) Suture-related factors

i) Suture materialii) Suture technique (interrupted, single running, double running, combinations)iii) Length of stitchiv) Depth of stitchv) Angle of stitch towards graft-host appositionvi) Suture tensionvii) “Depth disparity”

h) Simultaneous intraocular surgery (e.g., triple procedure, IOL exchange)i) Fixation rings and lid speculaj) Surgeon’s experience

3. Postoperative factorsa) Suture-related factors

i) “Cheese wiring” of suturesii) Suture looseningiii) Suture adjustment/selective suture removaliv) Time point of suture removal

b) Wound healing processesi) Wound dehiscenceii) Retrocorneal membraneiii) Incarceration of overlapping tissueiv) Focal vascularization

c) Medication (e.g., corticosteroids)d) Postoperative trauma

10.2.2.1Preoperative Determinants

Infant corneas have high refractive power(>50 diopters) and tend to steepen further aftertransplantation due to the biomechanical insta-bility of the tissue. Thus, Pfister and Breaud sug-gested using infant corneas to compensate foraphakia. However, the refractive outcome var-ied considerably and was not predictable [49].Thus, we do not recommend the use of infantdonor corneas for grafting.

Today, donor topography is still rarely per-formed. The higher the immanent preoperativeastigmatism of donor and recipient, the moreprobable it is that dysharmony between donorand recipient topography results in high astig-matism after suture removal [10, 15, 56]. Espe-cially high congenital astigmatism, keratoconusand previous corneal refractive surgery must beruled out in potential donors.

10.2.2.2Intraoperative Determinants (Fig. 10.3)

Asymmetrically placed fixation rings (e.g.,Flieringa or McNeill-Goldmann) may induce anastigmatism of up to 10 diopters [45]. Thus,post-PKP astigmatism is typically higher inaphakic than in phakic or pseudophakic PKP[48]. Even simple lid specula may be responsiblefor 3 diopters of with-the-rule astigmatism [45].

Decentration. Besides a higher incidence ofimmunologic graft reactions due to proximityto the limbal vessels, decentration of hosttrephination (>1 mm) may result in higherastigmatism. The flat axis of astigmatism pointstowards the direction of decentration [30, 75].Due to the thickness gradient from the center tothe periphery, donor decentration may alsohave a minor impact on post-PKP astigmatism[61].

“Vertical Tilt.” The amount of persisting post-PKP astigmatism after suture removal dependssignificantly on the incongruences (“mismatch-es”) of shape and cut angles of donor and recip-ient wounds [50,74,75].Theoretically,a trephinetilt of 5° (10°) can induce 1.6 (5.9) diopters of

astigmatism with an 8-mm-diameter graft [22].Especially tilted hand-held trephines and ne-glecting the horizontal position of the limbalplane are reasons for the “vertical tilt” phenom-enon. In addition, application of differenttrephine systems and different trephination di-rections (e.g., punching the donor from theendothelial side) in donor and host are crucialfactors.

“Horizontal Torsion.” One of the major predis-positions for regular all-suture-out curvatureafter PKP is the 360° symmetric apposition ofthe donor button in the recipient bed. Especial-ly the correct positioning of the second cardinalsuture opposite to the first one is crucial. Asym-metric placement of the second cardinal sutureresults in a tissue deficit on one side whichneeds to be compensated by forced suture adap-tation. In the case of long shallow suture bites, aregional flattening may result. In the case ofshort and deep suture bites, a central steepeningmay result, in analogy to sutured wedge resec-tions. On the other side a tissue surplus may re-sult in peripheral donor tissue compressionwith peripheral steepening and consecutivecentral flattening [74].


Fig. 10.3. Main reasons for high post-keratoplastyastigmatism: top decentration of donor and/or recip-ient trephination; middle “vertical tilt” due to incon-gruent cut angles; bottom “horizontal torsion” due toasymmetric suturing (modified from [42])

An analogous situation arises when the re-cipient bed is cut asymmetrically elliptical [34, 46, 78]. This may result from asymmetricbulging of the unstable cornea into the trephineopening or even by using an obturator in thecase of keratoconus [21]. Mechanical trephines,such as hand-held or motor trephines, may re-sult in oval-shaped host beds even if a circularround excision was intended [9].

Likewise, in donor trephination a trephinetilt of 20° may induce a difference of about0.5 mm between the maximal and minimal di-ameter, resulting in an elliptical donor button[45]. Suturing of such an elliptical donor buttonin a round bed will result in a peripheral steep-ening in the major axis due to tissue compres-sion and – consequently – a central flattening inthis (hemi-)meridian [8]. A wound disparity of0.1 mm is supposed to create an astigmatism ofabout 1 diopter [45, 74].

Undoubtedly, the technique for adequategraft-host adaptation by means of four to eightcardinal sutures is determined – at least in part– by the experience of the microsurgeon. Thesame holds true for the correct performance, in-terpretation and consequences of intraopera-tive keratoscopy. However, even if adequatesuture distribution and tension as well as intra-/postoperative suture adjustments compensatefor the fundamental intraoperative determi-nants of post-PKP astigmatism in the earlystage, suture removal – even after years – mayresult in major changes of topography and adramatic increase in astigmatism [36, 38].


Major intraoperative determinants for high/irregular astigmatism after suture removal include [42]:∑ Decentration (donor and/or recipient

trephination)∑ “Vertical tilt” (incongruent cut angles

between donor and host)∑ “Horizontal torsion” (horizontal discrepan-

cy of donor and host shape or asymmetricsuturing – second cardinal suture!)

10.2.2.3Postoperative Determinants

Postoperative suture adjustment or selectiveremoval of single sutures may have a favorableimpact on the early post-PKP astigmatism.However, changes of corneal curvature are un-predictable after suture removal [36, 38]. At thistime there is still no reliable indicator availableto the microsurgeon instructing him about theamount and direction of impending astigma-tism changes of the graft after suture removal.There is some evidence that a high coincidenceof the axes of refractive, keratometric and topo-graphic astigmatism with the suture in placespeaks in favor of decreasing astigmatism to beexpected after suture removal [54]. Thus, in thecase of intact sutures, lack of vascularization, alow amount of astigmatism, and high topo-graphic regularity resulting in good spectacle-corrected visual acuity, microsurgeons will tendto leave the suture in place for a longer period oftime under regular controls and adequate coun-seling of a compliant patient. However, it mustbe considered an illusion that keeping the su-tures in place for a longer time would help topreserve a favorable topography after final su-ture removal [11, 14, 36, 38, 70]. Especially stepformations after suture removal – often after inadequate trauma – will result in a flathemimeridian and irregular high astigmatism.For this reason, such steps at the graft-hostjunction need immediate surgical repair to pre-serve a good long-term refractive result even ifthe anterior chamber is not opened [18].


The pathomechanism of astigmatism increaseafter suture removal may be as follows:∑ A low quality of trephination wound

and geometric incongruences (horizontaland vertical) require a higher suture tension to guarantee:– Watertight wound closure– A pseudo-optimal topography early

postoperatively∑ Asymmetric regional forces between donor

and host may cause inhomogeneous woundhealing


∑ Removal of sutures liberates forces due to:(1) geometric incongruences and (2) inho-mogeneous wound healing

∑ Thus: horizontal, vertical and topographicdiscrepancies between donor and host intraoperatively are responsible for an increase in astigmatism after suture removal

10.2.3Prevention/Prophylaxis of Astigmatism After Keratoplasty

The large number of treatment options forastigmatism after PKP leads to the conclusionthat none of the methods is really convincing.Therefore, prophylaxis of high and/or irregularastigmatism is preferred over treatment [59].

10.2.3.1Alternatives “Without Sutures”

Alternatives “without sutures” include pho-totherapeutic keratectomy (PTK) in the case ofsuperficial corneal diseases. PTK yields goodresults especially with recurrences of cornealdystrophies after PKP. In order to avoid suturesinvolving Bowman’s layer, potentially self-seal-ing nut-bolt variants of donor-recipient apposi-tion have been investigated. One approach is di-vergent cut angles that may be created usinglasers [57]. The increased contact area reducesthe probability of wound dehiscence, the small-er diameter at the level of Bowman’s layer in-creases the distance from the limbal vessels withfavorable effects concerning immunologic graftreactions, and the larger diameter at the level ofDescemet’s membrane increases the amount oftransplanted endothelial cells with favorableeffects in Fuchs’ dystrophy and aphakic/pseudophakic bullous keratopathy. It has beenshown that the stability of the graft in the recip-ient bed increases with increasing divergence ofthe cut angles [57].Additional application of tis-sue glue, a temporary therapeutic contact lensor an intrastromal suture may further increasethe stability of the graft-host junction.

An analogous approach was followed by in-troducing an inverse mushroom-shaped trephi-nation with the larger diameter of the graft atthe level of Descemet’s membrane [7, 67].

In order to leave the architecture of the cen-tral cornea untouched, endothelial cell trans-plantation has been investigated and posteriorlamellar keratoplasty (PLKP) has been intro-duced into clinical routine by Melles [37] in Eu-rope in 1998 and later modified by Terry in theUnited States [71] in cases of sole endothelialfailure.

10.2.3.2Ten Precautions During Surgery

1. Donor topography should be attempted forexclusion of previous refractive surgery,keratoconus/high astigmatism, and “harmo-nization” of donor and recipient topography[16, 56, 59].

2. Donor and recipient trephination should beperformed from the epithelial side with thesame system, which – from our point of view– predisposes to congruent cut surfaces andangles in donor and recipient. For this pur-pose an artificial anterior chamber is usedfor donor trephination although the wholeglobe would yield even better results [27].

3. Orientation structures in donor and host fa-cilitate the correct placement of the first fourcardinal sutures to avoid horizontal torsion[2].

4. A measurable improvement seems possibleusing the Krumeich guided trephine system(GTS) [4], the second generation Hannatrephine [81] and our technique of nonme-chanical trephination with the excimer laser[58, 66].

5. Horizontal positioning of head and limbalplane is indispensable for state-of-the-artPKP surgery in order to avoid decentration,vertical tilt and horizontal torsion [59].

6. Graft size should be adjusted individually(“as large as possible, as small as necessary”)[60, 62].


7. Limbal centration should be preferred overpupil centration (especially in keratoconus –“optical displacement of pupil”) [31].

8. Excessive graft over- or undersize should beavoided to prevent stretching or compres-sion of peripheral donor tissue [19, 47, 82].

9. As long as Bowman’s layer is intact, a doublerunning cross-stitch suture (according toHoffmann [17]) is preferred since it results ingreater topographic regularity, earlier visualrehabilitation and less loosening of sutures,with suture replacement only rarely required.

10.Intraoperative keratoscopy should be ap-plied after removal of lid specula and fixa-tion sutures. Unstable donor epitheliumwould be better removed to allow for repro-ducible results. Adjustment of double run-ning sutures or replacement of single suturesmay be indicated [3].


Requirements for “the optimal trephination”include:∑ Full visual control∑ No contact∑ Optimal donor and host centration∑ Identical shape of donor and host

(typically circular)∑ Congruent cut angles∑ 360° symmetric donor host alignment∑ No necessity to complete trephination

by scissors∑ No damage to intraocular tissues∑ Future: self-sealing donor/host apposition

10.3Trephination Techniques

The principal indications for keratoplasty in-clude optical, curative and tectonic factors(Table 10.3). Overlaps between the different cat-egories may occur. But corneal transplants mayalso be classified according to the type of donor

material, the vertical shape of the graft, the hor-izontal shape of the graft and the location of thegraft within the host (Table 10.4) [40].

A few general technical details concerningPKP need to be mentioned [40, 42]:1. General anesthesia has advantages over local

anesthesia. The arterial blood pressureshould be kept low as the eye is opened(“controlled arterial hypotension”).

2. To protect the crystalline lens in phakic ker-atoplasty, usually the pupil is constricted.

3. Before recipient trephination, a stab-likeparacentesis at the limbus is performed.

4. The limbal plane must be horizontal duringtrephination.

5. An iridotomy prevents pupillary block andacute angle closure glaucoma (so-called Ur-rets-Zavalia syndrome in the case of dilatedpupil with iris sphincter necrosis [43]).

6. The second cardinal suture is crucial for graftalignment.


Table 10.3. Principal indications for keratoplasty(modified from [40])

1. Opticala) Opacitiesb) Pathologic curvature

2. Curativea) Deep keratitis (e.g., herpetic keratitis with

granulomatous reaction to Descemet’smembrane or Acanthamoeba keratitis)

b) Endothelial diseases (primary or secondary)

c) Perforated corneal ulcer

3. Tectonica) Traumatic corneal defectsb) Infectious corneal defectsc) Postoperative fistula after cataract

extraction or antiglaucomatous surgeryd) After “block excision” [44]

i) Uveal tumorsii) Localized epithelial downgrowth

(cysts)e) Reconstruction of the anterior segment

10.3.1Principal Considerations

10.3.1.1Donor Trephination

From a 16-mm corneoscleral button as providedby the Eye Bank, the transplant can be created intwo principal ways:1. The original method used is for the donor

button to be punched from the endothelialside against a firm surface (such as a paraffinor Teflon block) using special trephines(Lochpfeifentrepan) [6, 80]. Care must betaken to ensure a proper alignment whencutting since a beveled cut will result if theblade is not perpendicular to the cuttingblock. This risk may be decreased by the useof “guided donor trephine” systems (e.g.,“guillotines”) (Fig. 10.4).On histological evaluation, the cut surfaceswithout consideration of the cut angles seemto be almost “perfect.” However, deviation ofthe cut direction outwards results in conver-gent cut angles due to a smaller diameter atthe level of Descemet’s membrane and a larg-er diameter at the level of Bowman’s layer(“undercut”) (Fig. 10.4D) [76].

2. Since the development of “artificial anteriorchambers” [23], microsurgeons have had theopportunity to perform donor trephinationfrom the epithelial side, which is the same di-rection as in the host. If pressure in the arti-ficial anterior chamber is kept normal (e.g.,22 mmHg), the advantages with respect tocut angles are obvious [55]. However, fixingthe corneoscleral button in an artificial ante-rior chamber may induce a considerableamount of astigmatism. This problem can beovercome by using an artificial anteriorchamber with a larger central opening, leav-ing the limbus untouched during fixation for trephination from the epithelial side. Inthis setting the corneoscleral limbus seemsto have a protective effect concerning thecentral corneal topography of the fixatedcornea [27].


∑ Trephination of the donor button shouldpreferably be performed from the epithelialside using an artificial anterior chamberwith a large central opening

∑ Punching the donor from the endothelialside results in an undercut at the level ofDescemet’s membrane with convergent cutangles

10.3 Trephination Techniques 131

Table 10.4. Terminology of various types of keratoplasty (modified from [40])

Donor cornea Vertical shape Horizontal Location of graft shape of graft within the host

Lamellar (anterior vs. posterior)

Penetrating

Mushroom

Inverse mushroom [67]

Circular

Elliptical

Semilunar

Rectangular

Triangular

Ring-shaped

Autologous (autograft)

Homologous (allograft)

Heterologous (xenograft)

Alloplastic (keratoprosthesis)

Central

Eccentric

Marginal

10.3.1.2Recipient Trephination

For recipient trephination, the horizontal posi-tion of the head and especially the limbal planeis indispensable. To increase the overview andreduce vis à tergo, the Lieberman speculum is preferred. Any viscoelastic agent may be used

to stabilize the anterior chamber during tre-phination. A Flieringa ring is not necessary forPKP or the triple procedure, but is helpful incases of aphakic eyes, especially if a secondarysclera-fixated IOL is inserted. The ring can besutured temporarily onto the globe using 6-0Vicryl sutures through the conjunctiva andepisclera.


Fig. 10.4 A–D. Donor trephination from the endothelial side. A Correct position of hand-held trephine; B tilt-ed trephine; C “guillotine” to avoid trephine tilt; D smooth cut surface but “undercut” at the level of Descemet’smembrane

A

B D

C

Investigations by Van Rij and Waringdemonstrated that in recipient trephination alltrephine systems result in an opening largerthan the trephine size. In addition, the diameteris larger at the level of Descemet’s membrane,resulting in divergent cut angles [76]. This canbe explained by the “ballooning” of the corneato be excised into the trephine opening due tothe pressure executed. The higher the intraocu-lar pressure, the more divergent the angles to beexpected [55]. This phenomenon of “balloon-ing” is one of the major drawbacks of a mechan-ical trephine and can be prohibited – at least inpart – by the use of an “obturator.” However,Kaufman stresses that the use of an obturator inkeratoconus may result in other than roundhost openings such as pear-shaped holes [21].

The combination of a donor punched fromthe endothelial side with convergent cut anglesand a host opening with divergent cut angleswill result in a triangular-shaped tissue defect atthe level of Descemet’s membrane that has to becompensated for with increased suture tensionand – consequently – vertical tilt (Fig. 10.5).


∑ Horizontal positioning of limbal plane isindispensable

∑ Flieringa ring is only necessary in aphakiceyes

∑ The higher the intraocular pressure (iatrogenic!) the more divergent are the cut angles to be expected [55]

10.3.1.3Graft Size and “Oversize”

Graft Size. In a quantitative study we foundthat the corneal diameter of keratoconus pa-tients was larger than that of Fuchs’ patients(mean horizontal diameter of 11.8 mm in kera-toconus patients and 11.3 mm in Fuchs’ patients)[60]. In general, a good optical performancerequires a larger graft, whereas a low rate ofimmunologic graft reactions tends to be seenwith smaller grafts. Therefore, the graft shouldbe “as large as possible, but as small as neces-sary.” For many eyes with keratoconus an 8.0-mm diameter and in many eyes with Fuchs’dystrophy a 7.5-mm diameter prove to be goodoptions as a prerequisite for obtaining tissuefrom the Eye Bank. Today, graft diameters of5.5–7.0 mm are only rarely required and justi-fied.

It has been supposed that smaller graftsmight be associated with a higher post-kerato-plasty astigmatism. In a recent study we found[62]:1. A flatter curvature with smaller grafts2. A higher topographic irregularity with

smaller grafts3. A higher proportion of unmeasurable ker-

atometry mires with smaller grafts4. A tendency towards regularization of topo-

graphy after suture removal5. No difference concerning the amount of net

astigmatism between different graft sizeseither with or without sutures

The major reason for the flatter and more irreg-ular graft with smaller diameters seems to bethe closer position of the proximal suture endsin relation to the optical center of the graft. Thiswill be pronounced in particular with wider su-ture bites. After suture removal the potentiallytopography disturbing circular scar at the graft-host junction is located closer to the line of sightwith smaller grafts. This may explain that over-all the regularity of graft topography increases


Fig. 10.5. Combination of donor trephined from theendothelial side (convergent cut angle) and mechani-cally trephined recipient (divergent cut angle) resultsin a triangular-shaped tissue deficit at the level ofDescemet’s membrane which has to be compensatedby suture tension resulting in central flattening andvertical tilt

with suture removal but that major differencesbetween various graft sizes do persist.

Larger sizes may be considered for eccentrictectonic corneoscleral grafts (e.g., after theblock excision of tumors of the anterior uvea orcystic epithelial downgrowth [44]) and in buph-thalmos [73]. But we do not recommend graftsizes over 8.5 mm in buphthalmos for immuno-logic reasons [52].

Recent studies indicate that the rate ofchronic endothelial cell loss after PKP dependson the initial diagnosis [32, 53]. Endothelial mi-gration from donor to recipient in pseudopha-kic bullous keratopathy along a density gradientis thought to be the reason for this phenome-non. Therefore, eyes with bullous keratopathymay require a larger graft not just to improvethe optical performance but rather to transplantas many endothelial cells as possible. Neverthe-less, graft size has to be judged by the surgeon in-dividually in every single case before recipienttrephination to achieve the best compromisebetween immunologic purposes and opticalquality [59,60].A slit lamp with a measuring de-vice (scale), e.g., a Haag-Streit slit lamp, orcalipers for intraoperative application may behelpful. Prior removal of vascularized pannus(in contrast to vascularized stromal scars) mayrender a larger “individual optimal graft size”possible for transplantation of more endothelialcells and better graft topography.

Graft “Oversize.” In mechanical trephination,the diameter of the recipient bed tends to belarger and the diameter of the donor button,punched from the endothelial side, tends to besmaller than the trephine diameter, which mayaffect the resulting spherical equivalent [76].Thus, “oversizing” the donor button by 0.25–0.50 mm is commonly done to compensate forrefractive effects and to reduce crowding of thechamber angle and therefore postoperative“glaucoma” [47]. An oversize of 0.25 mm com-pared to one of 0 mm or 0.5 mm may accountfor a difference in keratometric readings of1.5 diopters after suture removal. Javadi et al.found no difference in astigmatism in compar-ing 0.25 mm and 0.50 mm graft oversize [19].However, Perl et al. stressed that oversizing thegraft by 0.5 mm (punched from the endothelial

side) may result in significantly increasedcorneal astigmatism [47]. In keratoconus, samesize donors were found to reduce resultingmyopia.We do not recommend undersizing of agraft!

In contrast, with guided trephines and lasertrephination (donor from the epithelial side),attempted diameters are indeed achieved withcongruent cut angles. Thus, donor oversize isnot necessary.


∑ Typically, keratoconus corneas are largerthan Fuchs’ dystrophy corneas

∑ Graft size has to be judged by the micro-surgeon individually in every single casebefore recipient trephination to achieve thebest compromise between immunologicpurposes and optical quality

∑ Donor trephination from the endothelialside results in a smaller donor button thantrephine size and convergent cut angles(“undercut”)

∑ Recipient trephination results in largeropenings than trephine size and divergentcut angles

∑ This discrepancy makes a donor “oversize”of ≥0.25 mm necessary

∑ Same size grafts are feasible if the donor iscreated by means of an artificial anteriorchamber from the epithelial side

∑ Undersizing the graft for simultaneouscorrection of myopia in keratoconus is not recommended (watertight wound! irregular astigmatism!)

10.3.1.4Pupil Versus Limbal Centration

Centration is crucial with respect to immuno-logic graft reaction and post-PKP astigmatism.Typically a compromise between limbal andpupil centration is attempted in the case of non-traumatized pupils. However, limbal centrationis preferred especially in keratoconus, scarsafter trauma or irregular astigmatism of otherorigins. In such eyes the center of the visible(“entrance”) pupil may be dislocated from thatof the real anatomic pupil [31].


An eight-line radial keratotomy marker maybe used to ensure centration (Fig. 10.6). Anadditional central dot-like mark may be helpful for certain trephine systems (e.g., Hessburg-Baron).

If the broadening of the superior limbus dueto a vascularized pannus is neglected intraoper-atively, an inferior decentration may be recog-nized on the next day at the slit-lamp.


∑ In doubt, limbal centration is preferred overpupil centration

10.3.1.5“Harmonization” of Donor and Patient Corneal Topography

Keratometric readings of the donor cornea arestill usually neglected. However, it might be bet-ter to consider them to improve predictability ofthe final refractive outcome after PKP [10,16,56].This may help to avoid transplantation of corneaswith unusual or abnormal curvatures. In addi-tion, it may allow a more accurate selection ofintraocular lens power in triple procedures.

The vertical difference at the graft-host junc-tion due to the different curvatures of donorand recipient must be compensated intraopera-tively by suture tension to avoid a step forma-tion. The resulting forces may be co-responsible

for the amount of relative change in curvatureafter suture removal. Therefore, “harmoniza-tion” of donor and recipient topography shouldallow for minimization of the residual astigma-tism for a given pair of donor and recipient [56].The use of an artificial anterior chamber en-ables donor topography analysis and allows the“contour line” of the trephination edges in bothdonor and recipient to be calculated. A comput-erized simulation of graft rotation in the recipi-ent bed may help to find an angle of graft rota-tion at which topographical misalignment isminimal.

Grütters et al. have proposed “astigmatism-oriented perforating keratoplasty”, i.e., match-ing the flat axis of the donor with the steep axisof the host cornea [16].


Consideration of donor topography may:∑ Eliminate the use of donors with abnormal

or unusual curvatures (such as high astig-matism, keratoconus, previous refractivesurgery)

∑ Allow for “harmonization” of donor andrecipient topography

10.3.1.6The Vascularized Cornea

Excessive bleeding after trephination of vascu-larized corneas with blood clots left in the ante-rior chamber may result in increased risk ofimmunologic graft reaction and peripheralanterior synechiae due to contraction. Thus, thefollowing precautions should be taken:

Before trephination the microsurgeonshould differentiate between vascularized pan-nus tissue (“plus”) and vascularized scars (“mi-nus”). Vascularized fibrous tissue between theepithelium and Bowman’s layer or the superfi-cial stroma in the case of defective Bowman’slayer can be removed easily with a hockey knife.Typically, bleeding stops after a few minuteswithout additional measures. In contrast, dis-tinct “feeder vessels” of vascularized scars maybe incised with a pointed scalpel at the limbus.Pillai et al. have proposed sophisticated kauteri-zation techniques for coagulation of afferentand efferent vessels [51]. In the case of diffusely


Fig. 10.6. An eight-line radial keratotomy marker(colored with methylene blue) may be used to facili-tate limbal centration

capillarized scars, ice-cold balanced salt solu-tion (BSS) or topical alpha-mimetic vasocon-stringent drops (such as naphazoline nitrate)may help to reduce bleeding during trephination.


∑ Removal of vascularized pannus tissue mayhelp to increase the “individually optimalgraft size”

∑ Incision or kauterization of distinct “feedervessels” of scars at the limbus may reducebleeding during trephination

10.3.1.7Keratoconus and Disabling HighAstigmatism of a Graft

Keratoconus. In keratoconus, a central roundPKP is indicated as soon as hard contact lensesare no longer tolerated. Excessively steepcorneas before surgery do not have less favor-able outcome than less deformed corneas afterPKP using the excimer laser for nonmechanicaltrephination [83].

Keratoconus eyes have larger corneas thannormal eyes and other dystrophies allowing forlarger graft diameters (typically 8.0 mm) [60].A larger graft diameter in keratoconus patientsmay help to preserve a sufficiently thick corneaat the trephination margin in the patient sincethe “cone” can be excised almost completely.Kauterization of the cone has been suggested toavoid divergent cut angles, but its effect may notbe reproducible. Thus, we do not advocate kau-terization of the cone. Kaufman has suggestednot using obturators in the case of keratoconusto prevent unintended creation of elliptical orpear-shaped openings [21].

We do not advocate centering the trephina-tion on the cone, thereby typically decenteringthe trephination with respect to the limbus. Inaddition, pupil centration may be misleadingdue to “optical displacement” of the visible pupilbecause of irregular refraction of incoming raysof light by the irregularly curved corneal sur-face in keratoconus [31].We do not advocate un-dersizing of the donor to reduce myopia, sinceirregular astigmatism is to be expected.

Due to inhomogeneous corneal thickness, anearly perforation at the site of the thinnedcornea is to be expected. This has to be takeninto account with conventional trephines toavoid inadvertent injury of the iris or even thelens.

Peripheral thinning of the host cornea, e.g.,with keratotorus (= pellucid marginal degener-ation) or Fuchs-Terrien marginal degeneration,is very rare but difficult to treat. Treatment op-tions include an eccentric semilunar lamellar/penetrating graft or an overdimensionedpreferably elliptical eccentric through-and-through graft.

Disabling High Astigmatism of a Graft. Eyeswith high disabling astigmatism after PKP areoften – but not always – associated with smalland/or decentered grafts. The re-graft should bewell centered and large enough to cut out theprevious graft entirely. However, in some casesthe previous graft-host junction cannot beexcised in toto (cf. Sect. 10.3.1.3,“Graft Size” and“Oversize”), leaving a “wedge” of the first donortissue in situ.

After second suture removal, astigmatismmay increase again and may no longer be signif-icantly different in comparison to the preopera-tive values [70].

Our own results suggest a potentially impor-tant role of the remaining second running su-ture in keeping corneal astigmatism values lowand topographic regularity high after repeatPKP in patients with high and/or irregular post-keratoplasty astigmatism. After removal of thelast suture, the curvature may change in an un-predictable and often unfavorable manner. Thepresumed original instability of the host rim,which on final suture removal may be trans-ferred to the center of the graft (“memory ef-fect”), is probably responsible for the increase inastigmatism and the increase in irregularity ofthe corneal surface. In addition, the host rim in-stability may be exacerbated by incomplete ex-cision of the previous graft-host junction in se-verely decentered first grafts. However, the exactrole of any such residual tissue has yet to beclarified.


The long-term value of so-called “intra-corneal rings” inside the graft-host junctionwith respect to stabilization of the topographyin such eyes has yet to be determined [13, 24].


∑ With keratoconus a large excision should becentered at the limbus (not the “cone”) andnon-contact laser trephination is preferredto prevent “other-than-round” recipientopenings

∑ Where repeat PKP is performed in eyeswith high and/or irregular astigmatism inclear grafts, visual rehabilitation may belimited by an increase in astigmatism andtopographic surface irregularity after re-moval of the last running suture

∑ In such eyes it may be advantageous topostpone final suture removal for as long as possible

10.3.1.8The Unstable Cornea

Unstable corneas include:1. Corneal perforations or descemtoceles typi-

cally arising from ulcerative necrotizingstromal keratitis of herpetic or bacterial ori-gin

2. Eyes after unfavorable keratorefractive sur-gery such as after radial keratotomy and ia-trogenic keratectasia after laser in-situ ker-atomileusis (LASIK)

In the “open eye” situation mechanical tre-phines may lead to compression and distortionof the cornea although a high-viscosity vis-coelastic agent is used to stabilize the anteriorchamber. Especially with large perforations thetrephine can only be used to mark the excision,the keratotomy has to be deepened with a dia-mond knife and the excision is completed withscissors. Nonmechanical laser trephination hasbeen advocated since it may allow non-contactround and elliptical trephinations (Fig. 10.7)[26]. One suggestion has been to insert atrimmed part of a soft contact lens via largeparacentesis, unrolling it inside the anteriorchamber and thus achieving a stable eye fortrephination after pressurizing the globe by in-

sertion of viscoelastic agent via paracentesis(“valve”). A larger than usual graft oversize(e.g., 0.5 mm) is recommended to avoid periph-eral synechiae in eccentric or even peripheralgrafts.

In the case of excisions involving the limbus,the scleral spur has to be preserved during(partly lamellar) trephination. In the case of pe-ripheral small perforations, an eccentric mini-keratoplasty may have immunologic advan-tages. Wide limbus-parallel perforations –typical of rheumatoid origin – may best betreated with a crescent graft. For this partly“freehand” procedure, an outer segmentaltrephination with a smaller diameter (e.g.,10 mm) is combined with an inner segmentaltrephination with a larger diameter (e.g.,16 mm). Adequate preparation of the slightlyoversized graft is best achieved from an intactdonor globe but is quite difficult using a cor-neoscleral button from the Eye Bank (protec-tion of endothelium!).


Fig. 10.7. A Descemetocele after ipsilateral autolo-gous keratoplasty for localized central herpetic scar;B eccentric elliptical triple procedure à chaud(7.0¥8.0 mm/7.1/8.1 mm, excimer laser trephination)

A

B

After excessive radial keratotomies resultingin irregular astigmatism and glare/halos due toscars in the optical field, deep epithelial plugsare typically present inside the original radialcuts for years. Instability leads to opening ofthese plugs during mechanical trephination.Certain types of circular sutures have been pro-posed before trephination. However, non-con-tact laser trephination seems to be the methodof choice for such eyes. In analogy, iatrogenickeratectasia after LASIK is prone to opening ofthe lamellar interface between the stromal bedand flap during conventional contact trephina-tion. This may result in oval host wounds anddifferent sizes of the excised button at the flapand bed levels [64]. Again, non-contact lasertrephination seems to be the method of choicefor such eyes, the incidence of which is sup-posed to increase over the next few decades.


∑ In the “open eye” situation conventionaltrephines typically only mark the host exci-sion which has to be completed freehandwith diamond knife and scissors

∑ With unstable corneas non-contact non-mechanical laser trephination has majoradvantages over conventional mechanicaltrephination

10.3.1.9The Triple Procedure

Since the introduction of the triple procedure[= simultaneous penetrating keratoplasty(PKP), extracapsular cataract extraction andimplantation of a posterior chamber intraocu-lar lens (PCIOL)] in the mid-1970s, there hasbeen an ongoing discussion among corneal mi-crosurgeons concerning the best approach (si-multaneous or sequential) for combined cornealdisease and cataract [65]. For the refractive re-sults after the triple procedure, some intraoper-ative details are crucial: trephination of recipi-ent and donor from the epithelial side withoutmajor oversize (guided trephine system or non-mechanical excimer laser trephination) shouldpreserve the preoperative corneal curvature.Graft and the PCIOL placed in the bag afterlarge continuous curvilinear capsulorhexis

should be centered along the optical axis(Fig. 10.8). If possible, performing the capsu-lorhexis under controlled intraocular pressureconditions prior to trephination may help tominimize the risk of capsular ruptures. In thecase of excessive corneal clouding, a capsu-lorhexis forceps is used via the “open sky” ap-proach. Delivery of the nucleus is achieved viathe “open sky” approach by means of manualirrigation, and removal of the lens cortex byautomated irrigation-aspiration.

The major advantage of the triple procedureis the faster visual rehabilitation achieved andless effort required for the mostly elderly pa-tients. In contrast, sequential cataract surgeryhas the potential for a simultaneous reductionof corneal astigmatism (appropriate location ofthe incision, simultaneous refractive kerato-tomies or implantation of a toric PCIOL). Dis-advantages may include the loss of graft en-dothelial cells and the theoretically increasedrisk of immunologic allograft reactions. Afterthe triple procedure, major deviations from tar-get refraction have been reported. However, in-dividual multiple regression analysis may helpto minimize this problem with appropriatemethods of trephination [77]. Since suture re-moval after PKP may result in major individualchanges of the corneal curvature, IOL power


Fig. 10.8. Well centered (1) trephination, (2) capsu-lorhexis, and (3) posterior chamber lens inside thecapsular bag after triple procedure in Fuchs’ dystro-phy (7.5/7.6 mm, excimer laser trephination with eight“orientation teeth/notches”)

calculation for the sequential approach requiresall sutures to be removed at the time of cataractsurgery. However, even after complete suture re-moval the abnormal proportions between ante-rior and posterior curvatures and/or the irregu-lar topographies after PKP may be responsiblefor marked IOL power miscalculations in the in-dividual eye [65].


∑ The postulated better prediction of refrac-tion after sequential keratoplasty andcataract surgery is opposed by a markedlydelayed visual rehabilitation

∑ We consider the triple procedure includingcataract extraction via “open sky” in gener-al anesthesia as the method of choice forcombined corneal and lens opacities

10.3.1.10Impact of Trephination on Suturing

The trephination modality may have a majorimpact on the correct placement of the first fouror eight cardinal sutures. The predominant pur-pose of the cardinal sutures is: (1) symmetrichorizontal distribution of donor tissue in the re-cipient bed, (2) good adaptation of graft andhost on Bowman’s level (external steps are to beavoided, internal steps may be tolerated in thecase of thin recipient corneas such as in pellucidmarginal degeneration or herpetic scars), and(3) stabilization of the anterior chamber forfurther homogeneous suturing.

Unintentionally other than round host open-ing may create a challenge even for the experi-enced PKP surgeon concerning the correctplacement of the second cardinal suture. Afterremoval of the cardinal sutures the quality ofthe trephination and graft positioning are majordeterminants for watertight wound closure. Thebetter the trephination, the smaller the final su-ture tension required for watertight wound clo-sure after removal of the cardinal sutures. Thesmaller the final suture tension, the better thevisual acuity as long as the sutures are in place.Generally, in cases where Bowman’s layer is intact, a 16-bite double-running diagonal cross-stitch suture (10-0 nylon) according to Hoff-mann (Fig. 10.9) is preferred. The more rapid

visual rehabilitation with these sutures in placein contrast to single sutures is due to a more reg-ular corneal topography avoiding cornea plana.


∑ The better the trephination the easierwatertight wound closure is achieved

∑ Inadequately high suture tension to achievewatertight wound closure may deterioratethe regularity of the topography after PKPand delay visual recovery

10.3.2Conventional Mechanical Trephines(Table 10.5)

In 1886 Arthur von Hippel was the first to use a mechanical clock-watch driven trephine(Fig. 10.10) for transplantation of a lamellarcorneal graft from a rabbit to a human [79]. Thesame trephine was used by Eduard Zirm for hisfirst successful PKP in a patient in 1905 [84].

Conventional mechanical trephination isassociated with deformation of corneal tissueincluding a distortion of the cut margin withrough-cut edges as a consequence of axial andradial forces induced by the trephine. The cutangle deviates from the perpendicular and itmay be different in donor and recipient, espe-cially if the donor trephination is undertakenfrom the endothelial side. The fitting of the


Fig. 10.9. Typical double running 10-0 nylon cross-stitch suture with 8 bites each (according to Hoff-mann [17]) in keratoconus (8.0/8.1 mm, excimer lasertrephination with eight “orientation teeth/notches”)


Table 10.5. Characteristics of mechanical trephines

Type Geuder Moria GTS Hessburg- Asmotom Micro-Keratron (Hanna) (Krumeich) Barron (Gliem & (discontinued) Franke)

Motorized cutter Yes No No No Yes

Vacuum fixation No Yes Yes Yes Doublefor recipient (limbus) (limbus) (cornea)

Cutter feed No No No No Yes

Depth adjustment No Yes Yes Limited Yes

Auto-retract No No No No Yes

Anterior chamber Yes Yes Yes Possible Nomaintainer required for donor

Automation No No No No Yes

Fig. 10.10 A, B. Mechanical trephines. A Arthur von Hippel’s clock-watch driven trephine. B “Modern”mechanical trephines (motor trephine, Lochpfeiffentrepan, hand-held trephine [39])

A B

Table 10.6. Trephines used in Germany in the year 2002 for 4583 penetrating keratoplasties (German Kerato-plasty Registry Erlangen) (122 institutions contributed) [5]

GTS Manual Barron Motor trephine Asmotom Excimer laser Unknown

Donor 1555 1040 716 415 393 313 151

% 33.9 22.7 15.6 9.1 8.6 6.8 3.3

Recipient 1570 818 745 640 346 313 151

% 34.3 17.8 16.3 13.9 7.6 6.8 3.3

donor tissue into the malleable recipient corneais extremely difficult to achieve in a perfectlysymmetric fashion. After suturing the incon-gruent cut edges in order to achieve watertightwound closure, wound healing may causemarked distortion of the surface topographyafter suture removal due to this “vertical tilt.” Inaddition, asymmetric cardinal suture place-ment may result in unequal donor tissue distri-bution in the host wound,particularly if the sec-ond cardinal suture is not placed exactlyopposite to the first (“horizontal torsion”) [42].

A questionnaire was sent to all German ker-atoplasty surgeons in 2002 asking for their pre-ferred technique of trephination. As outlined in Table 10.6 for recipient trephination, mostsurgeons use the GTS (34.3%), the hand-heldtrephine (17.8%) or the Hessburg-Barrontrephine (16.3%). Motor trephines are usedmore rarely and the laser trephination has stillnot entered many operating theaters because itis bulky and expensive. As many as 12% of allprocedures were performed with differenttrephine systems for donor and recipient [5]!

10.3.2.1Freestanding Blade/Hand-Held Trephines

Hand-held trephines are available in a widerange of diameters from very small (e.g.,1.5 mm) to very large (e.g., 16.0 mm). Hand-heldtrephines may be dull with reduced visual con-trol under the operating microscope despite re-cent improvements [39]. Thus, centration maybe a problem. Typically, the donor is punchedfrom the endothelial side (Lochpfeiffentre-pan). Francheschetti-type freestanding blades(Fig. 10.11) seem to create more reproduciblecuts than other hand-held trephines [72, 76].

10.3.2.2Motor Trephines (Mikro-Keratron, Asmotom)

Mikro-Keratron. The Geuder Micro-Keratrontrephine is a non-automated motor-driventrephine system for PKP. The depth of the cut isnot preadjustable, so that this trephine systemhas no impact on lamellar keratoplasty. Rota-tion (variable speed) may be started andstopped by pressing down and releasing a footpedal. Different blades mounted on the unit al-low for a wide range of trephination diameters.To trephine the donor cornea from the epithe-lial side, the tissue has to be mounted into anartificial anterior chamber maintainer. Motortrephine rotation may lead to “shifting” of thetrephine within the corneal stroma.

Asmotom. The Asmotom ATS is an automatedtrephine system for PKP. The trephination ofpatient and donor eyes as well as corneoscleraldisks is performed with separate instrumenta-tion sets. For non-perforating cuts the cuttingdepth is preadjustable with offset rings for thepatient. The cutter sets provided by the distrib-utors include five different diameters (6.0–8.2 mm). The ATS uses an innovative double fix-ation design.Vacuum is applied to both the cen-tral and the peripheral section of the cornea.The trephine rotates between the two concen-tric areas of fixation, using an automatic feed.Once the pre-set depth is reached, the cutter re-tracts back into its initial position, holding on tothe separated central portion, until vacuum isreleased. The ATS marker facilitates the center-ing of the trephination cut to the cornea. Thesystem does not require an artificial chambermaintainer for graft trephination.


Fig. 10.11. Francheschetti-type freestanding bladesare available in a wide range of diameters

10.3.2.3Suction Trephines (Hessburg-Barron)

The classical Hessburg-Barron trephine (HBT)has been on the market for over 25 years. TheHBT vacuum trephine is an easy to handle sin-gle-use product. The suction is applied to theperipheral cornea. The depth of the lamellartrephination can be predicted to a certain de-gree. One full rotation is presumed to achieve250 mm of corneal depth. Perforation is typical-ly limited to one-third to one-half of the cir-cumference of the excision. The recipienttrephine has cross-hairs for centration. No ob-turator is applied (Fig. 10.12A). The Hessburg-Barron trephine leads to divergent cut anglesand a larger diameter of the hole at the level ofDescemet’s membrane [72, 76].

In the classic version the donor is punchedfrom the endothelial side with the aid of a suc-tion device for fixating the donor epithelial sidedown. Tilt is avoided by four metal rods in the

periphery of the blade-containing part and fourcorresponding peripheral holes in the suction-containing part (Fig. 10.12B). In addition, foursmall holes inside the cut area which are coloredbefore the corneoscleral button is placed insidegive a reference with respect to the first four car-dinal sutures. The donor is typically oversizedby 0.25 mm [12].

Recently, a single-use artificial anteriorchamber has been available, to create donortrephination from the epithelial side using therecipient trephine for donor trephination first.

10.3.2.4Guided Trephines (GTS, Hanna)

The guided trephines result in the best cut qual-ities possible with mechanical trephines [72,76]. These new generation suction trephinessuch as the Hanna trephine [80] and theKrumeich trephine (“guided trephine system,”GTS) [4, 23] are preferred over the Hessburg-Barron trephine because they stabilize the globeby suction at the limbus – not the peripheralcornea. Thus – at least theoretically – the cutangles should be parallel to the optical axis, thedimensions for donor and recipient should beequal and, therefore, no graft oversize is re-quired [50]. Overall, handling of both trephinesrequires a special introduction to the micro-surgeon and the staff before application inpatients.

GTS (Fig. 10.13). The Krumeich guided trephinesystem (GTS) is designed for PKP, lamellar ker-atoplasty, and circular keratotomy. The GTS canbe used with and without an obturator prevent-ing ballooning of the excised tissue into thetrephine opening.

Advantages of the GTS include: (1) trephina-tion of donor and recipient from the epithelialside using an artificial anterior chamber, (2)pre-defined depth of trephination, e.g., forlamellar procedures, and (3) in experiencedhands through-and-through trephination with-out the necessity of cut completion with scissorscan be achieved.

Potential disadvantages of the GTS include:(1) it is difficult to apply in patients with narrowlid fissure or deeply set eyes with prominent or-


Fig. 10.12 A, B. Hessburg-Barron suction trephine.A Recipient trephine with cross-hairs for centration;B Donor trephination is performed from the endo-thelial side

B

A

bital bones (which is not an uncommon issue inkeratoconus), preexisting filtering blebs or con-junctival chemosis, (2) centration is difficultdue to the limited view, (3) injury if the iris andlens are not securely prohibited, and (4) eccen-tric mini-keratoplasty with a small diameter(e.g., 4 mm) cannot be accomplished.

Hanna Trephine (Fig. 10.14). The Hanna (Mo-ria) trephine system is one of the most advancedtrephines which is designed to create a properdonor/recipient match. The Hanna trephine at-taches firmly to the eye through suction appliedto the limbal conjunctiva. Uniform support overthe whole cornea during trephination preventscorneal vaulting. From a fully retracted position,the blade rotates while descending to a presetdepth, after which the blade rotates without fur-ther descent, cutting the displaced tissue and creating a uniform incision. The Hannatrephine in combination with the artificial ante-rior chamber allows the surgeon to trephine

both the recipient and the donor cornea fromthe epithelial side, thus reducing shape disparity.In the original version the donor trephinationwas performed from the endothelial side [81].


∑ If conventional trephines are used it is rec-ommended to use at least the same systemwith trephination of the donor from theepithelial side using an artificial anteriorchamber for placement of the corneoscleralbutton from the Eye Bank

∑ The trephine should be as sharp as possible

10.3.3Nonmechanical Laser Trephination

Hypothesizing that the properties of the woundbed are much more important for the final “all-suture-out” astigmatism and the final opticalperformance of the graft than various types ofsuture techniques or methods of suture adjust-ment, we have developed and optimized thetechnique of nonmechanical corneal trephina-tion since 1986.


Fig. 10.13. The Krumeich guided trephine system(GTS) is designed for PKP, lamellar keratoplasty, andcircular keratotomy. In patients, the GTS can be usedwith and without an obturator preventing ballooningof the excised tissue into the trephine opening

Fig. 10.14. The Hanna (Moria) trephine system. Inpatients this trephine attaches firmly to the eyethrough suction applied to the limbal conjunctiva.The Hanna trephine in combination with the artifi-cial anterior chamber allows the surgeon to trephineboth the recipient and the donor cornea from theepithelial side

10.3.3.1The 193-nm Excimer Laser

Since 1989 more than 1650 human eyes havebeen treated successfully with the MeditecMEL60 excimer laser (Fig. 10.15). Keratoconushas been by far the leading indication (around37%) for PKP with this non-contact technique(Table 10.7). For donor trephination from theepithelial side an artificial anterior chamber isused [41, 42, 58, 66].

Technique (Fig. 10.16). Before starting trephi-nation, the limbus is centered on the perpendi-cular HeNe aiming beam in donor and patientto ensure a reproducible position of the eye rel-ative to the laser and symmetric cut angles overthe entire circumference without tilt. The hori-zontal positioning of the limbal plane can becontrolled using the focusing device of the laserat 3, 6, 9, and 12 o’clock at the limbus beforefocusing the laser at the trephination edge (“tri-angulation”). “Horizontal torsion” of the graft

may be reduced by employing eight orientationteeth at the donor trephination margin andeight corresponding notches in the recipientbed (a technique which allows the use of eightsymmetric cardinal sutures) [2].

For donor trephination from the epithelialside using the 193-nm excimer laser MEL60


Fig. 10.15. Principle of excimer laser trephination in donor and recipient (schematic drawing, sagittal view)

Table 10.7. Indications for 1656 consecutive non-mechanical excimer laser keratoplasties (06/1989 to04/2005 in Erlangen)

Keratoconus 607 (36.7%)

Fuchs’ dystrophy 323 (19.5%)

Bullous keratopathy 275 (16.6%)

Avascular scars 181 (10.9%)

Graft failure 77 (4.6%)

Corneal ulcer 64 (3.9%)

Stromal dystrophies 48 (2.9%)

Disabling astigmatism 40 (2.4%)

Others 41 (2.5%)

(Carl Zeiss Meditec, Jena, Germany), a circularround metal aperture mask (diameter 5.6–8.6 mm, central opening 3.0 mm for centration,thickness 0.5 mm, weight 0.2 g, eight orientationteeth 0.15¥0.3 mm) is positioned on a cor-neoscleral button (16 mm diameter) fixed in an artificial anterior chamber (Polytech, Ross-dorf, Germany) under microscopic control(Fig. 10.16A, B). The pressure within the artifi-cial anterior chamber is adjusted to 22 mmHg.An automated rotation device for the artificialanterior chamber is used.

For recipient trephination exclusively per-formed with the manually guided excimer laser,a corresponding metal mask is used (diameter12.9 mm, central opening 5.5–8.5 mm), thickness0.5 mm, weight 0.4 g, eight orientation notches0.15¥0.3 mm (Fig. 10.16C, D). Before starting the

trephination, centration relative to the limbus isachieved by lining up the eight notches with theeight lines of a radial keratotomy marker undermicroscopic control (Fig. 10.6).

Advantages (Table 10.8). The main advantageof this novel laser cutting technique performedfrom the epithelial side in donor and recipient isthe avoidance of mechanical distortion duringtrephination, resulting in smooth cut edges(Fig. 10.17A) which are congruent in donor andpatient, potentially reducing “vertical tilt” [33].Such cut edges in combination with “orientationteeth” (Fig. 10.17B) at the graft margin [2] andcorresponding notches at the recipient marginfor symmetric positioning of the eight cardinalsutures minimize “horizontal torsion,” thuspotentially improving the optical performance


Fig. 10.16 A–D. Nonmechanical trephination usingthe 193-nm excimer laser in combination with metalmasks with “orientation teeth/notches.” A Curveddonor mask on top of corneoscleral button fixed in amodified Krumeich artificial anterior chamber;B metal donor mask with eight “orientation teeth”;

C laser arm and joystick for recipient trephination;D metal recipient mask with eight “orientation notch-es” on top of patient’s cornea.A 1.5¥1.5-mm laser spotis guided along the inner edge of the mask, half of thebeam on the mask and half of it on the cornea

A B

DC

after transplantation [42]. Furthermore, recipi-ent and donor decentration may be reduced [30,61]. The use of metal masks allows for arbitraryshapes of the trephination [28].

These favorable impacts on major intra-operative determinants of post-keratoplastyastigmatism (cf. Table 10.2) result in lowerkeratometric astigmatism, higher topographicregularity and better visual acuity after sutureremoval. After sequential removal of a doublerunning suture, keratometric astigmatism in-creased in 80% of eyes with conventionaltrephination, but further decreased in 52% ofeyes with laser trephination [58]. In addition toless blood-aqueous barrier breakdown duringthe early postoperative time course after PK[26], laser trephination induces neither cataractformation nor higher endothelial cell loss of thegraft. Likewise, the rates of immunologic graftrejection and secondary ocular hypertensionare comparable using either technique. In addi-tion, trephination of an unstable cornea, such asin (pre-)perforated corneal ulcers or after RK orLASIK, is facilitated [64].

Practical Considerations for the Microsurgeon[66]. The longer trephination time of around6 min for the donor and around 4 min for the re-cipient are by far compensated for by practicaladvantages for the microsurgeon during thesubsequent course of surgery: (1) injuries of in-


Table 10.8. Advantages of nonmechanical trephination with the 193-nm excimer laser along metal maskswith “orientation teeth/notches” [41, 42, 58, 66]

1. No trauma to intraocular tissues

2. Avoid deformation and compression of tissue during trephination

3. Reduction of horizontal torsion (“Erlangen orientation teeth/notches”)

4. Reduction of vertical tilt (congruent cut edges)

5. Reduction of host and donor decentration

6. Feasibility of “harmonization” of donor and host topography

7. Reduction of anterior chamber inflammation early after PKP

8. Reduction of astigmatism after suture removal

9. Higher regularity of corneal topography

10. Significantly better visual acuity with spectacle correction

11. Feasibility of trephination with unstable cornea (e.g.,“open eye”, descemetocele, after radialkeratotomy, iatrogenic keratectasia after LASIK)

12. Arbitrary shape (e.g., elliptical) [28]

Fig. 10.17 A, B. Donor trephination immediatelybefore perforation. A Histologic view with smoothalmost perpendicular cut edge; B macroscopic viewwith smooth cut surfaces and “orientation teeth”

A

B

traocular structures are impossible with thelaser – even in beginner’s hands – since the ab-lation stops as soon as aqueous humor fills thetrephination groove after focal perforation. (2)The need for completion of the cut by scissors isreduced to a minimum. (3) The localization of the first eight cardinal sutures is unequivo-cally given by the “orientation teeth/notches”(Fig. 10.18). (4) Crescent-shaped tissue deficitsat the graft-host junction (e.g., at other thanround recipient openings in keratoconus) areavoided, thus achieving a latent watertightwound closure often as soon as after four cardi-nal sutures. (5) During further suturing the an-terior chamber tends to remain stable. (6) Thefinal double running suture needs very littletension to keep a watertight wound after re-moval of the eight cardinal sutures. (7) There-fore, only very rarely are additional single su-

tures with adverse effects on graft topographyrequired at the end of surgery. (8) In addition,the so-called “barrel-top formation”at the prox-imal suture endings inducing a relative corneaplana and delaying optical rehabilitation can beavoided. (9) After removal of lid speculum andfixation sutures, the use of a Placido’s disk oftenenables an almost round projection image to beachieved during intraoperative suture adjust-ment.


∑ Nonmechanical trephination using the 193-nm excimer laser along metal maskshas improved functional outcome after PKPwith all-sutures-out

∑ The application of excimer lasers allowscontrolled trephination of unstable corneassuch as perforated ulcers or iatrogenic keratectasia after LASIK

10.3.3.2The 2.94-µm Erbium:YAG Laser

The erbium:YAG laser was investigated to im-prove handling, reduce acquisition and mainte-nance costs, and provide solid state laser safetybut keep the morphological advantages of theexcimer laser trephination [1]. However, shrink-age effects due to thermal damage of the cutedges especially in the free-running but evenwith Q-switched laser pulses are major draw-backs of this infrared laser [69]. The inducedthermal damage of the Q-switched mode er-bium:YAG laser has been detected to be around2–15 mm, in comparison to only 200 nm usingthe excimer laser [54, 68].


∑ The erbium:YAG laser will probably not substitute the excimer laser for non-mechanical trephination in the near futurewithout a loss of advantages


Fig. 10.18. Correct position of second cardinalsuture (arrow) is facilitated by orientation tooth(donor) and corresponding notch (host)

10.3.3.3The Femtosecond Laser

In contrast to the excimer laser, which allowsonly surface ablation, the femtosecond (=10–15 s) laser allows the cornea to be cut withinthe stroma, enabling truly three-dimensionalcuts without opening the eye and without ther-mal damage. No masks but an ultra-fast eyetracking system is required. There is no signifi-cant tissue loss to be compensated. For PKP es-pecially in keratconus a non-contact approachof laser application is favored to avoid deforma-tion.

Self-sealing keratoplasty wounds would be amajor step towards rapid visual rehabilitationin PKP. Various kinds of nut-and-bolt configu-rations to fit in the donor including “orientationteeth” of the graft in the recipient bed are feasi-ble using a femtosecond laser. We have intro-duced an inverse mushroom shaped trephina-tion with the larger diameter of the graft at thelevel of Descemet’s membrane (Fig. 10.19).Vari-ation of the diameter of the “stipe” and the “cap”may help to produce the best individual com-promise between the amount of transplantedendothelium and distance to limbal vessels andresistance to intraocular pressure [67].

In addition, posterior lamellar keratoplasty(PLKP) can be performed more easily with afemtosecond laser [63].


∑ Femtosecond laser application is the “excitement of tomorrow” in microsurgeryof the cornea

∑ New nut-and-bolt type variants for poten-tially self-sealing donor/host appositionsare on the horizon, offering a promising approach towards minimally invasive “no-stitch keratoplasty”

10.4Concluding Remarks

Today, expectations concerning the outcomeafter penetrating keratoplasty are not only re-stricted towards achieving a clear graft. Theonly criterion that counts for the patient is goodvision preferably without the need for contactlenses but with an easily tolerable need for cor-rection using spectacles. Therefore, transplantmicrosurgeons should not only consider all themeans available to prevent high or irregularpost-PKP astigmatism. Due to the lack of pre-dictability of the refractive result in an individ-ual patient after PKP, they should also familiar-ize themselves with the surgical techniques forcorrecting refractive errors after PKP in orderto achieve the individually best outcome for agiven patient.

References

1. Behrens A, Küchle M, Seitz B, Langenbucher A,Kus MM,Amann T, Schlötzer-Schrehardt U, Rum-melt C, Naumann GOH (1998) Stromal thermaleffects induced by nonmechanical (2.94 mm)Er:YAG laser corneal trephination.Arch Ophthal-mol 116:1342–1348

2. Behrens A, Seitz B, Küchle M, Langenbucher A,Kus MM, Amann T, Rummelt C, Naumann GOH(1999) “Orientation teeth”in nonmechanical lasercorneal trephination: 2.94-mm Er:YAG laser vs.193-nm ArF excimer laser. Br J Ophthalmol83:1008–1012

3. Belmont SC, Troutman RC, Buzard KA (1993)Control of astigmatism aided by intraoperativekeratometry. Cornea 12:397–400


Fig. 10.19. Macrophotograph of inverse mushroomshaped trephination using a femtosecond laser [67]

4. Belmont SC, Zimm JL, Storch RL, Draga A, Trout-man RC (1993) Astigmatism after penetratingkeratoplasty using the Krumeich Guided Tre-phine System. Refract Corneal Surg 9:250–254

5. Blüthner K, Seitz B (2004) Verwendete Trepansys-teme bei der perforierenden Keratoplastik – einBericht aus dem Deutschen Keratoplastikregister2002. 102. Tagung der Deutschen Ophthalmolo-gischen Gesellschaft (Vortrag), Berlin

6. Brightbill FS (1993) Corneal surgery. In: BrightbillFS (ed) Theory, technique and tissue. Mosby,St. Louis

7. Busin M (2003) A new lamellar wound configura-tion for penetrating keratoplasty surgery. ArchOphthalmol 121:260–265

8. Cohen KL, Tripoli NK, Pellom AC, Kupper LL,Fryczkowski AW (1984) Effect of tissue fit oncorneal shape after transplantation. Invest Oph-thalmol Vis Sci 25:1226–1231

9. Cohen KL, Holman RE, Tripoli NK, Kupper LL(1986) Effect of trephine tilt on corneal buttondimensions. Am J Ophthalmol 101:722–725

10. Dave AS, McCulley JP (1994) Demonstration offeasibility of application of a portable keratome-ter to cadaveric donor corneas. Cornea 13:379–382

11. Davis EA, Azar DT, Jakobs FM, Stark WJ (1998)Refractive and keratometric results after thetriple procedure. Experiences with early and latesuture removal. Ophthalmology 105:624–630

12. Duffin M, Olson RJ, Ohrloff C (1984) Analysis ofthe Hessburg-Barron vacuum trephine. Oph-thalmic Surg 15:51–54

13. Ehrich D, Duncker GI (2004) [The use of intra-corneal rings in penetrating keratoplasty.] KlinMonatsbl Augenheilkd 221:92–95

14. Filatov V, Alexandrakis G, Talamo JH, Steinert RF(1996) Comparison of suture-in and suture-outpostkeratoplasty astigmatism with single run-ning suture or combined running and interrupt-ed sutures. Am J Ophthalmol 122:696–700

15. Greenfield DS, Jivanjee SQ, Raizman MB (1995)The effect of trephination on corneal topography.Ophthalmic Surg Lasers 26:560–567

16. Grütters G, Reichelt JA, Nölle B (2001) [Astigma-tism-oriented perforating keratoplasty. A possi-bility for minimizing postoperative astigma-tism?] Ophthalmologe 98:397–401

17. Hoffmann F (1976) [Suture technique for perfo-rating keratoplasty.] Klin Monatsbl Augenheilkd169:584–590

18. Hoppenreijs VPT, Van Rij G, Beekhuis WH, Rijn-eveld WJ, Rinkel-Van Driel E (1993) Causes ofhigh astigmatism after penetrating keratoplasty.Doc Ophthalmol 85:21–34

19. Javadi MA, Mohammadi MJ, Mirdehghan SA, Saj-jadi SH (1993) A comparison between donor-re-cipient corneal size and its effect on the ultimaterefractive error induced in keratoconus. Cornea12:401–405

20. Judge D, Gordon L, Van der Zwaag R, Wood TO(1990) Refractive versus keratometric astigma-tism postkeratoplasty. Refract Corneal Surg6:174–178

21. Kaufman HE, McDonald MB, Barron BA, WilsonSE (1992) Corneal and refractive surgery. JB Lip-pincott, Philadelphia

22. Krumeich J, Binder PS, Knülle A (1988) The theo-retical effect of trephine tilt on postkeratoplastyastigmatism. CLAO J 14:213–219

23. Krumeich JH, Grasl MM, Binder PS, Knülle A(1990) Geführtes Trepansystem für perforierendeKeratoplastiken. In: Freyler H, Skorpik C, Grasl M(eds) 3. Kongreß der Deutschen Gesellschaft fürIntraokularlinsenimplantation. Springer, Vienna,pp 450–456

24. Krumeich JH, Daniel J (1999) Perforating kerato-plasty with an intracorneal ring. Cornea 18:277–281

25. Küchle M, Nguyen NX, Seitz B, Langenbucher A,Naumann GOH (1998) Blood-aqueous barrier af-ter mechanical or nonmechanical excimer lasertrephination in penetrating keratoplasty. Am JOphthalmol 125:177–181

26. Küchle M, Seitz B, Langenbucher A, NaumannGOH (1999) Nonmechanical excimer laser pene-trating keratoplasty for perforated or pre-descemetal corneal ulcers. Ophthalmology 106:2203–2209

27. Kus MM, Seitz B, Langenbucher A, NaumannGOH (1995) The effects of scleral and limbal fixa-tion on the topography of corneoscleral buttonsin an artificial anterior chamber. Invest Ophthal-mol Vis Sci 36(4 Suppl):S980

28. Lang GK, Naumann GOH, Koch JW (1990) A newelliptical excision for corneal transplantation us-ing an excimer laser. Arch Ophthalmol 108:914–915

29. Langenbucher A, Seitz B, Kus MM,Vilchis E, Nau-mann GOH (1996) [Regularity of corneal topo-graphy after penetrating keratoplasty – com-parison between nonmechanical (excimer laser193 nm) and mechanical trephination.] KlinMonatsbl Augenheilkd 208:450–458

30. Langenbucher A, Seitz B, Kus MM,Vilchis E, Nau-mann GOH (1998) Graft decentration in pene-trating keratoplasty – nonmechanical trephina-tion with the excimer laser (193 nm) versus themotor trephine. Ophthalmic Surg Lasers 29:106–113

References 149

31. Langenbucher A, Kus MM, Neumann J, Seitz B(1999) [Calculating the localization and dimen-sion of the real pupil in keratoconus with ray tracing of corneal topography data.] KlinMonatsbl Augenheilkd 215:163–168

32. Langenbucher A, Seitz B, Nguyen NX, NaumannGOH (2002) Graft endothelial cell loss after non-mechanical penetrating keratoplasty depends ondiagnosis: a regression analysis. Graefes ArchClin Exp Ophthalmol 240:387–392

33. Langenbucher A, Seitz B, Kus MM, NaumannGOH (1998) [Transplant vertical tilt after perfo-rating keratoplasty – comparison between non-mechanical trepanation with excimer laser andmotor trepanation.] Klin Monatsbl Augenheilkd212:129–140

34. Lieberman DM, Troutman RC (1992) The theoryof “other than round” cornea transplantation.Refract Corneal Surg 8:97–100

35. Lin DT, Wilson SE, Reidy JJ, Klyce SD, McDonaldMB, Insler MS, Kaufman HE (1990) Topographicchanges that occur with 10-0 running suture re-moval following penetrating keratoplasty. RefractCorneal Surg 6:21–25

36. Mader TH, Yuan R, Lynn MJ, Stulting RD, WilsonLA, Waring II GO (1993) Change in keratometricastigmatism after suture removal more than oneyear after penetrating keratoplasty. Ophthalmol-ogy 100:119–127

37. Melles GR, Eggink FA, Lander F, Pels E, RietveldFJ, Beekhuis WH, Binder PS (1998) A surgicaltechnique for posterior lamellar keratoplasty.Cornea 17:618–626

38. Musch DC, Meyer RF, Sugar A (1988) The effect ofremoving running sutures on astigmatism afterpenetrating keratoplasty. Arch Ophthalmol 106:488–492

39. Naumann GOH (1972) [Simple keratoplastytrephine to be used under the operating micro-scope.] Klin Monatsbl Augenheilkd 161:708

40. Naumann GOH, Sautter H (1991) Surgical proce-dures on the cornea. In: Blodi FC, Mackensen G,Neubauer H (eds) Surgical ophthalmology 1.Springer, Berlin, pp 433–497

41. Naumann GOH, Seitz B, Lang GK, LangenbucherA, Kus MM (1993) [193 excimer laser trepanationin perforating keratoplasty. Report of 70 pa-tients.] Klin Monatsbl Augenheilkd 203:252–261

42. Naumann GOH (1995) The Bowman lecture(Number 56) 1994. Part II: Corneal transplanta-tion in anterior segment diseases. Eye 9:395–421

43. Naumann GOH (1995) Iris ischaemia followingpenetrating keratoplasty for keratoconus (Urrets-Zavalia syndrome). Cornea 14:618

44. Naumann GOH, Rummelt V (1996) Block exci-sion of tumors of the anterior uvea. Ophthalmol-ogy 103:2017–2028

45. Olson RJ (1983) Modulation of postkeratoplastyastigmatism by surgical and suturing techniques.Int Ophthalmol Clinics 23(4):137–151

46. Pallikaris I (1980) [Preoperative Placido photog-raphy in keratoconus and its meaning in conjunc-tion with postoperative astigmatism.] Albrechtvon Graefes Arch Klin Exp Ophthalmol 213:87–99

47. Perl T, Charlton KH, Binder PS (1981) Disparatediameter grafting.Astigmatism, intraocular pres-sure and visual acuity. Ophthalmology 88:774–780

48. Perlman EM (1981) An analysis and interpreta-tion of refractive errors after penetrating kerato-plasty. Ophthalmology 88:39–45

49. Pfister RR, Breaud S (1983) Aphakic refractivepenetrating keratoplasty using newborn donorcorneas. A preliminary report on an alternativeapproach to refractive correction. Ophthalmolo-gy 90:1207–1212

50. Pflugfelder SC, Roussel TJ, Denham D, Feuer W,Mandelbaum S, Parel JM (1992) Photogrammet-ric analysis of corneal trephination. Arch Oph-thalmol 110:1160–1166

51. Pillai CT, Dua HS, Hossain P (2000) Fine needlediathermy occlusion of corneal vessels. InvestOphthalmol Vis Sci 41:2148–2153

52. Price FW Jr, Whitson WE, Johns S, Gonzales JS(1996) Risk factors for corneal graft failure.J Refract Surg 12:134–143

53. Reinhard T, Böhringer D, Hüschen D, Sundmach-er R (2002) [Chronic endothelial cell loss of thegraft after penetrating keratoplasty: influence ofendothelial cell migration from graft to host.]Klin Monatsbl Augenheilkd 219:410–416

54. Sarhan AR, Dua HS, Beach M (2000) Effect of dis-agreement between refractive, keratometric, andtopographic determination of astigmatic axis onsuture removal after penetrating keratoplasty. BrJ Ophthalmol 84:837–841

55. Sauer R, Seitz B, Mardin C, Langenbucher A,Hofmann-Rummelt C, Viestenz A, Küchle M,Naumann GOH (2003) [Impact of intracameralpressure on donor cut angles in nonmechanicalEr:YAG laser trephination for penetrating kerato-plasty.] Klin Monatsbl Augenheilkd 220:396–403

56. Seitz B, Langenbucher A, Kus MM, NaumannGOH (1994) Consideration of donor and recipi-ent corneal topography improves graft alignmentin penetrating keratoplasty. Invest OphthalmolVis Sci 35(4, Suppl):1879

57. Seitz B, Behrens A, Langenbucher A, Kus MM,Naumann GOH (1998) Experimental 193-nm ex-cimer laser trephination with divergent cut an-gles in penetrating keratoplasty. Cornea 17:410–416


58. Seitz B, Langenbucher A, Kus MM, Küchle M,Naumann GOH (1999) Nonmechanical cornealtrephination with the excimer laser improvesoutcome after penetrating keratoplasty. Ophthal-mology 106:1156–1165

59. Seitz B, Langenbucher A, Naumann GOH (2000)Astigmatismus bei Keratoplastik. In: Seiler T (ed)Refraktive Chirurgie. Enke, Stuttgart, pp 197–252

60. Seitz B, Langenbucher A, Zagrada D, Budde W,Kus MM (2000) [Corneal dimensions in varioustypes of corneal dystrophies and their effect onpenetrating keratoplasty.] Klin Monatsbl Augen-heilkd 217:152–158

61. Seitz B, Langenbucher A, Meiller R, Kus MM(2000) [Decentration of donor cornea in me-chanical and excimer laser trephination for pene-trating keratoplasty.] Klin Monatsbl Augenheilkd217:144–151

62. Seitz B, Langenbucher A, Küchle M, NaumannGOH (2003) Impact of graft diameter on cornealpower and the regularity of postkeratoplastyastigmatism before and after suture removal.Ophthalmology 110:2162–2167

63. Seitz B, Langenbucher A, Hofmann-Rummelt C,Schlötzer-Schrehardt U, Naumann GOH (2003)Nonmechanical posterior lamellar keratoplastyusing the femtosecond laser (femto-PLAK) forcorneal endothelial decompensation. Am J Oph-thalmol 136:769–772

64. Seitz B, Rozsival P, Feuermannova A, Langen-bucher A, Naumann GOH (2003) Penetrating ker-atoplasty for iatrogenic keratoconus after repeatmyopic laser in situ keratomileusis: histologicfindings and literature review. J Cataract RefractSurg 29:2217–2224

65. Seitz B, Langenbucher A, Viestenz A, Dietrich T,Küchle M, Naumann GOH (2003) [Cataract andkeratoplasty – simultaneous or sequential sur-gery?] Klin Monatsbl Augenheilkd 220:326–329

66. Seitz B, Langenbucher A, Nguyen NX, Kus MM,Küchle M, Naumann GOH (2004) [Results of thefirst 1000 consecutive elective nonmechanicalkeratoplasties with the excimer laser – a prospec-tive study over more than 12 years.] Ophthal-mologe 101:478–488

67. Seitz B, Brünner H, Viestenz A, Hofmann-Rum-melt C, Schlötzer-Schrehardt U, Naumann GOH,Langenbucher A (2005) Inverse mushroom-shaped nonmechanical penetrating keratoplastyusing a femtosecond laser. Am J Ophthalmol139:941–944

68. Stojkovic M, Küchle M, Seitz B, Langenbucher A,Viestenz A, Viestenz A, Hofmann-Rummelt C,Schlötzer-Schrehardt U, Naumann GOH (2003)Nonmechanical Q-switched erbium:YAG lasertrephination for penetrating keratoplasty: Exper-imental study on human donor corneas. ArchOphthalmol 110:2162–2167

69. Stojkovic M, Seitz B, Küchle M, Langenbucher A,Viestenz A, Viestenz A, Hofmann-Rummelt C,Naumann GOH (2003) Corneal shrinkage in-duced by nonmechanical Q-switched erbium:YAGlaser trephination for penetrating keratoplasty.Graefes Arch Clin Exp Ophthalmol 241:667–672

70. Szentmáry N, Seitz B, Langenbucher A, NaumannGOH (2005) Repeat keratoplasty for correction ofhigh or irregular postkeratoplasty astigmatism inclear corneal grafts. Am J Ophthalmol 139:826–830

71. Terry AM, Ousley PJ (2001) Endothelial replace-ment without surface corneal incisions or su-tures: topography of the deep lamellar endothe-lial keratoplasty procedure. Cornea 20:14–18

72. Tilanus M,van Rij G (1987) An experimental com-parison of three methods for trephination of thecornea and the consequent variations in the con-figurations of the trephine openings. GraefesArch Clin Exp Ophthalmol 225:50–52

73. Toker E, Seitz B, Langenbucher A, Dietrich T, Nau-mann GOH (2003) Penetrating keratoplasty forendothelial decompensation in eyes with buph-thalmos. Cornea 22:198–204

74. Troutman RC, Buzard KA (1992) Corneal astig-matism: etiology, prevention, and management.Mosby-Year Book, St. Louis, pp 349–451

75. Van Rij G, Cornell FM, Waring GO III, Wilson LA,Beekhuis H (1985) Postoperative astigmatism af-ter central vs eccentric penetrating keratoplas-ties. Am J Ophthalmol 99:317–320

76. Van Rij G, Waring GO III (1988) Configuration ofcorneal trephine opening using five differenttrephines in human donor eyes.Arch Ophthalmol106:1228–1233

77. Viestenz A, Seitz B, Langenbucher A (2005) IOLpower prediction for TRIPLE procedures inFuchs’ dystrophy using multiple regressionanalysis. Acta Ophthalmol Scand 83:312–315

78. Villacriz E, Rife L, Smith RE (1987) Oval hostwounds and postkeratoplasty astigmatism.Cornea 6:181–184

79. Von Hippel A (1886) Über Transplantationen derKornea. Ber Ophthalmol Ges Heidelberg 18:54

80. Waring GO III, Hanna KD (1989) The Hanna suc-tion punch block and trephine system for pene-trating keratoplasty. Arch Ophthalmol 107:1536–1539

81. Wilbanks GA, Cohen S, Chipman M, Rootman DS(1996) Clinical outcomes following penetratingkeratoplasty using the Barron-Hessburg andHanna corneal trephination systems. Cornea15:589–598

82. Wilson SE, Bourne WM (1989) Effect of recipient-donor trephine size disparity on refractive errorin keratoconus. Ophthalmology 96:299–305

References 151

83. Yi Liu, Seitz B, Langenbucher A, Nguyen NX,Naumann GOH (2003) Impact of preoperativecorneal curvature on the outcome of penetratingkeratoplasty in keratoconus. Cornea 22:409–412

84. Zirm E (1906) Eine erfolgreiche totale Kerato-plastik. Albrecht von Graefes Arch Klin Exp Oph-thalmol 64:580–593


11.1Introduction

Since its development in 1989 by Pallikaris fol-lowed later by FDA approval in the United Statesin 1999, LASIK (laser-assisted in situ ker-atomileusis) has become an extremely com-monly performed surgical procedure. Infectivecomplications are rare [4] but present specialchallenges. Infective keratitis following LASIKoften involves organisms unusual in otherforms of infective keratitis. It usually occurs inthe flap interface and may be relatively inacces-sible to topical antibiotics. Bilateral infection,

although not common,occurs at least partly dueto the common practice of performing bilateralsimultaneous LASIK procedures. Clusters of in-fection have also been reported [3, 7, 11]. Finallyit should be noted that it is a vision-threateningcomplication occurring in people with general-ly high visual expectations, adding to its gravi-tas.

11.2Frequency and Presentation

The reported frequency of infection followingLASIK ranges from 0% to 1.5%, with the fre-quency in most large case series being less than0.2% [4]. Gram-positive and non-tuberculousmycobacterial infections are commonest, withthese organisms accounting for 26% and 47%of culture-positive infections respectively in areview of published cases [4]. Mycobacterial in-fections are probably overrepresented due to re-porting bias but do represent a strikingly highproportion of cases of post-LASIK infection.Gram-negative organisms, by contrast, accountfor very few cases. Fungal and Acanthamoebainfections have also been described (Table 11.1).

There are almost certainly predisposing fac-tors for post-LASIK infection. Uncontrolledmeibomian gland dysfunction and blepharitisprobably contribute to staphylococcal infection[12]. Performing LASIK on eyes that have previ-ously undergone photorefractive keratectomy(PRK) seems to be a risk factor [4]. Post-LASIKtrauma is undoubtedly associated with infec-tion. However, the commonest association withreported infections is a breakdown in sterilityduring the procedure, with systematic contami-

Infective Complications Following LASIK

Adam Watson, Sheraz Daya

11

|

∑ Infective complications following LASIK are a rare, potentially sight-devastatingcomplication but often have good outcomes

∑ Early diagnosis helps prevent rapid steroid-related progression of infection

∑ Atypical organisms are common, especiallynon-tuberculous mycobacteria

∑ Early presenting cases (7–10 days) and latepresenting cases (>10 days) have a differentmicrobiological profile

∑ Intact epithelium inhibits antibiotic pene-tration. Flap lift, antibiotic soak and epithe-lial defect creation are useful strategies

∑ Reculture, biopsy and flap amputation may be necessary for worsening keratitisdespite treatment

∑ Informed consent and attention to risk factors are crucial

Core Messages

nation of the surgical field probably beingresponsible for three reported clusters of myco-bacterial infection [3, 7, 11].

11.3Characteristics

Patients with post-LASIK infective keratitistend to present with varying combinations ofpain, photophobia, discomfort, redness and dis-charge. Deterioration in postoperative visualacuity is commonly noted and may be the solepresenting symptom. Patients may also beasymptomatic with the infection identified at aroutine postoperative examination.

The timing of the onset of symptoms varies –between zero days and several months [4, 12].Post-LASIK infections may usefully be dividedinto early and late groups depending on thelength of time from surgery to the onset ofsymptoms. Those presenting early occur in thefirst 7–10 days and are more likely to be causedby “typical” Gram-positive bacteria. Late infec-tions, presenting beyond 10 days, are more like-

ly to be atypical infections, especially non-tu-berculous mycobacteria but also including fun-gal infection.

Flap interface infiltrate is the commonestsign evident on examination although infiltratemay be confined to the lamellar flap or the un-derlying corneal stroma [4]. Other features ofinfection that may be present are those found inother forms of infective keratitis, including an-terior chamber reaction, keratic precipitates,corneal abscess and epithelial defects. Epithelialdefects are found far less frequently in post-LASIK keratitis and tend to be associated withGram-positive infection. The lack of an epithe-lial defect has important implications for treat-ment, as topical antimicrobial penetration ispoorer in the absence of a defect. An intact ep-ithelium presents a relatively impermeable bar-rier to topical antibiotic penetration.

Crystalline keratopathy has also been report-ed in several cases associated with Mycobacteri-um chelonae infection [2,22].This appearance ishighly suggestive of M. chelonae infection.

11.4Differential Diagnosis

11.4.1Diffuse Lamellar Keratitis (DLK,“Sands of the Sahara”)

DLK, a non-infectious inflammation occurringafter LASIK in approximately 2–4% of cases[13], may present with mild pain, redness andphotophobia in the 1st week after surgery. In themilder stage 1 and stage 2 forms of DLK the in-filtrates are light and diffuse and unlikely to beconfused with infection. More severe stages ofDLK involve clumping of cellular infiltrates and,in stage 4 cases, stromal melting. The possibili-ty of infection should always be considered inthese cases and, since treatment of more severeDLK involves flap lift and irrigation, it is pru-dent to take a scrape sample for microbiologywhen lifting the flap [16]. Use of topical steroidsfor presumed DLK may lead to initial apparentimprovement in infective keratitis with subse-quent rapid progression of infection and de-structive stromal necrosis.

154 Chapter 11 Infective Complications Following LASIK

Table 11.1. Organisms reported to have caused post-LASIK keratitis

BacteriaStaphylococcus aureusStreptococcus pneumoniaeStreptococcus viridansCoagulase-negative staphylococciPseudomonas aeruginosaNocardia asteroides

MycobacteriaMycobacterium chelonaeMycobacterium mucogenicumMycobacterium abscessusMycobacterium szulgai

FungiCandida albicansCurvularia lunataScedosporium apiospermumFusarium solanFusarium oxysporumColleotrichum (Fusarium-like)

OtherAcanthamoeba

11.4.2Steroid-Induced Intraocular PressureElevation with Flap Oedema (Pseudo-DLK)

This uncommon phenomenon generally pres-ents with decreased visual acuity, flap oedemaand variable inflammation and may be mistak-en for DLK. Increased frequency of steroid usethen leads to worsening of the condition. Thecentrally measured intraocular pressure (IOP)is often normal and careful examination mayreveal a fluid cleft in the flap interface. Peripher-al IOP measured with a Tono-Pen (Medtronic-Solan) reveals an elevated IOP and the condi-tion will resolve with control of IOP, usuallywith topical agents, and tapering or cessation ofsteroids [10, 15].

11.5Management

The principles of management are similar tothose in regular infective keratitis, namely:∑ Suspect infection∑ Obtain a microbiological sample prior to

starting treatment∑ Give broad spectrum empirical therapy ini-

tially∑ Tailor therapy depending on clinical re-

sponse and microbiological results (Gramand other stains, culture, sensitivities)

∑ If there is a worsening clinical situation andno microbiological information to guide,consider temporary withdrawal of treatmentfor rescrape or corneal biopsy

Post-LASIK infective keratitis differs from regu-lar infective keratitis in that:∑ Atypical infections (non-tuberculous my-

cobacteria) are relatively common∑ Antibiotic penetration may be poor due to an

intact epithelium∑ Flap complications such as striae, epithelial

ingrowth, flap melt and dehiscence may beproblematic, related to infection or flap lift

We propose a management algorithm that takessome of these factors into account (Fig. 11.1).

11.5.1Flap Lift

This should be carried out in most circum-stances. An exception is if the focus of infectionis very peripheral and associated with overlyingflap necrosis allowing an adequate microbiolog-ical sample and debridement of infectious ma-terial (Fig. 11.2).

The flap may be lifted completely or partially,depending on the extent and location of infil-trate. Flap lift should be carried out beneath anoperating microscope under sterile conditionswith or without patient sedation. Some prefer toinitiate the flap lift at the slit lamp where the flapborder may be more easily identified. Initiationof flap lift is generally with a blunt spatula orSinskey hook to break the epithelium and openthe interface for one or two clock hours, thencompleted with non-toothed LASIK flap forceps.

11.5.2Specimen Taking

Gentle scraping of material for microbiologicalexamination and culture and to debride infec-tive debris follows this. A hypodermic needle,number 15 Bard-Parker blade or Kimura spatulamay be used. The authors prefer to plate thespecimens themselves on culture media imme-diately.We suggest as a minimum, if the amountof material allows, an air-dried slide for im-mediate Gram stain, blood, chocolate andSabouraud’s agar plates and brain-heart infu-sion broth. If Mycobacterium is suspected, thenculture on Lowenstein-Jensen medium shouldbe considered. Useful additional stains for late-presenting cases include auramine-rhodaminefor acid-fast bacilli [22] and periodic acid–Schiff (PAS) for fungi [21].


∑ A microbiological specimen prior to treatment is essential

∑ Flap lift is usually necessary

11.5 Management 155


Fig. 11.1. Algorithm outliningan approach to post-LASIKinfective keratitis

Fig. 11.2. Peripheral infiltrate 3 weeks after LASIKwith focal flap melt

Fig. 11.3. Arrow indicates an epithelial defect creat-ed over a peripheral interface infiltrate after raisingpart of the flap for an interface scrape. The defect aidsantibiotic penetration

11.5.3Treatment

A moistened lint-free sponge may be used toremove residual debris, followed by “soaking” ofthe flap and stromal bed in antibiotic solution.The choice of antibiotics may depend onwhether the keratitis falls into the early or lategroup (Fig. 11.1). Soaking should be for 2 min ormore with each antibiotic solution in turn, fol-lowed by careful relaying of the flap. If there islittle or no epithelial defect overlying the sus-pected infection, an epithelial defect should becreated to aid antibiotic penetration (Fig. 11.3).

Intensive topical antibiotics should then bestarted (hourly alternating around the clock).The choice of antibiotics will be partly deter-mined by the resistance characteristics of bacte-ria in the local region. Specialist microbiologistadvice should be sought if there is doubt.Suggestions for treatment choice are given inFig. 11.1.

Topical steroids should be avoided in theearly stages of treatment and only instituted,if at all, when there is clear clinical evidence ofimprovement (e.g. less pain, diminishing andcoalescing infiltrate, fewer keratic precipitates,healing epithelial defect), suggesting sterilisa-tion of the offending organism. Introduction ofany steroid should generally be in low dose (e.g.twice daily prednisolone sodium phosphate0.5%) and the response closely monitored forsigns of worsening infection, e.g. satellite infil-trates. Steroid use without concomitant antibi-otic has been implicated in the recrudescence of infection after apparent sterilisation ofPseudomonas keratitis [8].Steroid use should beavoided in cases of fungal keratitis.

Topical antibiotic choice may be alteredwhen microbial sensitivities are available. If theinfection is clinically improving, there may beno need to change the antibiotics other than ta-pering the frequency of use after 2–3 days. If theinfection is improving and sensitivity data areavailable, it may be reasonable to discontinueone of the antibiotics (e.g. gentamicin in a van-comycin/gentamicin combination when treat-ing a staphylococcal infection) to minimise ep-ithelial toxicity and promote healing.

The use of preservative free lubricants to pre-serve epithelial health should be considered. Acycloplegic (preservative free cyclopentolate orhomatropine) should be added if there is signif-icant anterior chamber inflammation.


∑ Choose antibiotics to cover atypical organisms in late-presenting cases

∑ Antibiotic penetration is aided by an epithelial defect

∑ Avoid steroid use unless there is unequivo-cal improvement suggesting sterilisation of infection

∑ Avoid steroids in fungal infection and without concomitant antibiotic use

11.5.4No Improvement

Failure of the infection to show signs of im-provement after several days of treatmentshould prompt a re-evaluation. An attempt atreculturing the infective agent is mandatory, byfurther corneal scrape or corneal biopsy. If theinfection is severe and judged to be threateningthe eye, flap amputation may be necessary, withhalf the flap being sent for histological exami-nation and staining for organisms, the otherhalf being sent for microbial culture.A high sus-picion for atypical infection exists at this pointand mycobacteria, fungi and Acanthamoebashould be specifically looked for.

Failure to control the infection despite treat-ment, as with regular infective keratitis, may re-quire further surgical intervention includingtherapeutic penetrating keratoplasty, and in-traocular instillation of antimicrobial drugs inthe case of perforation with suspected endo-phthalmitis, with or without lensectomy andvitrectomy depending on the involvement ofintraocular structures.

11.5 Management 157

11.6Special Considerations

11.6.1Mycobacteria

Topical clarithromycin and amikacin have gen-erally been the agents of choice for treatment ofM. chelonae keratitis. Tobramycin and the fluo-roquinolones are also often effective. There hasbeen recent interest in the fourth generation flu-oroquinolones, including moxifloxacin [1] andgatifloxacin, as having greater activity againstnon-tuberculous mycobacteria. The authorshave experience of treating a case (unpub-lished) of bilateral moxifloxacin-resistant M.chelonae post-LASIK keratitis (Figs. 11.4, 11.5).This highlights the benefit of using multipleagents to treat infection empirically until the or-ganism’s sensitivities are known, with contin-ued use of multiple antibiotics to which the or-ganism is sensitive to prevent recrudescence.Treatment may need to be continued for

6 months or more with a gradual taper, moni-toring closely for signs of recurrence. Viablemycobacteria have been cultured from an am-putated LASIK flap despite 9 weeks of appropri-ate treatment for M. chelonae keratitis [19].

11.6.2Fungal Keratitis

Fungal infections comprise about 14% of re-ported cases of post-LASIK keratitis [4]. Identi-fication of hyphae, pseudohyphae or yeasts maybe possible from direct microscopic examina-tion of appropriately stained slide preparationsof a scrape; or culture may yield fungal growth.An additional approach, maybe more applicablein the future, is PCR testing of specimens forfungal DNA, providing a quicker result thanfungal culture. This method, while sensitive,does suffer from poor specificity [21].

Treatment of fungal infections should be de-termined in collaboration with a microbiologistand based on the organism’s sensitivities when available. Common topical agents arenatamycin 5% and amphotericin B 0.15%, bothpolyenes with a broad spectrum of activityagainst filamentous fungi and yeasts althoughnatamycin may be slightly more effective andthe preferred choice where available [21]. Topi-cal econazole 1% is also being used where ap-propriate. Topical treatment should generally becombined with a systemic agent, e.g. one of theazoles such as ketoconazole or itraconazole.Voriconazole, a relatively new triazole agent, has


Fig. 11.4 A, B. Bilateral central Mycobacterium che-lonae post-LASIK keratitis (right eye A, left eye B).Note central interface infiltrates

A

B

Fig. 11.5. Subsequent right central flap melt in thecase of M. chelonae keratitis shown in Fig. 11.4

been reported to have superior activity againstScedosporium infections [18].

The use of topical steroids may cause fungalkeratitis to progress rapidly to widespreadcorneal involvement and perforation. Steroidsshould be avoided when treating fungal infec-tions, at least until effective antifungal treat-ment has been continued for several weeks. An-tifungal therapy needs to be prolonged for atleast 6 weeks – agents are generally fungistaticrather than fungicidal at the concentrationachieved in the corneal stroma, and eliminationof fungus depends ultimately on the host im-mune response.

11.6.3Viral Keratitis

Case reports of apparent reactivation of Herpessimplex keratitis following LASIK have beenpublished [5, 17]. It is not clear whether theLASIK procedure and/or the postoperative useof topical steroids were causative. However, ul-traviolet radiation exposure has been associat-ed with reactivation of latent Herpes simplex[20, 6]. In addition to a short-term topical an-tiviral, consideration should be given to longer-term systemic antiviral prophylaxis (e.g. oralacyclovir 400 mg twice daily).

11.7Visual Outcome

The visual outcome following post-LASIK ker-atitis is highly variable. Approximately 50% ofreported cases have no clinically significantworsening of best-corrected Snellen visual acu-ity. Twenty-five per cent suffer a severe reduc-tion [4]. Gram-positive infections are associat-ed with better visual outcomes while fungalinfections (excluding Candida albicans) aremore likely to be associated with severe visualreduction. Reported cases of C. albicans, on theother hand, had a good visual outcome – with abest corrected visual acuity average of 20/25[16]. Reported mycobacterial cases tend to beintermediate between Gram-positive and fun-gal infection in terms of visual outcome.

11.8Management of Sequelae

Common sequelae of post-LASIK infection in-clude scarring (Fig. 11.6), irregular astigmatismand varying degrees of epithelial ingrowth aris-ing from flap lift or flap melt (Fig. 11.7).

Once the infection has settled, the goal oftreatment is to optimise visual acuity in theaffected eye. How this is achieved will varymarkedly from case to case. Correction of re-fractive error should initially be explored usingglasses, soft contact lens and rigid gas perme-able lenses. Significant epithelial ingrowth in-

11.8 Management of Sequelae 159

Fig. 11.6. Eye 9 months following treatment for M.chelonae post-LASIK keratitis. Arrows point to stro-mal scarring (white) and stable interface epithelialinclusions (yellow). The uncorrected visual acuity is6/7.5

Fig. 11.7. Interface epithelial ingrowth arising froma flap defect in a case of M. chelonae post-LASIKkeratitis. Tongues of epithelium are progressing pe-ripherally

ducing astigmatism needs to be cleared fromthe flap interface prior to any further attemptsat surgical correction. Irregular astigmatismresulting from scarring may be amenable tocontact lens correction.

Consideration of further excimer laser re-fractive surgery should be approached withcaution. In addition to likely patient concernabout a repeat procedure, further LASIK will re-quire recutting of a deeper flap to avoid thescarred and irregular interface inevitably pres-ent, and PRK or laser epithelial keratomileusis(LASEK) is associated with a high risk of devel-opment of haze in an environment with activat-ed keratocytes.

Significant opacity affecting the visual axis,on the other hand, may need to be cleared.Options for this include homoplastic automatedlamellar therapeutic keratoplasty (HALTK, auseful technique for opacities limited to the an-terior one-third of the corneal stroma) [9], deepanterior lamellar keratoplasty [14] and pene-trating keratoplasty.

11.9Prevention

Rare cases of post-LASIK infective keratitis areinevitable. Attention to patient eyelid hygienewith control of blepharitis, careful patient in-struction regarding pre- and postoperative careand avoidance of trauma, and meticulous atten-tion to equipment sterility and operating envi-ronment hygiene are likely to lead to fewer cas-es. The authors strongly advise that separateblades and microkeratome heads be used ifcarrying out simultaneous bilateral LASIK todiminish the linked risk of bilateral infection.Above all, careful informed consent of thepatient prior to surgery is mandatory.

References

1. Abshire R, Cockrum P, Crider J et al. (2004) Topi-cal antibacterial therapy for mycobacterial ker-atitis: potential for surgical prophylaxis andtreatment. Clin Ther 26:191–196

2. Alvarenga L, Freitas D, Hofling-Lima AL et al.(2002) Infectious post-LASIK crystalline ker-atopathy caused by nontuberculous mycobacte-ria. Cornea 21:426–429

3. Chandra NS, Torres MF,Winthrop KL et al. (2001)Cluster of Mycobacterium chelonae keratitis cas-es following laser in-situ keratomileusis. Am JOphthalmol 132:819–830

4. Chang MA, Jain S, Azar DT (2004) Infections fol-lowing laser in situ keratomileusis: an integrationof the published literature. Surv Ophthalmol49:269–280

5. Davidorf JM (1998) Herpes simplex keratitis afterLASIK. J Refract Surg 14:667

6. Dhaliwal DK, Romanowski EG, Yates KA et al.(2001) Experimental laser-assisted in situ ker-atomileusis induces the reactivation of latentherpes simplex virus. Am J Ophthalmol 131:506–507

7. Freitas D, Alvarenga L, Sampaio J et al. (2003) Anoutbreak of Mycobacterium chelonae infectionafter LASIK. Ophthalmology 110:276–285

8. Gritz DC, Kwitko S, Trousdale MD et al. (1992) Re-currence of microbial keratitis concomitant withantiinflammatory treatment in an animal model.Cornea 11:404–408

9. Hafezi F, Mrochen M, Fankhauser F 2nd et al.(2003) Anterior lamellar keratoplasty with a mi-crokeratome: a method for managing complica-tions after refractive surgery. J Refract Surg19:52–57

10. Hamilton DR, Manche EE, Rich LF et al. (2002)Steroid-induced glaucoma after laser in situ keratomileusis associated with interface fluid.Ophthalmology 109:659–665

11. Holmes GP, Bond GB, Fader RC et al. (2002) Acluster of cases of Mycobacterium szulgai kerati-tis that occurred after laser-assisted in situ ker-atomileusis. Clin Infect Dis 34:1039–1046

12. Karp CL, Tuli SS, Yoo SH et al. (2003) Infectiouskeratitis after LASIK. Ophthalmology 110:503–510

13. McGhee CNJ, Brahma A (2001) Uncommon com-plications of LASIK: diffuse lamellar keratitis and epithelial ingrowth. CME J Ophthalmol 5:52–54

14. Melles GRJ, Lander F, Rietveld FJR et al. (1999) Anew surgical technique for deep stromal, anteriorlamellar keratoplasty. Br J Ophthalmol 83:327–333

15. Nordlund ML, Grimm S, Lane S et al. (2004) Pres-sure-induced interface keratitis: a late complica-tion following LASIK. Cornea 23:225–234

16. Peng Q, Holzer MP, Kaufer PH et al. (2002) Inter-face fungal infection after laser in situ kerato-mileusis presenting as diffuse lamellar keratitis.J Cataract Refract Surg 28:1400–1408


17. Perry HD, Doshi SJ, Donnenfeld ED et al. (2002)Herpes simplex reactivation following laser insitu keratomileusis and subsequent perforation.CLAO J 28:69–71

18. Shah KB, Wu TG, Wilhelmus KR et al. (2003) Activity of voriconazole against isolates ofScedosporium apiospermum. Cornea 22:33–36

19. Solomon A, Karp CL, Miller D et al. (2001) Myco-bacterium interface keratitis after laser in situkeratomileusis. Ophthalmology 108:2201–2208

20. Spruance SL (1985) Pathogenesis of herpes sim-plex labialis: experimental induction of lesionswith UV light. J Clin Microbiol 22:366–368

21. Thomas PA (2003) Fungal infections of thecornea. Eye 17:852–862

22. Verma S, Watson SL, Dart JKG et al. (2003) Bilat-eral Mycobacterium chelonae keratitis followingLASIK (letter). J Refract Surg 19:379–380

References 161

12.1Introduction

12.1.1Etiology and Clinical Course of OcularAdenoviral Infection

Adenoviral keratoconjunctivitis (AKC) was firstdescribed by Fuchs in 1889 [7]. In 1955, Jawetzidentified adenovirus as the cause of the disease[20]. Other authors isolated adenovirus sero-types 8, 19 and 37 as the most frequent causative

adenovirus subtypes [5]. Overall, many of themore than 40 adenovirus serotypes cause infec-tions of the ocular surface with and withoutgeneral symptoms. Pharyngoconjunctival feveris a conjunctivitis with upper respiratory tractinvolvement caused by serotypes 1, 3–7, and 14.The term “epidemic keratoconjunctivitis” de-scribes an infection of the ocular surface with-out general symptoms caused by serotypes 2–4,7–11, 14, 16, 19, 29, and 37. Unspecific follicularconjunctivitis is probably caused by all theabove and serotype 34. In principle, neither isthere a clear-cut distinction in this clinical

Treatment of Adenoviral Keratoconjunctivitis

Jost Hillenkamp, Rainer Sundmacher, Thomas Reinhard

12

∑ The clinical course of adenoviral kerato-conjunctivitis (AKC) should be divided intoan acute phase with conjunctival inflamma-tion of varying intensity with or withoutcorneal involvement and a chronic phasewith corneal opacities

∑ AKC is caused by many different serotypesand is highly contagious during the acutephase

∑ The economic and social price of AKC as an community epidemic is high

∑ Corneal opacities, the hallmark of thechronic phase, are usually self-limited

∑ Topical steroids should be avoided becausethey prolong viral replication, frequentlylead to long-lasting dry-eye symptoms, andcorneal opacities almost always recur afterdiscontinuation of topical steroids

∑ There is currently no effective and clinicallyapplicable topical antiviral agent for thetreatment of the acute phase of AKC

∑ Topical cidofovir is the first antiviral agentwhich has effectively reduced the incidenceof corneal opacities, but local toxicity rulesout its clinical application

∑ Recently, NMSO3, a sulfated sialyl lipid, hasdemonstrated a greater antiviral potencyagainst adenovirus in vitro than cidofovirexhibiting minimal cytotoxicity

∑ Topical cyclosporin A (CsA) appears to beeffective in the treatment of persistentcorneal opacities

∑ Topical interferon might be effective as aprophylaxis of infection

∑ Topical interferon is currently not commer-cially available due to unsettled patentissues

∑ Adequate infection control measuresshould be followed as prevention and to reduce epidemic AKC outbreaks

Core Messages

classification nor can the various serotypes beclearly associated with a distinct clinical pres-entation [5, 40].

The course of AKC must be divided into anacute phase and a chronic phase.


∑ Many different serotypes of adenoviruscause AKC

12.1.1.1The Acute Phase

The acute phase has a wide spectrum of dura-tion and intensity of local symptoms. It is prin-cipally self-limited. The intensity of symptomsvaries between a picture of a mild unspecificconjunctivitis and intense conjunctival injec-tion with marked chemosis and hemorrhagicinvolvement of the conjunctiva and the eyelids(Figs. 12.1–12.5). After an incubation period of2–14 days symptoms usually begin in one eyeand the other eye becomes symptomatic after2–4 more days (Fig. 12.1). The mild forms of ade-noviral conjunctivitis are clinically difficult todifferentiate from any other unspecific conjunc-tivitis. The more pronounced cases can be read-ily diagnosed by the clinician. They present witha typical picture of conjunctival hyperemia andchemosis, swelling of the conjunctival plica, andintense serous or muco-serous tearing (Fig. 12.2).Conjunctival pseudomembranes occur in somecases. Ipsilateral preauricular lymphadeno-pathy is a fairly typical sign observed in manypatients. The more severe cases of ocular aden-oviral infections are characterized by a highlydistressing morbidity [5] (Figs. 12.3–12.5). Thecornea is not necessarily involved in the acutephase. If it is, a superficial punctate keratitiswith small epithelial punctatae or larger stella-ta-type coarse punctatae and subepithelial infil-trates may develop [40]. Rarely, the cornea be-comes involved with large epithelial erosionsduring the acute phase of the infection. Alsorarely, the corneal endothelium becomes in-volved in the form of an endotheliitis withmarked temporary corneal edema which usual-ly subsides spontaneously [40].

164 Chapter 12 Treatment of Adenoviral Keratoconjunctivitis

Fig. 12.1. Acute phase of AKC 1 week after onset ofsymptoms. Symptoms obviously first appeared in theright eye; the left eye became involved 2 days later

Fig. 12.2. Typical presentation of the acute phase ofAKV with serous conjunctivitis

Fig. 12.3. This presentation with severe chemosis isfairly typical for the acute phase of AKC. An infectionwith herpesvirus cannot be clinically excludedwhereas this picture clearly differs from allergicchemosis


∑ The acute phase of AKC has a wide spectrumof intensity and duration of symptoms

∑ The acute phase of AKC is self-limited∑ The cornea may or may not be involved

in the acute phase of AKC

12.1.1.2The Chronic Phase

It is the corneal involvement during the chronicphase of the disease that sets AKC apart fromother forms of virus conjunctivitis. Typically,during the course of the infection, approxi-mately 10 days after onset of symptoms, cor-

neal subepithelial opacities frequently appear(Fig. 12.6). These nummular opacities or infil-trates can impair visual function, and may per-sist for months to years [40, 28]. Histopatholog-ical investigation of focal biopsies revealedsubepithelial infiltrates of lymphocytes, histio-cytes and fibroblasts accompanied by a disrup-tion of the collagen fibers of Bowman’s layer [15,25]. The pathogenesis of the nummular opaci-ties most likely includes a persisting viral repli-cation in subepithelial keratocytes triggering animmunological host reaction [40]. This hypoth-esis is supported by the clinical observation thatopacities usually resolve with topical steroidtreatment but recur when steroids are discon-tinued [39].


∑ Corneal opacities are the hallmark of thechronic phase of AKC

∑ Corneal opacities are probably caused by an immunological host reaction againstpersisting virus in keratocytes

∑ Corneal opacities almost invariably sponta-neously resolve mostly within 1 year


Fig. 12.4. This presentation with pronounced hem-orrhagic chemosis is also fairly typical for the acutephase of AKC

Fig. 12.5. This is an exceptionally severe hemor-rhagic involvement of the eyelids in the acute phase ofa case of proven AKC

Fig. 12.6. Corneal opacities are the hallmark ofthe chronic phase of AKC. Histopathology revealedsubepithelial infiltrates of lymphocytes, histiocytesand fibroblasts accompanied by a disruption of thecollagen fibers of Bowman’s layer [6, 7]. The patho-genesis of the nummular opacities most likely in-cludes a persisting viral replication in subepithelialkeratocytes triggering an immunological host reac-tion which causes the visible, steroid-sensitive opaci-ties

12.2Socioeconomic Aspect

AKC is a highly contagious disease which occursworldwide sporadically and epidemically.Whilenot permanently blinding, adenoviral ocular in-fections remain the most common external oc-ular viral infection worldwide. The economicand social price of this community epidemicalso remains high [6]. Public institutions suchas schools or kindergartens must be closed fol-lowing the outbreak of an epidemic. Many workhours are lost every year.


∑ The economic and social price of AKC as a community epidemic is high

12.3Treatment

The clinical investigation of candidate treat-ments of AKC in patient studies has been ham-pered in the past by the very variable intensityand duration of the clinical symptoms and theself-limited nature of the acute phase of AKC.Furthermore, most studies have failed to applylaboratory tests for adenovirus as the cause ofthe disease and therefore the etiology of thetreated diseases remains questionable. A possi-ble treatment must therefore be evaluated byadequately designed large prospective con-trolled clinical trials with reliable tests of aden-ovirus as the underlying cause of the treateddisease. Also, treatment of the acute phase ofAKC must be distinguished from treatment andprophylaxis of corneal opacities, the hallmarkof the chronic phase. Currently, no clinically ap-plicable specific antiviral therapy is available toshorten the course of the infection, to improvethe distressful clinical symptoms, to stop viralreplication, and to prevent the development ofcorneal opacities. The effect of several topicalagents has been investigated. Only steroids [42]and cidofovir [16, 17] have demonstrated a cer-tain therapeutic effect and will therefore befurther discussed. Other measures, such astopical povidone-iodine, topical interferon [26,

30, 31, 38, 41, 47], topical non-steroidal anti-in-flammatories [13, 37], topical cyclosporin A ortopical trifluridine, have been shown not to be more effective than topical lubrication [16,17, 19, 46].

12.3.1Treatment of the Acute Phase

12.3.1.1Topical Steroids

Treatment of the acute phase of the infectionwith topical steroids has been widely recom-mended because of the clinical experience thatthe distressing local symptoms subside earlierwith steroids. The effect of topical steroids hasbeen investigated in a prospective randomizedclinical study [42]. The results of this study con-firm the clinical impression that the improve-ment of the local symptoms is accelerated withsteroids. Furthermore, the number of cornealopacities per affected cornea in the chronicphase was reduced as compared to controls[42]. However, the number of patients affectedby corneal opacities was not reduced andcorneal opacities appeared later in the course ofthe disease. This finding suggests that steroidsmay prolong the persistence of adenovirus inthe cornea. This undesired effect was confirmedby Romanowski et al., who found an increasedviral replication under topical steroids [32].Also, a significantly greater number of patientstreated with topical steroids experienced long-lasting, distressing dry eye symptoms as com-pared to controls [42].These findings lead to theconclusion that the negative effects of topicalsteroids outweigh the positive effect of an earli-er relief of the distressing local symptom [32,40,42]. Therefore, topical steroids should be avoid-ed in the treatment of the acute phase of AKC.


∑ Local symptoms subside with topicalsteroids but viral replication is probablyprolonged

∑ Topical steroids frequently lead to long-lasting dry-eye symptoms

∑ Topical steroids should be avoided


12.3.1.2Topical Cidofovir

Principally, an effective antiviral agent to short-en the course of the infection, to improve thedistressful local symptoms of the acute phase,and to prevent the development of cornealopacities of the chronic phase would representthe ideal treatment of AKC.

Topical ganciclovir was only mildly effectivein vitro and in the cotton rat model and wasthus not further investigated as a potentialtreatment [43, 44].

Cidofovir or HPMPC, a broad-spectrum an-tiviral agent, demonstrated a significant in-hibitory effect on adenovirus types 1, 5, 8 and 19isolated from patients with AKC in vitro [8]. Theefficacy of cidofovir was also documented invivo. Topical cidofovir demonstrated significantantiviral activity in the AD 5 McEwen/NZ rabbitocular model with 0.2% as the lowest effectiveconcentration [11, 10]. Gordon et al. first report-ed clinical efficacy and safety of topical cido-fovir 0.2% in the treatment of a single patientwith proven AKC [12].

These results encouraged us to investigatethe effect of topical cidofovir in a 0.2% concen-tration in the same preparation as described byGordon et al. [16] Topical cidofovir 0.2% provedto be a well tolerated drug which did not causeany discomfort but did not have a statisticallysignificant effect on the course of the acute orthe chronic phase of adenoviral keratoconjunc-tivitis. In particular, the frequency of corneal in-filtrates at the end of the 21-day treatment peri-od was not altered by cidofovir 0.2% [16].

There are several possible explanations forthe failure of cidofovir 0.2% to show the clinicalefficacy that may have been expected from theantiviral activity demonstrated in vitro [8] andin the rabbit ocular model [11, 10].

Concentration of Cidofovir. Cidofovir 0.2%was administered 4 times daily. Gordon andRomanowski et al. showed that cidofovir 0.2%administered 4–5 times daily limited adenoviralreplication in the adenovirus type 5/New Zea-land rabbit ocular model. Cidofovir 0.5% and1% administered only twice daily was not supe-rior, but equally effective [11].

Serotype Dependency. Adenovirus demon-strated serotype-dependent differences in in-vitro infectivity titers and clinical course. AD8was the most frequent serotype in a total num-ber of 106 positive adenoviral cultures, causingsignificantly more often a severe clinical coursewith marked eyelid edema than other sero-types. AD3 and AD4 were associated with high-er infectivity titers than other serotypes. Infec-tivity and the clinical course of AKC areserotype dependent [33].

Cidofovir proved to be effective against ade-novirus types 1, 5 and 6 in the rabbit model [11,29], but a relative resistance of serotype 19 tocidofovir was reported [8]. Variants of aden-ovirus serotype 5 with different sensitivity totopical treatment with cidofovir 0.5% in therabbit ocular model have been described byAraullo-Cruz et al. [1]. Consequently, the effica-cy of cidofovir in patients may also be serotypedependent. The patients enrolled in clinicalstudies were probably infected with variousadenovirus types.

Pharmacokinetics. Eyedrops may be washedout by intense tearing; additionally regularconjunctival absorption of eyedrops may be im-paired in patients with severe conjunctivalchemosis and swelling of the conjunctival plica.

Viral Replication and Onset of Treatment. TheNew Zealand rabbit ocular model of adenovirustype 5 infection showed a duration of viral repli-cation of 9 days with a peak on day 3. The symp-toms of AKC after the phase of viral replicationare thought to be caused by the host’s immuneresponse [9]. Treatment with cidofovir in therabbit model began 24 h after inoculation [29].The treatment of study patients began 1–7 days(mean: 3.5 days) after onset of symptoms whenthey first presented to the clinic [16]. Cidofovirmay have been effective in the rabbit ocularmodel because treatment began early in thecourse of the infection while treatment of ourpatients may have failed because treatment didnot start until the phase of viral replication hadalmost been completed.

The fivefold concentration of cidofovir test-ed in our second pilot study did not alter thecourse of the acute phase but appeared to be

12.3 Treatment 167

effective in the prevention of severe cornealopacities [17]. So far, only topical immunosup-pressants such as steroids [39] or cyclosporin A[28] have been shown to be effective in the treat-ment of existing corneal opacities. These agentsprobably act by suppressing the immunologicalresponse directed against viral antigens whichpersist in the cornea. By contrast, as an antiviralagent, cidofovir treats the underlying cause ofcorneal opacities. As opposed to a treatmentwith steroids, or less so with topical cyclosporinA [28], it could therefore possibly prevent theiroccurrence, thereby avoiding the problem of re-currences after discontinuation of treatment[39]. Unfortunately, local toxicity forbids theclinical application of cidofovir in the 1% con-centration. We observed an increased preva-lence of conjunctival pseudomembranes as wellas conjunctivitis and erythematous inflamma-tion of the skin of the eyelids. These changessubsided completely after 7–28 days (mean13.7 days) with topical lubrication and dexpan-thenol ointment applied to the affected skinareas [17] (Fig. 12.7). Others have additionallydescribed lacrimal blockade following the ap-plication of topical cidofovir [34]. These resultsare disappointing, however, the study demon-strated for the first time the principal efficacy ofa topical antiviral agent as a treatment of AKC.This positive aspect of the study encouragesfurther research for an effective yet tolerableantiviral agent.


∑ Topical cidofovir is the first antiviral agentwhich effectively reduces the incidence of corneal opacities

∑ Local toxicity rules out the clinical applica-tion of topical cidofovir in an antivirallyeffective concentration

12.3.2Treatment of the Chronic Phase

12.3.2.1Topical Steroids

The pathogenesis of the nummular cornealopacities of the chronic phase most likely in-cludes a persisting viral replication in sub-epithelial keratocytes triggering an immunolog-ical host reaction. Topical steroids suppress thehost reaction and thus lead to a quick disappear-ance of the opacities. Unfortunately, opacities al-most invariably reappear when topical steroidsare discontinued. The opacities frequently recurprobably because steroids effectively suppressthe immunological host reaction but lack a con-comitant antiviral effect [40, 42]. Suppressingthe immunological host reaction may thereforeeven enhance and prolong viral persistencewithin keratocytes. Even prolonged tapering oftopical steroids failed to prevent recurrences ofcorneal opacities, even years after the initialacute phase of the disease [39]. For this reasonand because of unwanted side effects of pro-longed use of topical steroids such as cataractformation and secondary glaucoma, steroidtreatment for the chronic phase of AKC cannotbe recommended. There are no controlled clini-cal studies documenting the natural course ofcorneal opacities in AKC, but in the pre-steroidera Thygeson reported that corneal opacities inAKC invariably spontaneously disappear within1 year. This is in accordance with our own clini-cal experience. Persistence of opacities withouttreatment for more than 1 year is a rarity. Alsorarely, persistent, scarred corneal opacities causeirregular astigmatism which can be effectivelycorrected with hard contact lenses. Penetratingkeratoplasty is usually not required [40].


Fig. 12.7. Local side effects of topical cidofovir 1%:conjunctivitis and erythematous inflammation of theeyelids


∑ Corneal opacities almost always recur afterdiscontinuation of topical steroids

∑ Even careful and prolonged tapering of topical steroids failed to prevent the recurrence of corneal opacities

∑ Corneal opacities mostly resolve sponta-neously within 1 year

12.3.2.2Topical Cyclosporin A

Cyclosporin A (CsA) is a well-established im-munosuppressant which has been used in theprevention of transplant rejection for 25 years[3]. Topical CsA was used effectively in the treat-ment of Mooren’s ulcer [48], vernal keratocon-junctivitis [2], ulcerative keratitis associatedwith rheumatoid arthritis [24], anterior uveitis[18], and Thygeson’s punctate keratitis [27]. Sideeffects of topical CsA have not been described.CsA may also have some antiviral potency asCsA has been shown to inhibit herpes simplexvirus in vitro [45]. In a non-controlled studytopical CsA 2% led to a disappearance ofcorneal opacities in two-thirds of 56 treated pa-tients with persisting opacity [28]. The opacitiesslowly responded to this treatment over severalweeks, a response significantly slower than therapid response to topical steroids [28]. TopicalCsA was well tolerated, and only 7% of all treat-ed patients discontinued the medication be-cause of a local burning sensation. After slowtapering of topical CsA, one-third of the initialresponders suffered a recurrence of the opaci-ties. In all of these patients the opacities couldbe effectively abolished with another course oftopical CsA on a low maintenance level of 1–2drops/day [28].


∑ Topical CsA appears to be effective in thetreatment of persistent corneal opacities

∑ So far, no side effects of topical CsA havebeen described

12.3.3Prophylaxis

12.3.3.1Topical Interferon

Interferons have multiple immunomodulativeeffects and therefore their effect on the course ofviral infections is difficult to predict [40]. Treat-ment of the acute phase of AKC with topical in-terferon has been shown not to be effective [26,30, 31, 38, 41, 47]. However, the consequent appli-cation of one drop per day of topical interferon(Berofor®) to all unaffected hospital staff andinpatients in a nosocomial viral epidemic kera-toconjunctivitis which was recurrent in spite of hygienic prophylactic measures seemed toeffectively prevent further spread of the infec-tion [35]. In view of the epidemic character ofAKC, prophylactic topical interferon seems tobe an effective measure to protect yet unaffect-ed individuals. However, topical interferon iscurrently not commercially available. Berofor®has been removed from the market due tounsettled patent issues.


∑ Interferons have multiple immunomodula-tive properties

∑ Topical interferon might be effective as aprophylaxis of infection

12.3.3.2Infection Control and Hygienic Measures

It has been shown that infection control pro-grams including specified methods of patientscreening and isolation, handwashing, instru-ment disinfection, medication distribution, andfurlough of infected employees are associatedwith decreased rates of nosocomial AKC out-breaks and outbreak morbidity in a large teach-ing eye institute [14]. A number of precautionsshould be observed to ensure safety in any oph-thalmologist’s office. Paramount among these ishand washing, both immediately after contactwith patient’s eyes and again between patients[4]. Highly concentrated alcohol-based handrubs and hand gels have been demonstrated to

12.3 Treatment 169

inactivate adenovirus within 2 min [21], povi-done-iodine, peracetic acid, and formaldehydehave also demonstrated antiviral activity as dis-infectant agents against adenovirus althoughthe resistance of individual serotypes varies andthe genomes of adenoviruses showed consider-ably more chemical resistance than the com-plete viral particle [36].

Careful disinfection of contact tonometrydevices between patients is important. Further-more, contact lens fitting is a typical procedurecarrying the risk of transmitting adenovirus [4, 21]. A recent study investigated the effect ofchemical, hydrogen peroxide, and heat steriliza-tion systems on contaminated hard and softcontact lenses. Only heat sterilization was effec-tive.As heat sterilization is not readily available,it may be prudent for patients with AKC to dis-pose of unclean contact lenses [21].

12.4Conclusion and Outlook

Following topical cidofovir’s failure in clinicaldevelopment, the need for an antiviral to treatAKC persists.

It must be the aim of future work to investi-gate the therapeutic properties of an effectiveyet non-toxic topical antiviral agent for thetreatment of all non-herpetic viral infections of the ocular surface. The acute phase of AKC calls for an antiviral monotherapy whereas thechronic phase with corneal opacities may re-quire supplementation with an immunosup-pressant such as topical cyclosporin A.

Because of the wide spectrum of durationand intensity of local symptoms of the naturalcourse of the acute phase of AKC, a possible top-ical treatment will ultimately have to be investi-gated in an adequately designed prospectivecontrolled clinical trial once the necessary invitro studies and investigations in the alreadyestablished rabbit ocular model have been com-pleted.

Recently, NMSO3, a sulfated sialyl lipid,has demonstrated a greater antiviral potencyagainst adenovirus in vitro than cidofovir ex-hibiting minimal cytotoxicity [22]. The applica-

tion of NMSO3 10% also effectively inhibitedviral replication in the established Ad5/NZWrabbit ocular model although to a lesser degreethan cidofovir [34]. Further dosage and toxicitystudies of NMSO3 are required before this agentcan be tested in humans.

12.5Current Clinical Practice and Recommendations

Until treatment with an adequate antiviral agenthas been established the following therapeuticstrategies can be recommended:∑ The acute phase of AKC may be treated with

copious topical lubrication alone to alleviatedistressful local symptoms.

∑ Topical cyclosporin A may be applied in pa-tients with functionally relevant cornealopacities that fail to spontaneously disappearafter months. This treatment must be ta-pered very slowly over several weeks to avoidrecurrences.

∑ Prophylaxis of infection of exposed,yet unaf-fected individuals with topical interferon canbe recommended.

∑ Topical steroids should be avoided in boththe acute and the chronic phase of the dis-ease.

∑ Adequate infection control measures shouldbe followed as prevention and to reduce epi-demic AKC outbreaks.

References

1. Araullo-Cruz T, Gordon YJ, Romanowski EG et al.(2000) The efficacy of topical cidofovir treatmenton cidofovir-resistant variants of adenovirus inthe AD5/NZW rabbit ocular model. Invest Oph-thalmol Vis Sci 41:ARVO Abstract, S59

2. BenEzra D, Pe’er J, Brodsky M, Cohen E (1986)Cyclosporine eyedrops for the treatment of se-vere vernal keratoconjunctivitis. Am J Ophthal-mol 101:278–282

3. Calne RY, White DJG, Thiru S et al. (1978) Cyclo-sporin A in patients receiving renal allograftsfrom cadaver donors. Lancet 2:1323–1327

4. Cohen EJ (1990) Is your office safe? Yes. Cornea9:S41–43


5. Dawson CR, Sheppard J (1995) Follicular con-junctivitis. In: Tasman W, Jaeger EA (eds) Duane’sclinical ophthalmology, vol 4, chap 7. Lippincott-Raven, Philadelphia, pp 2–8

6. Ford E, Nelson KE, Warren D (1987) Epidemiolo-gy of epidemic keratoconjunctivitis. EpidemiolRev 9:244–261

7. Fuchs E (1889) Keratitis punctata superficialis.Wien Klin Wochenschr 44:837–842

8. Gordon YJ, Romanowski E, Araullo-Cruz T et al.(1991) Inhibitory effect of (S)-HPMPC, (S)-HPM-PA, and 2’ nor-cyclic GMP on clinical ocular ade-noviral isolates is serotype-dependent in vitro.Antiviral Res 16:11–16

9. Gordon YJ, Romanowski EG, Araullo-Cruz T(1992) An ocular model of adenovirus type 5 in-fection in the NZ rabbit. Invest Ophthalmol VisSci 33:574–580

10. Gordon YJ, Romanowski E, Araullo-Cruz T, DeClercq E (1992) Pretreatment with topical 0.1%(S)-1-(3-hydroxy-2-phosphonylmethoxypropyl)cytosine inhibits adenovirus type 5 replication inthe New Zealand rabbit ocular model. Cornea11:529–533

11. Gordon YJ, Romanowski EG, Araullo-Cruz T(1994) Topical HPMPC inhibits adenovirus type 5in the New Zealand rabbit ocular replicationmodel. Invest Ophthalmol Vis Sci 35:4135–4143

12. Gordon YJ, Naesens L, DeClercq E et al. (1996)Treatment of adenoviral conjunctivitis with topi-cal cidofovir. Cornea [letter] 15:546

13. Gordon YJ, Araullo-Cruz T, Romanowski EG(1998) The effects of topical non-steroidal anti-inflammatory drugs on adenoviral replication.Arch Ophthalmol 116:900–905

14. Gottsch JD, Froggatt JW, Smith DM et al. (1999)Prevention and control of epidemic keratocon-junctivitis in a teaching eye institute. OphthalmicEpidemiol 6:29–39

15. Günther R (1939) Pathologisch-anatomischer Be-fund einer Hornhaut bei Keratitis epidemica.Klin Monatsbl Augenheilkd 103:309–314

16. Hillenkamp J, Reinhard T, Ross RS, Bohringer D,Cartsburg O, Roggendorf M, De Clercq E, Gode-hardt E, Sundmacher R (2001) Topical treatmentof acute adenoviral keratoconjunctivitis with0.2% cidofovir and 1% cyclosporine. Arch Oph-thalmol 119:1487–1491

17. Hillenkamp J, Reinhard T, Ross RS, Bohringer D,Cartsburg O, Roggendorf M, De Clercq E, Gode-hardt E, Sundmacher R (2002) The effects of cid-ofovir 1% with and without cyclosporine A 1% asa topical treatment of acute adenoviral kerato-conjunctivitis. Ophthalmology 109:845–850

18. Holland EJ, Olsen TW, Ketcham JM et al. (1993)Topical cyclosporin A in the treatment of anteri-or segment inflammatory disease. Cornea 12:413–419

19. Hutter H (1990) Keratokonjunktivitis epidemica:Therapieergebnisse während einer Epidemie.KlinMonatsbl Augenheilkd 197:214–217

20. Jawetz E, Kimura SJ, Hanna L et al. (1955) Newtype of APC virus from epidemic keratoconjunc-tivitis. Science 122:1190–1191

21. Kampf G, Rudolf M, Labadie JC et al. (2002) Spec-trum of antimicrobial activity and user accept-ability of the hand disinfectant agent SteriliumGel. J Hosp Infect 52:141–147

22. Kaneko H, Kato K, Mori S et al. (2001) Antiviralactivity of NMSO3 against adenovirus in vitro.Antiviral Res 52:281–288

23. Kowalski RP, Sundar-Raj CV, Romanowski EG etal. (2001) The disinfection of contact lenses con-taminated with adenovirus. Am J Ophthalmol132:777–779

24. Liegner JT, Yee RW, Wild JH (1990) Topical cyclo-sporine therapy for ulcerative keratitis associatedwith rheumatoid arthritis. Am J Ophthalmol109:610–612

25. Lund OE, Stefani FH (1978) Corneal histologyafter epidemic keratoconjunctivitis. Arch Oph-thalmol 96:2085–2088

26. Reilly S, Dhillon BJ, Nkanza KM et al. (1986) Ade-novirus type 8 keratoconjunctivitis – an outbreakand its treatment with topical human fibroblastinterferon. J Hyg (Lond) 96:557–575

27. Reinhard T, Sundmacher R (1999) Topical cyclo-sporine A in Thygeson’s superficial punctatekeratitis. Graefes Arch Clin Exp Ophthalmol 237:109–112

28. Reinhard T, Godehardt E, Pfahl HG, SundmacherR (2000) Lokales Cyclosporin A bei Nummulinach Keratoconjunctivitis epidemica. Eine Pilot-studie. Ophthalmologe 97:764–768

29. Roba LA, Kowalski RP, Gordon AT et al. (1995)Adenoviral ocular isolates demonstrate serotypedependent differences in in-vitro infectivity titersand clinical course. Cornea 14:388–393

30. Romano A, Revel M, Guarai-Rotman D et al.(1980) Use of human fibroblast-derived (beta) in-terferon in the treatment of epidemic adenoviruskeratoconjunctivitis. J Interferon Res 1:95–100

31. Romano A, Sadan Y (1988) Ten years of experi-ence with human fibroblast interferon in treat-ment of viral ophthalmic infections. Metab Pedi-atr Syst Ophthalmol 11:43–46

32. Romanowski EG, Roba LA, Wiley L et al. (1996)The effects of corticosteroids on adenoviral repli-cation. Arch Ophthalmol 114:581–585

References 171

33. Romanowski EG, Gordon YJ (2000) Efficacy oftopical cidofovir on multiple adenoviral sero-types in the New Zealand rabbit ocular model. In-vest Ophthalmol Vis Sci 41:460–463

34. Romanowski EG,Bowlin TL,Yates KA et al. (2004)Topical NMSO3 inhibits adenovirus replication inthe Ad5/NZW rabbit ocular model. Invest Oph-thalmol Vis Sci 45: ARVO Abstract 1657

35. Rossa V, Sundmacher R (1991) Lokale Interferon-Prophylaxe einer “epidemischen” Konjunktivitisdurch ein bislang nicht identifizierbares Virus.Klin Monatsbl Augenheilkd 199:192–194

36. Sauerbrei A, Sehr K, Eichhorn U et al. (2004)Inactivation of human adenovirus genome bydifferent groups of disinfectants. J Hosp Infect57:67–72

37. Shiuey Y, Ambati BK, Adamis AP, and the ViralConjunctivitis Study Group (2000) A random-ized double-masked trial of topical ketorolac ver-sus artificial tears for treatment of viral conjunc-tivitis. Ophthalmology 107:1512–1517

38. Sundmacher R (1982) Interferon in ocular viraldiseases. Interferon 4:177–200

39. Sundmacher R, Engelskirchen U (1991) ZumProblem der rezidivierenden und persistieren-den Nummuli nach Keratoconjunctivitis epidem-ica. Klin Monatsbl Augenheilkd 198:550–554

40. Sundmacher R, Hillenkamp J, Reinhard T (2001)Perspektiven von Therapie und Prophylaxe derAdenoviruskeratokonjunktivitis. Ophthalmologe98:991–1009

41. Sundmacher R, Wigand R, Cantell K (1982) Thevalue of exogenous interferon in adenovirus ker-atoconjunctivitis. Preliminary results. GraefesArch Clin Exp Ophthalmol 218:139–140

42. Trauzettel-Klosinski S, Sundmacher R, Wigand R(1980) Die Wirkung von Steroiden bei Keratocon-junctivitis epidemica. Ergebnisse einer kontrol-lierten prospektiven Studie. Klin Monatsbl Au-genheilkd 176:899–906

43. Trousdale MD, Goldschmidt PL, Nobrega R (1994)Activity of ganciclovir against human adenovirustype 5-infection in cell culture and cotton rateyes. Cornea 13:435–439

44. Tsai JC, Garlinghouse G, McDonnel PJ et al. (1992)An experimental animal model of adenovirus-in-duced ocular disease. The cotton rat. Arch Oph-thalmol 110:1167–1170

45. Vahlme A, Larsson PA, Horal P et al. (1992) Inhi-bition of herpes simplex virus production in vit-ro by cyclosporine A. Arch Virol 122:61–75

46. Ward JB, Siojy LG,Waller SG (1993) A prospective,masked clinical trial of trifluridine, dexametha-sone, and artificial tears in the treatment of epi-demic keratoconjunctivitis. Cornea 12:216–221

47. Wilhelmus KR, Dunkel EC, Herson J (1987) Topi-cal human fibroblast interferon for acute aden-oviral conjunctivitis. Graefes Arch Clin Exp Oph-thalmol 225:461–464

48. Zhao JC, Jin XY (1993) Immunological analysisand treatment of Mooren’s ulcer with cyclosporinA applied topically. Cornea 12:481–488


13.1Introduction

With good reason, biomicroscopy in ophthal-mology is dominated by slit-lamp examina-tions. The foundations laid by Gullstrand wereoutlined in exemplary fashion in Vogt’s Lehr-buch und Atlas der Spaltlampenmikroskopie deslebenden Auges [Textbook and Atlas of Slit-Lamp Microscopy of the Living Eye], first pub-lished in 1921 [19]. The second (1930) edition ofthat standard textbook placed particular em-phasis on information yielded by focal illumi-nation of scatters, described as Hornhautkör-perchen (corpusculi corneae) [78]. However, themaximum magnification achievable with thistechnique – approx. ¥40 – limited further subd-ifferentiation and did not reveal clinicomor-phologic correlations at the cellular level.

Specular microscopy was also first describedby Vogt [78], but did not gain popularity untilphotographic and video techniques alloweddocumentation and quantification of cornealendothelial cells [5, 7, 8, 31, 41]. Although thespecular microscope is useful for the in vivoexamination of the cornea, its applications arerestricted to the endothelial cell layer.

In 1968, the same year that Maurice de-scribed the first high-powered specular micro-scope [41], the first scanning confocal micro-scope was proposed [54]. This device wascharacterized by high z-axis resolution and pro-vided high-resolution microscopic images ofcells within living tissues of patients without theneed for fixation or staining.

In Vivo Micromorphology of the Cornea:Confocal Microscopy Principles and Clinical Applications

Rudolf F. Guthoff, Joachim Stave

13

|

∑ In vivo confocal microscopy permits celldensity quantification for all corneal subpopulations

∑ Tear film dynamics can be studied usingnon-contact confocal biomicroscopy,and the generation of dry spots can beevaluated

∑ Corneal epithelial thickness can be measured with high precision

∑ Dendritic cells can be displayed and quantified

∑ The influence of contact lenses on epithe-lial cell densities can be evaluated

∑ Inflammatory cells associated with bacteri-al and viral infections of the cornea can bedisplayed and quantified

∑ Follow-up after refractive corneal surgerypermits exact measurement of flap thick-ness and wound healing reactions in the interface

∑ Functional imaging at the cellular level using sodium fluorescein may be helpfulfor understanding metabolic activities incorneal epithelium

∑ Corneal nerves can be displayed three-dimensionally and perhaps quantified on the basis of nerve fiber density

Core Messages

Alongside this tandem scanning confocalmicroscope, a slit-scanning confocal design hasalso been produced [35, 58, 64, 73]. Currently, thetandem scanning design is manufactured by theAdvanced Scanning Corporation (New Orleans,LA) and the most recently developed version ofthe slit-scanning design is produced by NidekTechnologies Srl (Vigonza, Italy). The authorshave been accumulating experience with theslit-scanning design since 1994 [58, 64] and haveattempted to modify the system in order toachieve reproducible and quantifiable results[4].

13.2Principle of In Vivo Confocal MicroscopyBased on the Laser-Scanning Technique

The development of in vivo confocal laser-scan-ning microscopy in the late 1980s permittedprecise three-dimensional (3-D) visualization ofmicrostructures of the ocular fundus in partic-ular, with its optic nerve head and the peripap-illary retina. Modern digital image processingtechnology enables quantitative data to be col-lected non-invasively, rapidly and with a low

level of illumination. The precision of laser-scanning ophthalmoscopy in this context isbased on the principle of the confocality of theexamined object with the light source and thedetector plane. A laser light source is focusedthrough a pinhole diaphragm to one point onthe object. The reflected laser light is separatedby a beam splitter from the incident laser beampath and deflected through a second confocaldiaphragm to reach a photosensitive detector.Because of the confocal design, light originatingfrom outside of the focal plane is highly sup-pressed, and only the object layer located at thefocal plane contributes to the image. In order tobuild up a two-dimensional image perpendicu-lar to the optical axis of the device, the laserbeam has to scan the sample point by point.Thisis achieved by introducing two oscillating mir-rors into the beam path. Figure 13.1 provides a schematic illustration of this principle. By moving the focal plane optically, an image canbe acquired from a deeper layer of the examinedobject, thus enabling a data cube to be built upin a successive series.

By contrast, in slit-lamp biomicroscopyexamination of the cornea, an optical sectionthat is essentially perpendicular to the corneal

174 Chapter 13 In Vivo Micromorphology of the Cornea: Confocal Microscopy Principles and Clinical Applications

Fig. 13.1. Principle of confocal laser-scanning microscopy

surface is seen in up to ¥50 magnification or,with an additional lens for endothelial viewing(specular microscopy), in up to ¥200 magnifi-cation. Nowadays, documentation is generallyperformed using digital photography. All othercellular structures, e.g., the epithelium, cannotbe imaged with this technique because of thehigh proportion of scattered light.

Optical tomography perpendicular to the in-cident light path has only become possible withthe adaptation of confocal microscopy, as de-scribed above, for the examination of the livingeye [6, 9, 10, 28, 43, 73, 74]. This yields images ofthe endothelium, for example, that are com-parable to those obtained with specular mi-croscopy. In this case the most pronouncedsource of scattered light is the cytosol of theendothelial cells, with the result that the cellborders appear dark. Only light reflected fromthe focal plane contributes to the image. In thisway, cell structures in the stroma, nerves andcorneal epithelium [81] can also be imaged infine optical sections.

13.2.1Slit-Scanning Techniques

Techniques based on the principles of the rotat-ing NIPKOW disk or tandem slit-scanning wereused initially for confocal microscopy of theanterior segment of the eye. Figure 13.2 shows aslit-scanning microscope of this type incorpo-rating the use of a halogen lamp. For the pur-poses of corneal assessment these microscopescan be used to image confocal sections with anoptical layer thickness of approx. 5–10 mm.Synchronization of slit-scanning [35, 73, 80]with the video rate of a residual light camerayields sharp and motion-independent imagesequences at 25 frames/s [34, 36] (Fig. 13.3).

Three-dimensional image distortion due toeye movements with this non-contact micro-scopy technique can only be minimized byfaster image acquisition. However, more rapidmovement of the confocal plane along the opti-

cal axis (z-scan) is accompanied by a loss of res-olution. Loose optical coupling to the cornea us-ing a gel (Fig. 13.4) limits the precision of depthinformation relating to the optical section in thecornea, and hence the 3-D reconstruction of cellstructures [44, 53, 64]. Systematic errors such asinhomogeneous image illumination and imagedistortion also arise as a result of the electro-mechanical slit-scanning technique used. Lin-ear movement along the z-axis during individ-ual image acquisition also produces distortionalong the z-axis. These fundamental sources oferror can only be prevented by a rapid laserscanning system unhampered by mass inertiaand incorporating a serial dot raster techniquewith stepwise advance of the confocal planeduring the z-scan.

13.2 Principle of In Vivo Confocal Microscopy Based on the Laser-Scanning Technique 175

Fig. 13.2. Confocal slit-scanning microscope (ConfoScan 3/NIDEK)

13.2.2Laser-Scanning Microscopy and Pachymetry

As an alternative to confocal slit-scanning mi-croscopes, a confocal laser-scanning micro-scope for the anterior segment of the eye wasdeveloped at the Rostock Eye Clinic on the basisof an already commercially available laser-scan-ning system. Not least because of its compactconstruction, the Heidelberg Retina TomographII (HRT II, Heidelberg Engineering GmbH,Germany) was selected as the basic device for adigital confocal corneal laser-scanning micro-scope.

In laser-scanning ophthalmoscopy of theposterior segment, the optically refractive me-dia of the eye forms part of the optical imagingsystem. For anterior segment applications, ahigh-quality microscope lens is positioned be-tween the eye and the device, providing a laser


Fig. 13.3 A–D. Slit-scanning microscopy: A superficial and B basal cells; C nerve plexus with keratocytes and D endothelium

A B

DC

Fig. 13.4. Gel coupling between lens and eye (ConfoScan/NIDEK)

focus less than 1 mm in diameter. The result is ahigh-resolution, high-speed, digital confocallaser-scanning microscope permitting in vivoinvestigation of the cornea (Fig. 13.5). Move-ment of the confocal image plane inside thecornea can be achieved manually at the micro-scope lens or by using the automatic internal z-scan function of the HRT II. Laser-scanningtomography is consequently possible in theanterior segment of the eye.

This technique permits rapid and reliable vi-sualization and evaluation of all the microstruc-tures in the cornea, including the epithelium,nerves, and keratocytes, as well as the endothe-lium and bulbar conjunctiva. For the first time,the dendritic (or Langerhans’) cells can nowalso be visualized in vivo with an image qualitythat permits quantification [70,83]. In principle,any body surface that can be reached by the lenssystem is a suitable candidate for examination,with the result that potential applications alsoexist outside ophthalmology (e.g., the skin,tongue surface and oral mucosae).

The original functions of the basic HRT IIdevice for evaluating the optic nerve head inglaucoma are fully retained when the system ismodified into the confocal laser microscope.Software adapted to the special requirements ofscanning microscopy of the cornea has alsobeen developed. This permits the acquisition ofindividual section images, image sequences ofsection images, and volume images with inter-nal z-scan over a distance of approx. 80 mm. Thedigital properties of the device offer good pa-

tient and image data administration, also deliv-ering rapid access to, and hence comparisonwith, data from previous examinations.

The Heidelberg Retina Tomograph HRT IIhas been modified with a lens system attach-ment known as the Rostock Cornea Module(RCM; J. Stave, utility model no. 296 19 361.5,licensed to Heidelberg Engineering GmbH).The module is combined with a manual z-axisdrive to move the focal plane inside the cornea.This enables a cell layer at any depth to be im-aged and, for example, selected as the startingplane for the automated internal z-scan. Duringthe examination, pressure-free and centeredcontact with the cornea can be monitored visu-ally using a color camera.

The distance from the cornea to the micro-scope is kept stable using a single-use contactelement in sterile packaging (TomoCap). Opti-cal coupling is achieved via the tear film or byapplying protective gel to the eye (Fig. 13.6). TheTomoCap is a thin cap with a planar contact sur-face made from PMMA and is coupled opticallyto the lens with the aid of a gel.

In the 3-D imaging mode, the distance be-tween two subsequent image planes is approxi-mately 2 mm in the cornea. A 3-D image consistsof 40 image planes, thus covering a depth rangeof 80 mm. The acquisition time for a volume im-age is 6 s, and each individual section image isrecorded in 0.024 s. In the image sequence ac-quisition mode up to 100 images can be storedwith variable frame rates (1–30 frames/s). It thusbecomes possible to document dynamic pro-


Fig. 13.5. Rostock Cornea Module (RCM) (confocal laser-scanning microscope)

Fig. 13.6. TOMOCAP contact cap (Rostock Cornea Module)

cesses in the tissue (e.g., blood flow in the scle-ra).When the Rostock Cornea Module is used toset a plane manually at a desired depth, e.g., atthe LASIK interface after laser surgery to cor-rect refraction, image series from this depth canbe acquired with almost 100% image yield andprecise depth allocation [32, 45, 46, 77].

With the HRT II, z-axis movement betweenimages in the internal z-scan is performed – forthe first time – in a stepwise manner, i.e., duringacquisition of one section image the z-settingremains constant. This is a major advance and a prerequisite for generating distortion-free images of structure from one plane during the z-scan. A crucial prerequisite for undistort-ed 3-D reconstructions has therefore beenachieved [26].

A short focal length water immersion micro-scope lens with a high numerical aperture wasused to achieve high magnification (Achroplan63¥ W/NA 0.95/AA 2.00 mm, 670 nm, Carl Zeiss;alternatively, LUMPLFL 60¥ W/NA 0.90/AA2.00 mm, Olympus).

To optimize image quality, the Zeiss micro-scope lens was customized with special anti-re-

flection coating appropriate for the laser wave-length.

With the aid of an additional lens positionedbetween the HRT II and the microscope lens, thefield of view of the scanning system (fixed at 15°with the HRT II) is reduced to approx. 7.5° toallow for the necessary magnification (Fig. 13.7).Depending on the microscope lens and addi-tional lens used, the size of the field of view inthe contact technique can be 250 mm ¥ 250 mm,400 mm ¥ 400 mm, or 500 mm ¥ 500 mm. Dry mi-croscope lenses can be used in non-contact, forexample to image the tear film [66].

In particular, the compact construction ofthe HRT II simplifies its use as a confocal in vivomicroscope because the view is virtually unob-scured when monitoring the patient and bring-ing the microscope up to the cornea.

The precise perpendicular positioning of thecornea in front of the microscope within the mi-crometer range is facilitated by color cameracontrol. By observing the laser reflex on thecornea, even when bringing the microscope tothe eye, it is possible to make a lateral or verticalcorrection so that contact with the cornea is


Fig. 13.7. Rostock Cornea Module (RCM)

exactly in the optical axis (Fig. 13.8). As a result,images are captured only of cell structures thatare in a plane parallel to the surface, i.e., trans-verse sectional images are acquired. The contacttechnique guarantees a fixed distance between

microscope and cornea. The precise movementof the focal plane through the cornea with si-multaneous digital recording of depth positionrelative to the superficial cells of the epithelium(at the corneal surface) thus makes exactpachymetry possible.

13.2.3Fundamentals of Image Formation in In Vivo Confocal Microscopy

The laser light emitted constitutes an electro-magnetic wave with a defined wavelength lamb-da (l). This light wave is modified on passagethrough the cornea. Part of it migrates un-changed through all layers (transmission). Atinterfaces with changing refraction indices(nD), the wave alters direction due to scatter andrefraction. Light scattering is the basis of imageformation in confocal microscopy where onlybackscattered light is used for the image. Theamount of backscattered light is very small be-cause scatter occurs predominantly in a forwarddirection [14]. Light-scattering interfaces arefound in the cornea, for example, at the junctionbetween cytoplasm or extracellular fluid withnD=1.35–1.38 and lipid-rich membranes withnD=1.47 in the form of cell borders, cell nucleusmembranes and mitochondrial membranes[63] (Fig. 13.9). The amount of backscatteredlight depends on the structure of the interfacesurfaces.Rough surfaces scatter light in a broad-ly diffuse pattern, whereas directed waves withnarrow scatter cones are formed on smoothstructures. In addition, the confocal image is in-fluenced both by the number as well as the sizeand orientation of scattering organelles or par-ticles. Objects whose diameters are of the sameorder of magnitude as the wavelength of thelaser light display Mie scatter, and very muchsmaller molecules display Rayleigh scatter,which is directed backwards to a greater extentthan Mie scatter. Different orientations of elon-gated cell organelles can place different particlecross-sections in the path of the incident lightbeam, the result being that the light is backscat-tered to varying degrees. A high proportion ofcell organelles also increases the amount ofbackscattered light [14].


Fig. 13.8 A–C. Camera control: laser reflexes on thecornea – producing the immersion gel bridge be-tween the objective lens/cap and the cornea

A

B

C

13.3General Anatomical Considerations

The corneal epithelium consists of five to six lay-ers of nucleated cells that can be subdividedfunctionally and morphologically into threezones:∑ Superficial cells: approx. 50 mm frontal diam-

eter and approx. 5 mm thick. About 1/7 ofthese cells is lost by desquamation within24 h. Before detachment the cytoplasm andnucleus undergo a change in their opticalcharacteristics.

∑ Intermediate cells: 50 mm diameter and10 mm thick. These cells form a contiguouspolygonal, wing-shaped pattern (wing cells).

∑ Columnar basal cells have a flat basal surface,adjacent to Bowman’s membrane, a frontalheight of approx. 20 mm and a frontal diame-ter of 8–10 mm. Like endothelial cells, theycan be quantified accurately because of theirdefined location in relation to the basementmembrane (Fig. 13.10).

Bowman’s membrane – which is clearly distincthistologically from the epithelial basementmembrane – is 10–16 mm thick and remains

amorphous on light microscopy. Its location onin vivo confocal microscopy is well defined bythe subepithelial plexus (SEP).

The stroma accounts for some 90% of totalcorneal volume. Ninety-five percent of the stro-ma consists of amorphous ground substance(glycoproteins, glycosaminoglycans: keratansulfate and chondroitin sulfate) and collagenfibers. The remaining 5% of stromal volume isaccounted for by cellular structures known askeratocytes, which are specialized fibroblasts.Besides the nerves, their irregularly shapednuclei are the only well-defined sources of scat-tered light in corneal stroma detected on confo-cal microscopy. Their widely branching cyto-plasmic extensions are not visible in the healthycornea (Fig. 13.11).

The cornea is the most densely innervatedtissue in the human body. It is supplied by theterminal branches of the ophthalmic nerve inthe form of 30–60 non-myelinated ciliary nerves(Fig. 13.12).


Fig. 13.9. The confocal image of one section (x, y) isproduced by the sum of the backscattered light inten-sities (IR) from the focal depth range (z) (IV forwardscatter, IT transmission)

Fig. 13.10. Schematic illustration of the corneal epi-thelium and upper corneal stroma

In the limbus region these are seen aswhitish, filigree-like structures; their complexstromal and epithelial branchings are not visi-ble by slit-lamp microscopy, but are relativelyclear on confocal microscopy.

Like Bowman’s membrane, Descemet’s mem-brane, which should be regarded as the base-ment membrane of the endothelium, remainsamorphous on light microscopy. It is 6–10 mmthick. On confocal microscopy it is defined opti-cally by the easily identifiable endothelial cells.

The endothelium consists of about 500,000hexagonal cells approx. 20 mm in diameter, 5 mmthick and with large, flattened, central nuclei.The high concentration of cell organelles is in-dicative of very intensive metabolic activity.

13.4In Vivo Confocal Laser-ScanningMicroscopy

The layered structure of the epithelium of theeye can be visualized with high contrast usingthe Rostock Cornea Module (RCM) attachmentwith the HRT II and, because of the good quali-ty of depth resolution, can be imaged in opticalsections a few micrometers thick. The same istrue for the subepithelial plexus, the entire stro-ma including the keratocytes, and the endothe-lial fine structure (Fig. 13.13).

13.4 In Vivo Confocal Laser-Scanning Microscopy 181

Fig. 13.12. Schematic illustra-tion of the subepithelial plexus(SEP) and its branches in thecorneal epithelial layers (BEPbasal epithelial plexus)

Fig. 13.11. Schematic illustration of the layeredstructure of the human cornea. The differentlyshaped keratocyte nuclei can be distinguished on invivo confocal microscopy (adapted from Krstiè)

13.4.1Confocal Laser-Scanning Imaging of Normal Structures

13.4.1.1Tear Film

The pre-ocular tear film with its complex fluidstructure bathes the cornea and conjunctiva.Tear film structure and function are maintainedby a highly differentiated system of secretory,distributive and excretory interactions [49, 60].In particular, these function to smooth thecorneal surface and maintain its optical clarity.

The water content of the cornea is regulated byevaporation and the resultant osmotic gradient.The oxygen in the air is dissolved in the tear flu-id and thus supports the aerobic metabolism ofthe epithelium.

The tear film is 7–10 mm thick and is charac-terized by a three-layered structure. The ex-ternal lipid layer, which is produced chiefly bythe meibomian glands close to the margin of the eyelids, prevents rapid evaporation ofthe aqueous layer and renders the surfacehydrophobic. The inner mucin layer consists ofglycoproteins. Its task is to make the epithelialsurface hydrophobic and thus to guaranteewettability.


Fig. 13.13. A Superficial cells; B wing cells; C basal cells;D nerve plexus; E keratocytes/anterior stroma;F keratocytes/posterior stroma; G endothelium

A B C

FED

G

Replacing the contact system in the confocallaser-scanning microscope with a dry objectivelens (Fig. 13.14) enables the fine structure of thetear film to be imaged (Fig. 13.15). The rapidimaging sequence in the device also permitsdynamic processes to be recorded [39, 65, 75].

13.4.1.2Epithelial Layer

13.4.1.2.1Superficial Cells (up to approx. 50 µm in Diameter)

In the case of the most superficial epithelialcells, bright cell borders and a dark cell nucleusand cytoplasm are readily visualized on con-focal laser-scanning microscopy. The cells characteristically display a polygonal – oftenhexagonal – shape.Cells undergoing desquama-tion are characterized by a highly reflectivecytoplasm, in the center of which the brightlyappearing (pyknic) cell nucleus with its darkperinuclear space is clearly visible (Fig. 13.16).The average density of superficial cells in thecentral and peripheral cornea is approx. 850cells/mm2.


Fig. 13.14. Non-contact microscopy: laser reflex onthe cornea

Fig. 13.15. Normal tear film

Fig. 13.16. Superficial cells: the cytoplasm and cellnuclei are visualized; cells in the process of desqua-mation possess a highly reflective cytoplasm, in thecenter of which the bright (pyknic) cell nucleus withits dark perinuclear space is clearly visible (z=50 mm)

13.4.1.2.2Intermediate Cells/Wing Cells (up to Approx. 20 µm in Diameter)

The cells of the intermediate layers are charac-terized by bright cell borders and a dark cyto-plasm. The cell nucleus can be distinguishedonly with difficulty. In terms of size and appear-ance, wing cells in healthy subjects exhibit onlyminimal variation (Fig. 13.17). The average celldensity is approx. 5,000 cells/mm2 in the centralcornea and approx. 5,500 cells/mm2 in the pe-riphery.

13.4.1.2.3Basal Cells (up to Approx. 10 µm in Diameter)

The basal cells are located immediately aboveBowman’s membrane. They present as brightlybordered cells in which the cell nucleus is notvisible. Between-cell comparison reveals inho-mogeneous reflectivity of the cytoplasm. Likethe wing cells above them, the basal cells displayonly minimal variation in shape and size(Fig. 13.18). The average cell density is approx.9,000 cells/mm2 in the center of the cornea and10,000 cells/mm2 in the periphery. In normalsubjects, therefore, in terms of cell densities, the

ratio between superficial cells, intermediatecells and basal cells is 1:5:10.

13.4.1.3Langerhans’ Cells

Confocal microscopy permits in vivo evaluationof Langerhans’ cells (LCs) within the humancornea, with a particular emphasis on cell mor-phology and cell distribution.

LCs present as bright corpuscular particleswith dendritic cell (DC) morphology and adiameter of up to 15 mm. LC distribution followsa gradient from low numbers in the center tohigher cell densities in the periphery of thecornea. Moreover, in vivo confocal microscopypermits differentiation of LC bodies lackingdendrites, LCs with small dendritic processesforming a local network, and LCs forming a wirenet via long interdigitating dendrites (Fig. 13.19).While almost all the cells located in the periph-ery of the cornea demonstrate long processes in-terdigitating with the corneal epithelium, thosein the center of the cornea often lack dendrites,most probably underlining their immature phe-notype [21]. Immature LCs are equipped to cap-ture antigens, while mature forms are able tosensitize naive T-cells through MHC moleculesand secretion of interleukin-12 as well as costim-


Fig. 13.17. Intermediate cells: the cells of the interme-diate layers are characterized by bright cell bordersand a dark cytoplasm. The cell nucleus can be identi-fied only with difficulty. The wing cells display onlyminimal variation in terms of size and appearance

Fig. 13.18. Basal cells: these are regularly arrangedcells with bright borders, but the cell nucleus is notvisualized. Intercellular comparison reveals inhomo-geneous cytoplasmic reflectivity

ulatory molecules,and thus represent an integralpart of the immune system [3].

The average density of LCs in normal sub-jects is 34±3 cells/mm2 (range: 0–64 cells/mm2)in the central cornea and 98±8 cells/mm2

(range: 0–208 cells/mm2) in the periphery [83].In contact lens wearers, LC density varies from60±16 cells/mm2 (range: 0–600 cells/mm2) inthe central cornea to 159±18 cells/mm2 (range:0–700 cells/mm2) in the periphery. LC densitiesdiffer significantly between healthy volunteersand contact lens wearers both in the central(p=0.03) and in the peripheral cornea(p=0.001),while the gradient of LC density fromperiphery to center was almost identical in bothgroups (unpublished data).

It has been suggested that LCs participate inimmune and inflammatory responses, therebydetermining cell-mediated immunity. In light ofthis theory, the present data on LCs in the hu-man cornea provide a helpful basis for furtherinvestigations in ocular pathology.

13.4.1.4Corneal Nerves

The cornea is one of the most sensitive struc-tures in the human body, and even the mostminimal contact provokes the lid reflex to pro-tect the eye. This sensitivity is attributable to thelarge numbers of nerve fibers that pass throughthe cornea. Furthermore, the corneal nervesexert an influence on the regulation of epithelialintegrity and on wound healing.

In vivo visualization of these nerve struc-tures is possible by confocal corneal micro-scopy.

The cornea is innervated primarily by senso-ry fibers arising from the ophthalmic nerve, aside branch of the trigeminal nerve. Humancorneal nerves are non-myelinated and vary inthickness between 0.2 and 10 mm.

The nerve fiber bundles, which enter the an-terior and central stroma in the corneal periph-ery, run parallel to the corneal surface in a radi-al pattern before making an abrupt 90° turn inthe direction of Bowman’s membrane [47]. Onconfocal corneal microscopy these nerve fibersmostly present as thick, almost always stretched,highly reflective structures (Fig. 13.20). Fre-quently, the stromal nerves are found in closeproximity to keratocytes. The deep stroma isdevoid of nerves that can be visualized on con-focal microscopy.

In the anterior stroma, immediately beforeBowman’s membrane, the nerve fiber bundlesdisplay three different patterns. Some of thenerve fibers ramify before reaching Bowman’smembrane without penetrating it and form thesubepithelial plexus [51] (see schematic illustra-tion in Fig. 13.12). Other nerves penetrate Bow-man’s membrane either directly following aperpendicular or slightly oblique course, or justbefore penetration they ramify into several finebranchlets. After they have penetrated Bow-man’s membrane, they again make a 90° direc-tional change and pass between the basal celllayer of the epithelium and Bowman’s mem-


Fig. 13.19. In vivo confocal microscopic images, representing different forms of Langerhans’ cells: A individ-ual cell bodies without processes; B cells bearing dendrites; C cells arranged in a network via long interdigitat-ing dendrites

A B C

brane toward the corneal center and form thebasal epithelial plexus (see schematic illustra-tion in Fig. 13.12). In so doing, they give off manysmall side branchlets directed both toward thecorneal surface, where they end freely, and to-ward the center [47, 48]. The nerve fibers of thebasal epithelial plexus mostly run parallel toeach other and often form Y- or T-shapedbranches. Their predominantly granular,“stringof pearl” structure is characteristic; more rarelythey display a smooth surface. Unlike the stro-mal nerves, they are characterized by lesser re-flectivity and frequently follow a meanderingpath (Fig. 13.21). Occasionally, a thicker nervefiber bundle will divide into two finer nervefibers, before these then reunite after a shortdistance into a single nerve fiber with the samethickness as before (Fig. 13.21A). The finerbranchlets also form connections between larg-er nerve fibers (Fig. 13.21A, B).


Fig. 13.20. Nerve fibers (arrowed) in the anteriorcorneal stroma. The stretched pattern of the stromalnerves is characteristic. The keratocyte nuclei areidentifiable as hyperreflective oval structures, someof which are in close proximity to the nerve (star)

Fig. 13.21 A, B. Highly reflective nerves from thebasal epithelial plexus located between Bowman’smembrane and the basal cell layer of the corneal ep-ithelium. The nerve fibers have their characteristic

granular “string of pearl” appearance. In most casesthey run parallel to each other and show T (circle)-and Y (star)-shaped nerve branchings that may pro-duce connections between larger nerve fibers

A B

13.4.1.5Bowman’s Membrane

The anterior limiting membrane has an amor-phous appearance. Its location can be estab-lished from the nerves of the basal epithelialplexus, which ramify there (Fig. 13.22).

13.4.1.6Stroma

Apart from neural structures, only the highlyreflective, sharply demarcated cell nuclei of thekeratocytes are visualized on examination ofthe stroma. The cytoplasm of this fibroblastsubpopulation and the collagen fibers producedby them are not visible. Keratocyte nucleus den-sity is higher in the anterior stroma close toBowman’s membrane than in the central anddeep stroma (Figs. 13.23, 13.24). Keratocytedensity is highest in the anterior stroma, clearlydeclines toward the central stroma, and increas-es again slightly in the region immediatelybefore Descemet’s membrane.


Fig. 13.22. Subepithelial nerve Fig. 13.23. Anterior stroma: in corneal stroma onlythe keratocyte nuclei are visualized; the density of thecell nuclei is highest of all in the anterior stroma (seeFig. 13.24); the size of the cell nuclei shown is approx.15 mm

Fig. 13.24. Central stroma: clearly demarcated,highly reflective, oval-shaped nuclei of keratocytes inthe central stroma; here cell nucleus density is thelowest in corneal stroma

13.4.1.7Descemet’s Membrane

Like Bowman’s membrane, Descemet’s mem-brane has an amorphous appearance and istherefore not visualized in healthy subjects.

13.4.1.8Endothelial Cells

The endothelium consists of a regular pattern ofhexagonal reflective cells. The cell nuclei cannotusually be visualized. The cell borders reflectless light than the cytoplasm,with the result thata network of dark cell borders appears betweenareas of bright cytoplasm. Endothelial cell den-sity can be determined by counting (Fig. 13.25).

13.4.1.9Limbal Region

Because it forms a junctional zone with the con-junctiva, the limbal region is especially impor-tant. Inflammatory cells migrate across it intothe cornea in immunological disease, it is thesource of new inbudding corneal vessels and,not least, it also plays an important role incorneal regeneration as the site of origin ofcorneal stem cells.

The limbal region is where the cornealepithelium forms a junction with the conjuncti-val epithelium which comprises approx. 10–12cell layers. This region also contains a radialarrangement of trabecular conjunctival pro-cesses (the limbal palisades of Vogt) that areconsidered to be the site of origin of cornealstem cells [12, 61]. Overall, the organization ofthe conjunctival epithelium is less uniform be-cause different epithelial cell types (e.g., gobletcells) occur here and the arrangement of the in-dividual cell layers is also not so strictly parallelwith the surface [13].

On confocal microscopy, the epithelial cellsof the conjunctiva, unlike those of the cornea,are more reflective, smaller and less well demar-cated. Their nucleus is relatively large andbright. The junctional zone is characterized byinhomogeneous reflectivity and marked varia-tion in cell shape and size (Fig. 13.26). The lim-bal palisades of Vogt can often be visualized asparallel trabecular extensions of the conjunc-tival epithelium (Fig. 13.27). In the immediatejunctional zone the conjunctival epitheliumalso commonly exhibits tongue-like extensionswhich are mostly well demarcated, especially inthe deeper layers, and at the end of which arelocated isolated cells or cell groups with verybright cell borders and a bright cytoplasm(Fig. 13.28). These may be secretory cells. Sub-epithelially, in the region of the conjunctivaclose to the limbus, are the blood vessels of thelimbal vascular plexus, in the lumen of whichflowing blood cells can be seen (Fig. 13.29).


Fig. 13.25. Endothelium: a monolayer of regularlyarranged hexagonal cells completely covering theposterior surface of the cornea. Unlike the basal cells(see Fig. 13.18), these cells have a brightly reflectingcytoplasm and dark cell borders. The cell nucleus isnot visible


Fig. 13.26. Superficial epithelium of: A conjunctiva (cells up to approx. 30 mm in diameter, bright large cellnuclei; B transitional zone (variable morphology; C cornea (cells up to approx. 50 mm in diameter, bright cellborders) (x= Position on the cornea)

Fig. 13.27. Limbal palisades of Vogt. Trabecular ex-tensions of the conjunctiva growing from outside (inthis case from below) in a radial pattern toward thecornea

Fig. 13.28 A–C. Extensions of conjunctival epithelium at different depths. In the center of the image in A a cellgroup with strikingly bright cell borders is shown (arrowed). The basal conjunctival epithelium in C is muchbrighter with the result that structures are no longer identifiable

Fig. 13.29. Branched conjunctival vessel close to thelimbus with erythrocytes visible (z= 100 mm, depth oflaser focus in the cornea)

A

A B C

B C

13.5Clinical Findings

13.5.1Dry Eye

Disturbances of tear film secretion or tear filmstructure give rise to a condition known as dryeye. On microscopy such disturbances are evi-dent as altered reflection or dry spots on theepithelium (Fig. 13.30).

As the most important component of thecorneal diffusion barrier, the corneal epitheli-um displays differing permeability for aqueousionic substances such as sodium fluorescein(NaF). Patients with diabetes, for example,have significantly increased permeability forNaF. NaF also penetrates areas of micro-ero-sions and pathologically altered cells. The liter-ature reveals discrepant views concerning thenature of the penetration process. Most authorssubscribe to the view that the fluorescein fillsthe “footprint” spaces vacated by cells that havebeen lost. However, others assume that fluores-cein fills the intercellular space. At present onlyconfocal slit-scanning microscopes and fluo-rophotometers are used to analyze this phe-nomenon.

The Heidelberg Retina Angiograph (HRA) incombination with the Rostock Cornea Modulecan be used for confocal laser-scanning fluores-cence microscopy of the microstructure of thecorneal epithelium and tear film using contactand non-contact techniques (Fig. 13.31) with alateral resolution of 1 mm and up to ¥1,000 mag-nification. The red-free reflection and fluores-cence images display the intercellular mi-crostructure with stained cell nuclei and alteredcell surfaces and borders. The same area exam-ined on the cornea can be visualized simultane-ously in reflection and fluorescence mode. Thepenetration profile of NaF can be measuredwith precise depth resolution over a prolongedtime using the contact technique. Autofluores-cence measurements are also possible. Fig-ure 13.32 shows the tear film (a) and epithelium(b) in fluorescence mode.

13.5.2Meesmann’s Dystrophy

Meesmann’s dystrophy is a rare, bilaterally sym-metrical epithelial condition inherited as anautosomal dominant trait and attributable tomutations in the keratin 3 (K3) gene [11] orkeratin 12 (K12) gene [82] on chromosome 12[27] or chromosome 17 [71]. Due to the ruptureof mainly interpalpebral epithelial cysts, thecondition causes episodic pain, photophobia,epiphora, blepharospasm and fluctuating visualacuity or a moderate decline in visual acuity,even in young children. No causal therapyexists.

13.5.2.1Summary Evaluation

Clinically, Meesmann’s dystrophy is character-ized bilaterally by small cystic changes in the corneal epithelium, particularly in the interpalpebral zone, and individual superficialpunctate opacities. Histological assessment andelectron microscopy reveal exclusively intra-epithelial cysts with cell debris (clumpedkeratin) which migrate to the corneal surfaceduring normal epithelial regeneration andrupture there (Fig. 13.33). In addition, there are


Fig. 13.30. Tear film/dry spots

irregular cell arrangements and granular de-posits in the basal cells and a thickened base-ment membrane. The other corneal layers donot show any changes.

Cystic epithelial changes consistent with thehistological findings can be visualized in vivousing the Rostock confocal laser-scanning mi-croscope (Figs. 13.34, 13.35), and thus this non-invasive method contributes to confirming thediagnosis.

13.5 Clinical Findings 191

Fig. 13.31. Heidelberg Retina Angiograph HRA/RCM. Fluorescence mode: blue argon laser line; reflectionmode: green argon laser line

Fig. 13.32. A tear film (HRA Classic); reflection mode; B epithelium/NaF-stained (HRA Classic); fluorescencemode

A B

13.5.3Epithelium in Contact Lens Wearers

Distinct changes in corneal morphology,pachymetry and structure in contact lens wear-ers can be demonstrated by in vivo confocallaser scanning microscopy. These findings arebest interpreted as resulting from mechanicalor metabolic disturbances of the cornea.

All cell layers (superficial, intermediate andbasal cells) are present and characterized bybright cell borders and uniformly dark cyto-


Fig. 13.35. Histologic findings in Meesmann’s dys-trophy (after Naumann: Pathologie des Auges [49])with intraepithelial cysts containing cellular debris

Fig. 13.33. A B.N., 12 years old: microcystic epithelial changes in the interpalpebral zone, isolated superficialpunctate opacities, otherwise normal corneal structures. B Retroillumination (cf. A)

A B

Fig. 13.34. A B.N., 12 years old: confocal microscopyof the epithelium in vicinity of the superficial cells(depth: 5 mm) showing cystic structures with spheri-cal, highly reflective contents similar to the histologic

sections (see Fig. 13.35). B B.N., 12 years old: confocalmicroscopy of the epithelium at a depth of 30 mm withincreased visualization of spherical highly reflectiveand cystic structures

A B

plasm. The cell count increases with layer depthdue to a decrease in cell diameter. Bowman’smembrane and the subepithelial plexus displayborder structures between epithelium and stro-ma (Fig. 13.36A).

Superficial cells are characterized by a darknucleus, and the cytoplasm is generally darkerthan in the normal cornea. The polygonal struc-ture is retained, but cell bodies are generallysmaller (30 mm in contact lens wearers and up to50 mm in the normal cornea) (Fig. 13.36B).

Our data (unpublished results) show a sig-nificant increase in superficial cell density(p<0.05) both centrally and peripherally.

The intermediate cells do not show any morphological changes by comparison withfindings in normal subjects: pale cell borders,invisible nucleus and dark cytoplasm weredetected in both the lower and upper wing cells (Fig. 13.36C). A significant reduction in thecell count was noted only in the periphery(p<0.05).


Fig. 13.36. Epithelium in contact lens wearers: A oblique corneal section: superficial, intermediate and basalcells, Bowman’s membrane and anterior stroma; B superficial cells; C intermediate cells; D basal cells

A B

DC

Basal cell structure is characterized by an in-homogeneous cytoplasm and invisible nucleus;cell diameters are approx. 8–10 mm (Fig. 13.36D).A significant reduction in the cell count was alsodetected in the peripheral cornea (p<0.05).

Analysis of the pachymetry data revealed re-duced corneal thickness in the periphery com-pared to that in normal volunteers, especially inpatients who had worn contact lenses for longerthan 10 years. There were no age-related changesin cell count or epithelial thickness, but stromalthickness was reduced.

The type of contact lens (hard vs. soft) has noinfluence on corneal morphology; duration ofcontact lens wear was the factor with the great-

est impact. Corneal microdeposits in stroma(Fig. 13.37A) and signs of polymegathism, pleo-morphism (Fig. 13.37B, C) and endotheliumprecipitates (Fig. 13.37B) are the most commonfindings.

As demonstrated in Fig. 13.38, alterations inLangerhans’ cells also occur due to contact lenswearing.

In light of this, in investigations of the corneain contact lens wearers, attention must focus onthe cell density of each layer and on the thick-ness of the corneal epithelium, and results mustalways be compared between the center andperiphery.


Fig. 13.37. Micromorphologic changes in thecornea of contact lens wearers: A corneal micro-deposits; B signs of polymegathism, pleomorphismand endothelium precipitates; C signs of poly-megathism and pleomorphism

A

C

B

13.5.4Epidemic Keratoconjunctivitis

Epidemic keratoconjunctivitis (EKC) is a highlycontagious infection caused by type 8, 19, 37adenoviruses; one of its chief complications isthe development of nummular areas of sub-epithelial corneal opacity which, in exceptionalcases, may lead to years of reduced visual acuityand to increased glare sensitivity [30, 69].Histopathologic examination reveals that thenummular lesions consist of an accumulation ofcells from the monocyte-macrophage system,such as lymphocytes, histiocytes and fibro-blasts [18].

Viral persistence in the keratocytes is sus-pected as the cause for the continuing presenceof the nummular lesions. The immune responseinduces focal infiltration of immune cellsaround the infected keratocytes. The complexes

form the slit-lamp microscopic substrate of thenummular lesions [57] (Fig. 13.39A, B).

Hyperreflective punctate structures can bevisualized in the intermediate layer of the epi-thelium on confocal Rostock laser-scanning mi-croscopy (Fig. 13.40A). These may be lympho-cytes, histiocytes and/or fibroblasts [33].

By contrast with physiologic findings, thebasal cell layer is hardly distinguishable as such.In addition to a network of hyperreflectivedendritic structures (Fig. 13.40B), which be-comes clearly less dense with increasing depth(Fig. 13.40C, D), corpuscular changes with den-dritic extensions are visualized (Fig. 13.40C, D),some of which appear to be spread out betweenthe nerve fibers (Fig. 13.40E). Considering theirlocation, size and shape, these are most proba-bly the antigen-presenting Langerhans’ cells[59], which are responsible for the induction ofcell-mediated delayed-type immune responses.They assume an important role in triggering


Fig. 13.38. Langerhans’ cells and reflective keratocytes after contact lens wearing (3 years)

contact allergies, rejection reactions and viraldefense, and in the healthy cornea they are lo-cated in the epithelial layers of the conjunctiva,the limbus and peripheral cornea, but not in thecentral cornea. Migration of the Langerhans’cells into the central cornea may occur in re-sponse to traumatic, chemical or inflammatorystimuli [1, 18].

The changes visible beneath the nerve fiberlayer are possibly scatter artifacts in the vicinityof the ruptured Bowman’s membrane [33].

The superficial epithelial layer, stroma andendothelium do not display any abnormalities.

13.5.5Acanthamoeba Keratitis

Numerous free-living phagotrophic amoebaecause opportunistic infection in humans. Acan-thamoeba keratitis has been recognized as apotentially blinding disease, which is often onlydiagnosed at a late stage. The condition is some-times confused with other types of infectiouskeratitis, particularly those of fungal and her-petic origin.


Fig. 13.39 A, B. Slit-lamp microscopy photograph: righteye of a 28-year-old female patient on day 14 after the onsetof symptoms of epidemic kera-toconjunctivitis, showing thesubepithelial nummular lesionsas fleecy-fused areas of opacitywith unclear margins: A slit-lamp microscopy and B with the “Pentacam” Scheimpflugcamera (Oculus)

Fig. 13.40 a–e. Confocal image of the central corneain epidemic keratoconjunctivitis; edge length of theimage in vivo,250 mm; focal planes moved axially fromepithelium to endothelium. a Intermediate epitheliallayer with isolated hyperreflective round structures

located between the cells; b basal cell layer with hy-perreflective dendritic network; c, d transition frombasal cell layer to nerve plexus layer with dendriticcell structures, some of which appear to be spread outbetween the nerve fibers; e Bowman’s membrane

A B

Although not widely available, the confocalmicroscope can be helpful in establishing thediagnosis of Acanthamoeba keratitis, based onthe visualization of pear-shaped cysts approx.10 mm in length and irregular trophozoites [38,55] (Figs. 13.41–13.43).

In vivo confocal microscopy permits identi-fication of Acanthamoeba cysts in the cornea[40, 2]. The identity of findings with those fromconventional ex vivo microscopy and PCR pro-vides a basis for simple and reliable in vivodiagnosis (Fig. 13.44).


Fig. 13.42 A, B. Corneal microcysts (cystic stage of life cycle, round in shape, up to 10 mm, double wall) are visible at the level of the deep intermediate and basal cells (z=32 mm) (A) and in the anterior stroma (z=93 mm) (B)

BA

Fig. 13.41. Slit-lamp photograph from a 42-year-old female patient with a unilateral red, painful eye:A with epithelial defects, stromal ring infiltrate; B fluorescein staining positive; sensibility decreased. PCR (herpes zoster) and corneal scrapings (pathological agents including Acanthamoeba) negative

A B


Fig. 13.43 A, B. The same areas of the cornea 3 months after specific therapy (propamidine isethionate/Brolene): no signs of cysts either in the epithelium (z=22 mm) or in the stroma (z=70 mm) . The stromal archi-tecture is highly irregular (A, B)

A B

Fig. 13.44 A–C. Confocal microscopy as a non-inva-sive diagnostic method for in vivo identification ofAcanthamoeba cysts in the cornea. The identity offindings with conventional ex vivo microscopy pro-

vides a basis for easy and reliable in vivo diagnosis ofAcanthamoeba cysts. A Light microscopy in vitro;B confocal microscopy ex vivo; C confocal micro-scopy in vivo

A B C

Fig. 13.45 A, B. Slit-lamp photograph from a 70-year-old female patient with a unilateral red, painful eye withepithelial and stromal defects: A infiltrated and blurred cornea in the ulcer area; B fluorescein staining positive

A B

13.5.6Corneal Ulcer

Little experience has been gained with confocalmicroscopy in unspecific corneal ulcers. Leuko-cyte infiltration may be demonstrated in theulcer margins in both the epithelial and the

superficial stromal region. Figure 13.45 is a slit-lamp photograph and Fig. 13.46 shows confocalmicroscopy. In vivo confocal microscopy ofcorneal ulcers provides additional informationabout corneal healing processes, and permitsevaluation of epithelialization and reinnerva-tion at the cellular level.


Fig. 13.46. A Oblique section of central cornea nearthe ulcer: regular epithelial structure, absence ofsubepithelial plexus, and distortion of anterior stro-ma; B at the level of deeper intermediate and basalcells: bright cellular structures, most probably leuko-

cytes; C distortion of basal cell layer and other struc-tures of the subepithelial plexus; D anterior stroma inulcer area: keratocytes or cell nuclei are not visible,severe destruction of stromal structure

A B

DC

13.5.7Refractive Corneal Surgery

The different methods of refractive corneal sur-gery are designed to reduce ametropia, wherepresent. Depending on the technique used, re-fractive corneal surgery may result in morpho-logic changes and sometimes also in irritationand complications in the vicinity of the cornealepithelium or stroma that may lead to sub-jective disorders [56, 50]. The morphology and mechanism of wound healing processes following refractive corneal surgery are there-fore of particular interest in this context [25, 15](Figs. 13.47–13.49).

In vivo confocal microscopy of the cornea af-ter refractive surgery yields information aboutthe functional status of the keratocytes and thereinnervation of stroma and epithelium [37]. Itis possible to define the precise depth locationfor corneal opacities and to measure changes incorneal thickness [45]. Even years later, thedepth of the interface zone following laser in-situ keratomileusis (LASIK) can be identifiedon the basis of the morphologic changes visiblethere.


Fig. 13.47. Slit-lamp photograph from a 25-year-oldmale patient 1 year after laser in-situ keratomileusis(LASIK): uncorrected visual acuity 20/20, normalcorneal morphology apart from a small circular stro-mal scar, representing the border of the former flapzone (arrow)

Fig. 13.48 A, B. Confocal images from the same patient: A level of the interface zone with diffuse hyperreflec-tion and “microdot” structures; B region of the circular stromal scar with evidence of reinnervation of the flap(arrow) (z=170 mm (A); z=85 mm (B))

A B

13.6Future Developments

13.6.1Three-Dimensional Confocal Laser-Scanning Microscopy

Many researchers have investigated the corneawith in vivo confocal microscopy [4, 22, 23, 24,37, 72]. This sophisticated tool has been useful inaugmenting our understanding of anatomy inthe healthy and diseased human cornea. Thelimitations in magnification due to slight,unavoidable eye movements are obvious andtherefore 3-D reconstruction is restricted onpractical grounds. The step size is too coarseand magnification is too small. However, 3-D vi-sualization and modeling would improve ourunderstanding of the morphology of cornealarchitecture, e.g., of epithelial nerve structure.

This was our motivation for developing afast, non-invasive, high-resolution method for

the detailed, 3-D investigation of the humancornea. The further development of the confo-cal microscope [79] took the form of a modifiedconfocal laser-scanning ophthalmoscope [66]based on a commercially available instrument(Heidelberg Retina Tomograph II, HeidelbergEngineering GmbH, Germany) [62]. A water-immersion microscope lens (Achroplan 63¥/0.95W/AA 2.00 mm, Carl Zeiss, Germany) witha long working distance and high numericalaperture was used and coupled to the cornea viaa PMMA cap by interposing a transparent gel(Vidisic, Mann Pharma, Germany) for in vivoimaging [66]. For 3-D imaging, an internal scan-ning device moves the focal plane perpendi-cularly to the x-y plane, in the same way as in optic disk tomography performed with theoriginal HRT II configuration. During imagecapture the z-movement is stopped and theimage plane is exactly perpendicularly to the z-axis. For the investigations presented, anacquisition time of 1 s with a scanning depth of 30 mm was used for all subjects; this is

13.6 Future Developments 201

Fig. 13.49. Epithelium and keratocytes after LASIK

currently thought to be the maximum when pa-tient and examiner movements are taken intoaccount.

Images are presented in the form of a seriesof 2-D grayscale images (384¥384 pixels, 8 bit)representing optical sections through thecornea. The original raw image stacks were con-verted using ImageJ (NIH, USA) for 3-D recon-struction using Amira 3.1 (TGS Inc., USA). Thevoxel size is around 0.8¥0.8¥0.9 mm using theabove-mentioned acquisition parameters. TheAmira volume-rendering software package pro-vides an interactive environment allowing fea-tures such as volume orientation for viewingplanes and 3-D perspectives, segmentation anddetermination of distances and surfaces. Theimage stacks were carefully aligned and modi-fied to eliminate unspecific information byadapting the gray values in the depicted spec-

trum. Shadows and illumination were manipu-lated after assigning density values to grayvalues to more clearly visualize the spatialarrangement without loss of information.

As a first in vivo application of the new de-vice in combination with 3-D reconstructiontechniques, nerve fiber distribution was charac-terized in healthy human corneal epithelium.The spatial arrangement of epithelium, nervesand keratocytes was visualized by in vivo 3-Dconfocal laser-scanning microscopy (CLSM)(Fig. 13.50). The 3-D reconstruction of thecornea in healthy volunteers yielded a picture ofthe nerves in the central part of the humancornea. Thick fibers arise from the subepithelialplexus, and the nerves further subdivide di- and trichotomously, resulting in five to six thinner fibers arranged parallel to Bowman’smembrane and with partial interconnections


Fig. 13.50 A–D. Schematic illustration (A) and 3-Dreconstruction (B–D) of the corneal epithelium withanterior stroma and nerves (healthy human subject):B anterior view, C posterior view, D anterior view with

virtual removal of the epithelium. Thin nerves run-ning parallel to Bowman’s membrane in the basalepithelial plexus. Thicker fibers originating from thesubepithelial plexus

A B

C D

(Fig. 13.51). Branches penetrating the anteriorepithelial cell layer cannot be visualized.

In conclusion, 3-D CLSM is the first tech-nique to permit visualization and analysis of thespatial arrangement of the epithelium, nervesand keratocytes in the living human cornea.Themethod developed provides a basis for furtherdevice refinements and for studies of changes incellular arrangement and epithelial innervationin corneal disease. For example, CLSM may helpto clarify gross variations of nerve fiber pat-terns under various clinical and experimentalconditions.

13.6.2Functional Imaging

In conventional microscopy the possibility ofusing dyes to visualize specific anatomic struc-tures yields major information gains. This is es-pecially true when techniques of fluorescencemicroscopy or immunohistochemistry are usedin combination with confocal techniques [67].Because these methods are well suited for inves-tigating the functional status of tissues, they arealso interesting for in vivo microscopy in hu-mans, for example, for studies of wound healingor inflammation processes. However, problemsarise due to the necessity for “real time” investi-gation because of involuntary movements onthe part of the subjects and due to the selectionof suitable non-toxic vital stains. Nevertheless,successful initial steps have already been takentoward confocal in vivo fluorescence micro-scopy of the anterior eye segments [20, 29], and

these may be regarded as a further enhance-ment of corneal assessments by slit-lamp mi-croscopy following fluorescein or rose bengalstaining [68, 17, 16, 42, 52].

To achieve this, a Heidelberg Retina Angio-graph (HRA/C, Heidelberg Engineering GmbH,Germany) has been modified with a lens attach-


Fig. 13.51. Nerve fibers of the basal epithelial plexus,in strict alignment parallel to Bowman’s membrane

Fig. 13.52 A, B. Fluorescence micrographs of the su-perficial corneal epithelial cells: A intact corneal epi-thelium without appreciable fluorescence; B stippledepithelium after contact glass examination, bluemarked area with fluorescein-stained cells, orangemarked area with unstained intact cells

A

B

ment (Rostock Cornea Module) so that the laserfocus is shifted to the anterior eye segments.This enables fluorescence microscopy images tobe obtained after staining with the non-specificstain sodium fluorescein and excitation withblue argon laser light (wavelength 488 nm) andaddition of a barrier filter (500 nm). Using agreen argon laser (514 nm) in reflection modewith the same device, it is also possible to visu-alize break-up phenomena of the tear film [29,75, 76] (Figs. 13.52–13.54).

The result is a technique that enables fur-ther-reaching investigations of damaged cornealepithelium and of the associated wound healingprocesses. In future, confocal fluorescence mi-croscopes specially designed for in vivo investi-gations in humans will perhaps permit high-quality functional imaging that is even morecomprehensive and specific.


Fig. 13.53 A–D. Patient with intact corneal epithelium: A intact corneal epithelium, slit-lamp microscopyphotograph; B reflection mode, tear film still intact 20 s after eyelid opening; C fluorescence mode, superficialcorneal epithelium, only minimal fluorescence of individual cells; D fluorescence mode, higher magnification

A B

DC


∑ Confocal high-resolution biomicroscopywill be used for the in vivo description ofcorneal pathology at the cellular level

∑ It will enable degeneration and repairmechanisms under various conditions to beexamined so that the findings can be corre-lated with those from conventional slit-lamp biomicroscopy

∑ This will generate enhanced quality in clini-cal evaluation

∑ The use of vital staining substances, e.g.,sodium fluorescein or etidium homodimeror calcein, may give insights into the meta-bolic activities of a variety of cells underdifferent wound healing or degenerativeconditions


Fig. 13.54 A–D. Same patient as in Fig. 13.52 after ap-plication of a local anesthetic and following applana-tion tonometry: A slit-lamp microscopy photographwith corneal stippling; B reflection mode, tear film

defect (dry spot) just 3 s after eyelid opening; C fluo-rescence mode, same area as in B, superficial cornealepithelium, area with marked fluorescence; D fluores-cence mode, higher magnification

A B

DC

Acknowledgements. The authors are gratefulfor the cooperation and detailed contributionsof Alexander Eckard, Steffi Knappe, RobertKraak,Petra Schröder,Oliver Stachs,Hans-PeterVick, and Andrej Zhivov.

References

1. Ardjomand N, Komericki P, Radner H, Aigner R,Reich ME (1997) Korneale Langerhanszellen –Verhalten während der Lagerung in Organkultur.Ophthalmologe 94:703–706

2. Auran JD, Starr MB, Koester CJ, LaBombardi VJ(1994) In vivo scanning slit confocal microscopyof Acanthamoeba keratitis. A case report. Cornea13:183–185

3. Banchereau J,Steinman RM (1998) Dendritic cellsand the control of the immunity. Nature 392:245–252

4. Berlau J, Becker HH, Stave J, Oriwol C, Guthoff RF(2002) Depth and age-dependent distribution ofkeratocytes in healthy human corneas: a studyusing scanning-slit confocal microscopy in vivo.J Cataract Refract Surg 28:611–616

5. Bigar F, Thaer A (1994) Untersuchung der Horn-haut intakter Spenderaugen mit konfokalemSpalt-Scanning Video-Mikroskop. Klin MonatsblAugenheilkd 204:421–423

6. Böhnke M, Thaer AA (1994) Untersuchung derKornea mit einem neuen konfokalen Mikroskop.In: Lund OE, Waubke TN (eds) Bildgebende Ver-fahren in der Augenheilkunde. Methoden und In-dikationen. Enke, Stuttgart, pp 47–53

7. Böhnke M, Masters BR (1999) Confocal micro-scopy of the cornea. Prog Retin Eye Res 18:553–628

8. Bourne WM, Kaufman HE (1976) Specular mi-croscopy of human corneal endothelium in vivo.Am J Ophthalmol 81:319–323

9. Cavanagh HD, Jester JV, Essepian J, Shields W,Lemp MA (1990) Confocal microscopy of theliving eye. CLAO J 16:65–73

10. Cavanagh DH, Petroll WM, Alizadeh H, He YG,McCulley JP, Jester JV (1993) Clinical and diag-nostic use of in vivo confocal microscopy inpatients with corneal disease. Ophthalmology100:1444–1454

11. Coleman CM, Hannush S, Covello SP, Smith FJ,Uitto J, McLean WH (1999) A novel mutation inthe helix termination motif of keratin K12 in a USfamily with Meesmann corneal dystrophy. Am JOphthalmol 128:687–691


13. Duke-Elder S, Wybar KC (eds) (1958) System ofophthalmology, vol II. Henry Kimpton, London,pp 113–127; 543–557

14. Dunn A, Smithpeter C, Welch AJ, Richards-Kor-tum R (1997) Finite-difference time-domain sim-ulation of light scattering from single cells. J Bio-med Opt 2:262–266

15. Fagerholm P (2000) Wound healing after pho-torefractive keratectomy. J Cataract Refract Surg26:432–447

16. Feenstra RP, Tseng SC (1992) Comparison of fluo-rescein and rose bengal staining. Ophthalmology99:605–617

17. Fromm, Groenouw (1891) Ueber die diagnostis-che Verwendbarkeit der Flouresceïnfärbung beiAugenerkrankungen. Arch Augenheilkd 22:247–257

18. Gillette TE, Chandler JW, Greiner JV (1982)Langerhans cells of the ocular surface. Ophthal-mology 89:700–711

19. Gullstrand A (1930) Demonstration der Nern-stspaltlampe. Verslg ophthalm Ges Heidelberg1911, 374, cited in: Vogt A (ed) Lehrbuch und Atlasder Spaltlampenmikroskopie des lebenden Auges,2nd edn. Verlag von Julius Springer, Berlin, p 1

20. Hahnel C,Somodi S,Weiss DG,Guthoff RF (2000)The keratocyte network of human cornea: athree-dimensional study using confocal laserscanning fluorescence microscopy. Cornea19:185–193

21. Hamrah P, Huq SO, Liu Y, Zhang Q, Dana MR(2003) Corneal immunity is mediated by hetero-geneous population of antigen-presenting cells.J Leukoc Biol 74:172–178

22. Hara M, Morishige N, Chikama T, Nishida T(2003) Comparison of confocal biomicroscopyand noncontact specular microscopy for evalua-tion of the corneal endothelium. Cornea 22:512–515

23. Harrison DA, Joos C, Ambrosio R (2003) Mor-phology of corneal basal epithelial cells by in vivoslit-scanning confocal microscopy. Cornea 22:246–248

24. Holopainen JM, Moilanen JA, Tervo TM (2003) Invivo confocal microscopy of Fleck dystrophy andpre-Descemet’s membrane corneal dystrophy.Cornea 22:160–163

25. Ivarsen A, Laurberg T, Moller-Pedersen T (2004)Role of keratocyte loss on corneal wound repairafter LASIK. Invest Ophthalmol Vis Sci 45:3499–3506

26. Kaufmann T, Bodanowitz S, Hesse L, Kroll P(1997) Corneal reinnervation after photorefrac-tive keratectomy and laser in situ keratomileusisand in-vivo study with a confocal video micro-scope. German J Ophthalmol 5:508–512


27. Klintworth GK (2003) The molecular genetics ofthe corneal dystrophies – current status. FrontBiosci 8:d687–713

28. Koester CJ, Roberts CW, Donn A, Hoefle FB (1980)Wide field specular microscopy. Clinical and re-search applications. Ophthalmology 87:849–860

29. Kraak R, Stave J, Guthoff RF (2004) In vivo Unter-suchung des Verteilungsmusters von Na-Fluores-cein an menschlichen Hornhäuten mittels kon-fokaler Laser-Scanning Fluoreszenzmikroskopie.Paper presented at the 102nd Meeting of theDeutsche Ophthalmologische Gesellschaft, Berlin

30. Laibson PR, Dhiri S, Oconer J, Ortolan G (1970)Corneal infiltrates in epidemic keratoconjunc-tivitis. Response to double-blind corticosteroidtherapy. Arch Ophthalmol 84:36–40

31. Laing RA, Sandstrom MM, Leibowitz HM (1975)In vivo photomicrography of the corneal endo-thelium. Arch Ophthalmol 93:143–145

32. Li HF, Petroll WM, Moller-Pedersen T, Maurer JK,Cavanagh HD, Jester JV (1997) Epithelial andcorneal thickness measurements by in vivo con-focal microscopy through focusing (CMTF). CurrEye Res 16:214–221

33. Lund O-E, Stefani FH (1978) Corneal histologyafter epidemic keratoconjunctivitis. Arch Oph-thalmol 96:2085–2088

34. Masters BR, Thaer AA (1974) Real-time scanning-slit confocal microscopy of the in-vivo humancornea. Appl Optics 33:695–701

35. Masters BR, Thaer AA (1994) In vivo humancorneal confocal microscopy of identical fields ofsubepithelial nerve plexus, basal epithelial, andwing cells at different times. Microsc Res Tech29:350–356

36. Masters BR, Böhnke M (2001) Three-dimensionalconfocal microscopy of the human cornea invivo. Ophthalmic Res 33:125–135

37. Mastropasqua L, Nubile M (2002) Confocal mi-croscopy of the cornea. Slack Inc., Thorofare, NJ

38. Mathers WD, Sutphin JE, Folberg R, Meier PA,Wenzel RP, Elgin RG (1996) Outbreak of keratitispresumed to be caused by Acanthamoeba. Am JOphthalmol 121:129–142

39. Mathers WD, Lane JA, Zimmerman MB (1997) As-sessment of the tear film with tandem scanningconfocal microscopy. Cornea 16:162–168

40. Mathers WD, Nelson SE, Lane JL, Wilson ME,Allen RC, Folberg R (2000) Confirmation of con-focal microscopy diagnosis of Acanthamoebakeratitis using polymerase chain reaction analy-sis. Arch Ophthalmol 118:178–183

41. Maurice DM (1968) Cellular membrane activityin the corneal endothelium of the intact eye.Experientia 24:1094–1095

42. Maurice DM (1967) The use of fluorescein in ophthalmological research. Invest Ophthalmol6:464–477

43. Maurice DM (1974) A scanning slit optical micro-scope. Invest Ophthalmol 13:1033–1037

44. Moller-Pedersen T, Ehlers N (1995) A three-di-mensional study of the human corneal keratocytedensity. Curr Eye Res 14:459–464

45. Moller-Pedersen T, Vogel M, Li HF, Petroll WM,Cavanagh HD, Jester JV (1997) Quantification ofstromal thinning, epithelial thickness, andcorneal haze after photorefractive keratectomyusing in vivo confocal microscopy. Ophthalmolo-gy 104:360–368

46. Moller-Pedersen T, Li HF, Petroll WM, CavanaghHD, Jester JV (1998) Confocal microscopic char-acterization of wound repair after photorefrac-tive keratectomy. Invest Ophthalmol Vis Sci39:487–501

47. Müller LJ, Marfurt CF, Kruse F, Tervo TM (2003)Corneal nerves: structure, contents and function.Exp Eye Res 76:521–542

48. Müller LJ, Vrensen GF, Pels L, Cardozo BN,Willekens B (1997) Architecture of human cornealnerves. Invest Ophthalmol Vis Sci 38:985–994

49. Naumann OHG (1997) Pathologie des Auges.Springer, Berlin, Heidelberg, New York, pp 22–30

50. O’Brart DP, Lohmann CP, Fitzke FW, Smith SE,Kerr-Muir MG, Marshall J (1994) Night vision af-ter eximer laser photorefractive keratectomy:haze and halos. Eur J Ophthalmol 4:43–51

51. Oliveira-Soto L, Efron N (2001) Morphology of corneal nerves using confocal microscopy.Cornea 20:374–384

52. Parazza F, Humbert C, Usson Y (1993) Method for3D volumetric analysis of intranuclear fluores-cence distribution in confocal microscopy. Com-put Med Imaging Graph 17:189–200

53. Patel SV, McLaren JW, Camp JJ, Nelson LR,Bourne WM (1999) Automated quantification ofkeratocyte density by using confocal microscopyin vivo. Invest Ophthalmol Vis Sci 40:320–326

54. Petran M, Hadravsky M, Egger MD, Galambos R(1968) Tandem scanning reflected-light micro-scope. J Opt Soc Am 58:661–664

55. Pfister DR, Cameron JD, Krachmer JH, Holland EJ(1996) Confocal microscopy findings of Acan-thamoeba keratitis. Am J Ophthalmol 121:119–128

56. Probst LE, Doane JF (2001) Refractive surgery – acolour synopsis. Thieme, New York

57. Reinhard T, Godehardt E, Pfahl H-G, Sundmach-er R (2000) Lokales Cyclosporin A bei Nummulinach Keratokonjunktivitis epidemica. Ophthal-mologe 97:764–768

References 207

58. Richter A, Slowik C, Somodi S, Vick H-P, GuthoffR (1997) In-vivo-Darstellung der Hornhautinner-vation des Menschen mit Hilfe der konfokalenMikroskopie. Ophthalmologe 94:141–146

59. Rosenberg ME,Tervo TM,Müller LJ,Moilanen JA,Vesaluoma MH (2002) In vivo confocal micro-scopy after herpes keratitis. Cornea 21:265–269

60. Ruben M (1978) Kontaktlinsenanpassung. G. Fis-cher Verlag, Stuttgart, pp 23–45

61. Sangwan VS (2001) Limbal stem cells in healthand disease. Biosci Rep 21:385–405

62. Scheuerle AF, Schmidt E (2004) Atlas of laserscanning ophthalmoscopy. Springer, Berlin Hei-delberg New York

63. Smithpeter C, Dunn A, Drezek R, Collier T,Richards-Kortum R (1998) Near real time confo-cal microscopy of cultured amelanotic cells:sources of signal, contrast agents and limits ofcontrast. J Biomed Opt 3:429–436

64. Stave J, Slowik C, Somodi S, Hahnel C, GrümmerG, Guthoff R (1998) Keratozytendichte der In-vivo-Kornea. Automatische Messung mit einemmodifizierten konfokalen Mikroskop Micro-phthal®. Klin Monatsbl Augenheilkd 213:38–44

65. Stave J, Guthoff R (1998) Darstellung von Tränen-film und In-vivo-Kornea. Erste Untersuchungs-ergebnisse mit einem modifiziertem konfokalenLaser-Scanning-Ophthalmoskop. Ophthalmologe95:104–109

66. Stave J, Zinser G, Grümmer G, Guthoff R (2002)Der modifizierte Heidelberg-Retina-TomographHRT. Ophthalmologe 99:276–280

67. Stelzer EH, Wacker I, De Mey JR (1991) Confocalfluorescence microscopy in modern cell biology.Semin Cell Biol 2:145–152

68. Straub M (1888) Flourescinlösung als diagnostis-ches Hülfsmittel für Hornhauterkrankungen.Centralbl Augenheilkd 12:75–77

69. Sundmacher R, Engelskirchen U (1991) ZumProblem der rezidivierenden und persistieren-den Nummuli nach Keratokonjunktivitis epi-demica. Klin Monatsbl Augenheilkd 198:550–554

70. Suzuki T, Sano Y, Kinoshita S (2000) Conjunctivalinflammation induces Langerhans cell migrationinto the cornea. Curr Eye Res 21:530–533

71. Swensson O, Swensson B, Nolle B, Rochels R,Wannke B, Thiel HJ (2000) Keratin-Gen-Mutatio-nen als Ursache der Epitheldystrophie Meesman-Wilke und autosomal dominanter Verhornungs-störungen der Haut. Klin Monatsbl Augenheilkd217:43–51

72. Tervo T, Moilanen J (2003) In vivo confocal mi-croscopy for evaluation of wound healing follow-ing corneal refractive surgery. Prog Retin Eye Res22:339–358

73. Thaer A, Geyer OC (1997) Die Anwendung derkonfokalen Spaltscanning-Mikroskopie auf dieIn-vivo-Untersuchunge der Kornea-Mikrostruk-tur im Zusammenhang mit dem Tragen von Kon-taktlinsen. Contactologia 19:157–177

74. Tomii S, Kinoshita S (1994) Observations ofhuman corneal epithelium by tandem scanningconfocal microscope. Scanning 16:305–306

75. Torens S, Berger E, Stave J, Guthoff R (2000)Darstellung der Mikroarchitektur und Dynamikder Aufreißphänomene des präkornealen Trä-nenfilms mit Hilfe der Laser-Rastermikroskopie.Ophthalmologe 97:635–639

76. Torens S,Berger E,Stave J,Guthoff R (2000) Light,reflecting and laser scanning microscopy in theimaging and assessment of the precorneal tearfilm. New perspectives on the break-up of the tearfilm. Contactologia 22:49–67

77. Vesaluoma M, Perez-Santonja J, Petroll WM, Lin-na T, Alio J, Tervo T (2000) Corneal stromalchanges induced by myopic LASIK. Invest Oph-thalmol Vis Sci 41:369–376

78. Vogt A (1921) Lehrbuch und Atlas der Spaltlamp-enmikroskopie des lebenden Auges. 2nd edn1930. Springer, Berlin, p 50

79. Webb RH (1996) Confocal optical microscopy.Rep Prog Phys 59:427–471

80. Wiegand W, Thaer AA, Kroll P, Geyer OC, GarciaAL (1995) Optical sectioning of the cornea with anew confocal in vivo slit-scanning videomicro-scope. Ophthalmology 102:568–575

81. Wirbelauer C, Scholz C, Engelhardt R, Laqua H,Pham DT (2001) Biomorphometrie des Horn-hautepithels mittels spaltlampenadaptierter op-tischer Kohärenztomographie OCT. Ophthal-mologe 98:848–852

82. Yoon MK,Warren JF, Holsclaw DS, Gritz DC, Mar-golis TP (2004) A novel arginine substitution mu-tation in 1A domain and a novel 27 bp insertionmutation in 2B domain of keratin 12 gene associ-ated with Meesmann’s corneal dystrophy. Br JOphthalmol 88:752–756

83. Zhivov A, Stave J,Vollmar B, Guthoff RF (2005) Invivo confocal microscopic evaluation of Langer-hans cell density and distribution in the normalhuman corneal epithelium. Graefes Arch Clin ExpOphthalmol (in press)


14.1Introduction

Owing to the fact that the eye is one of the firstorgans to encounter environmental allergens,allergic eye disease has become a common ocu-lar problem, estimated to affect about 20% ofthe population worldwide [51]. Allergic eye dis-ease is one of a spectrum of diseases that sharea common initiating mechanism and pattern of

inflammation and is a problem that is wide-spread among individuals who suffer with aller-gies. Although the incidence of allergic eyedisease varies by geographical location, itsprevalence is difficult to gauge as allergies tendto be underreported.A recent survey conductedby the American College of Allergy,Asthma andImmunology found that 35% of families inter-viewed in the United States experienced aller-gies, 50% of whom reported associated eyesymptoms [48]. However, this prevalence is setto increase probably as a result of environmen-tal factors. For example, the morbidity and mor-tality of asthma have increased with this, coin-ciding with the increase in house dust mitelevels, and are greatest in communities exposedto high allergen levels [32].

Geographical variations, the lack of anyclear-cut objective diagnostic criteria and thedifficulty over the diagnosis – especially when itis the sole manifestation of atopy – have made itdifficult to report the incidence rates for differ-ent forms of allergic eye disease. In the past,clinical features were used to classify allergiceye disease, but recent work that has defined theunderlying pathogenic mechanisms has provid-ed an understanding of the cellular and media-tor mechanisms involved, thereby enabling abetter understanding of the disease process andthe development of more effective treatments.

Allergic conjunctivitis is typically dividedinto five types: seasonal allergic conjunctivitis(SAC), perennial allergic conjunctivitis (PAC),vernal keratoconjunctivitis (VKC), atopic kera-toconjunctivitis (AKC) and giant papillary con-junctivitis (GPC). The latter is an iatrogenic dis-ease associated with foreign bodies on the eye,such as contact lenses and ocular prostheses.

Allergic Eye Disease:Pathophysiology, Clinical Manifestations and Treatment

Bita Manzouri, Thomas Flynn, Santa Jeremy Ono

14

|

∑ Allergic eye disease affects a reported 20% of the population worldwide and may be increasing in line with other atopicdiseases, such as asthma, as a result of environmental factors

∑ Other pathological mechanisms, in addi-tion to the standard type I hypersensitivityreaction, have been recently implicated inthe pathogenesis of allergic eye disease

∑ Established treatments have targeted mastcells, but as a result of our greater under-standing of the mechanisms involved ineye allergy, researchers are now concentrat-ing on other cell types, such as eosinophilsand dendritic cells, as potential targets forimmunomodulation

∑ Other areas of investigation to elucidatenovel treatment strategies include thestudy of the genetics of ocular allergy,the role of environmental factors in thepathogenesis of ocular allergy, and the use of immunostimulatory DNA sequencesthat can inhibit the allergic response

Core Messages

Although not always included in this grouping,it is thought to have a possible allergic mecha-nism because of the predominance of mastcells. GPC invariably resolves when the cause isremoved and keratopathy is rare.

The aim of this review will be to focus on theunderlying mechanisms of allergic eye diseaseand the current classification of the various dis-ease manifestations. Treatment modalities, bothwell established and new innovations, will alsobe discussed.

14.2Pathophysiology

Ocular allergic disease is typically associatedwith immunoglobulin E mediated mast cellactivation (type I immediate hypersensitivityreaction) in the conjunctival tissue. However,recent data from several groups indicate thatother additional mechanisms can also be in-volved in causing a red, allergic eye.

14.2.1Type I Hypersensitivity

The allergic response begins when allergen isencountered by an antigen presenting cell (APC),either directly or as part of an immune complexwith immunoglobulin. The APCs then processand present the allergen to CD4+ T cells as apeptide fragment in association with the majorhistocompatibility (MHC) class II molecule.These T cells are then polarized into T helpertype 1 (Th1) cells and T helper type 2 (Th2) cells.The Th2 cells produce a variety of interleukins,two of which – IL-4 and IL-13 – stimulate im-munoglobulin class switching of B cells fromproducing IgM to producing IgE. This im-munoglobulin binds to high affinity receptors(FceRI) on the surface of mast cells and ba-sophils. Subsequent encounter with this aller-gen results in the cross linkage of IgE bound toFceRI on the surface of mast cells and a cascadeof signal transduction with a resultant release ofpreformed and newly synthesized mediators.Tissue fibroblasts and epithelial cells are alsotriggered by Th2 cells to produce chemokines

such as monocyte chemoattractant protein-1(MCP-1), eotaxin-1, or the protein regulated onactivation normal T-cell expressed and secreted(RANTES), resulting in the migration of inflam-matory cells into the site of allergen exposure[5].

This sensitized mast cell mediated responseis responsible for many of the symptoms seen inSAC and PAC – such as itching, redness and eye-lid swelling – with most of these patients havinga positive family history of atopy and raised lev-els of allergen specific IgE in the serum andtears [32]. Immunohistochemical studies haveshown that in SAC there is a significant increasein the numbers of conjunctival mast cells, whichcorrelates with the patient’s severity of symp-toms [32]. A number of proinflammatory cy-tokines are released by mast cells and these in-clude histamine, leukotriene C4, prostaglandinD2, platelet-activating factor (PAF), tryptase,chymase, cathepsin G and other eosinophil andneutrophil chemoattractants in what is termedthe early phase response [32]. This responselasts for a maximum of 20 min after allergenactivation and includes enhanced tear levels ofhistamine, protease tryptase, and leukotrienes,and an increase in the number of eosinophils[46]. At about 6 h a late phase response occurswhich includes a second peak of tear histamine(without an increase in tryptase) and an in-crease in tissue adhesion molecules E-selectinand interstitial cell adhesion molecule 1 (ICAM-1), which is followed by an influx of inflammato-ry cells such as neutrophils, T cells, basophilsand eosinophils [46]. The presence of tear hista-mine and the absence of tear tryptase in the latephase response may indicate that basophils, asopposed to mast cells, are involved.

Mast cells are also known to synthesize, storeand release a number of cytokines such as IL-4,IL-5, IL-8, IL-13 and TNFa [46]. Cytokine in-volvement, particularly the Th2 cytokines, hasbeen the focus of many studies recently lookinginto the mechanisms of ocular allergy. It isknown, for example, that IL-4 plays a key role inallergic inflammation by promoting T-cellgrowth, by inducing the production of IgE fromB cells, by upregulating the adhesion moleculevascular cell adhesion molecule 1 (VCAM 1),and by regulating the differentiation of the Th2

210 Chapter 14 Allergic Eye Disease: Pathophysiology, Clinical Manifestations and Treatment

subset, which is essential for the allergic reac-tion [19, 31].

Physiologically,mast cells represent a hetero-geneous population. They are subdivided on thebasis of their ultrastructural characteristics,protease content, and T-lymphocyte dependen-cy [49]. In humans, mast cells that containtryptases, chymases, carboxypeptidase A, andcathepsin G are designated MCTC and those thatcontain tryptase only are designated MCT. Al-though both subtypes develop from the sameCD34+ mononuclear precursor, the MCT sub-type is dependent on the presence of T lympho-cytes, present at mucosal surfaces, and increas-es in number in aeroallergen driven allergicdisease, whilst the MCTC subtype appears to beindependent of T cells but its developmentrequires fibroblastic derived growth factors,which are predominant in connective andperivascular tissues, and is characteristic offibrotic processes [32]. Normally, approximately80% of conjunctival mast cells are of the MCTCphenotype and are mainly subepithelial in dis-tribution, with the rest being MCT, but duringallergic inflammation such as that seen in SAC,VKC or AKC, the numbers of the latter type in-crease in the epithelial and subepithelial layers[37]. In the chronic and fibrosing conditionAKC, however, the MCTC subtype predominates,perhaps indicating an important transitionfrom a simple mediator driven disorder to thatof chronic inflammation leading to conjunctivalfibrosis [37].

14.2.2Ocular Inflammatory Reaction:Late Phase

A late phase reaction sustained by a complexnetwork of inflammatory cells and mediatorscan also occur in the eye. This has been demon-strated in humans using allergen for conjuncti-val provocation of allergic subjects [10]. Aller-gen challenge caused the typical early-phasereaction within 20 min, with the initial reactionbeing dose dependent. With smaller doses ofallergen the reaction was not so pronouncedand spontaneous recovery occurred within abrief period.With larger doses, the reaction was

more persistent and progressed to a late-phasereaction. Typically, high doses of allergen in-duced a continuous reaction manifested byburning, redness, itching, tearing and a foreignbody sensation that began 4–8 h after challengeand persisted for up to 24 h. This clinical reac-tion was accompanied by a significant recruit-ment of inflammatory cells in tears. Neutrophilsfirst appeared about 20 min after challenge,with eosinophils and lymphocytes increasing inprominence 6–24 h after challenge.

The eosinophil predominates in the latephase reaction. It is a powerful effector cell,releasing arginine rich toxic proteins capable ofcausing corneal epithelial damage [32]. Normal-ly, eosinophils are not found in the conjunctivalepithelium of non-atopic subjects but the num-bers are increased in the conjunctival epitheli-um, subepithelium and tears of patients withAKC and, to a greater extent, VKC patients.Furthermore, this increase in eosinophils andeosinophil products [e.g. eosinophil peroxi-dase, eosinophil cationic protein (ECP)] is alsopresent in both skin test positive and skin testnegative VKC and is not confined to ocular tis-sues.This suggests that, in at least some forms ofallergic conjunctivitis such as VKC, eosinophilicinfiltration – and not IgE sensitization – is themore relevant feature of the disease and is asso-ciated with signs of systemic activation ofeosinophils [10].

14.2.3Non-specific Conjunctival Hyperreactivity

Non-specific stimuli can also cause target organhyperreactivity and this is thought to play a rolein allergic diseases of the eye. It is postulatedthat “non-specific conjunctival hyperreactivity”may represent a distinct pathophysiological ab-normality in allergic eye disease [10]. The vari-ability of symptoms experienced in allergicconjunctivitis which do not correlate with envi-ronmental changes such as the levels of sensitiz-ing allergens, as well as the ocular reaction in-duced by non-sensitizing stimuli, may well beexplained by this non-specific hyperreactivity.Natural non-specific stimulation with agentssuch as wind, dust, and sunlight may act only as

14.2 Pathophysiology 211

triggers of an abnormal non-specific reactivityof the conjunctiva in allergic patients [10].

Furthermore, multiple physical, chemical, in-fectious, or antigenic factors may stimulate thebiological responses of mast cells, leading to therelease of several mediators. Rubbing of theeyes, exposure to UV light, and increase of ocu-lar surface temperature may lead to acute de-granulation of the mast cells and release of theirmediators. The local generation of stimuli thatinduce different patterns of mast cell cytokinerelease may represent another method of bio-logical, non-specific activation of mast cells[42]. It has been observed that whenever a pa-tient with VKC is exposed to the sun, signs andsymptoms recur. Furthermore, the symptoms ofallergy become most severe in children withVKC who develop bacterial conjunctivitis. Cer-tain types of lipopolysaccharides of bacteriamay cause degranulation of mast cells, leadingto the release of their mediators that cause exac-erbation of the allergic process.

14.2.4T-Cell-Mediated Hypersensitivity in Allergic Eye Disease

Both CD4+ and CD8+ T cells populate thesubepithelial tissue of the normal human con-junctiva. In the active forms of SAC and PAC, theT cell profile remains virtually unchanged com-pared to the normal milieu, but in chronic aller-gic disorders such as VKC,AKC and GPC, CD4+T cells but not CD8+ T cell numbers are in-creased, with a mixed cellular infiltrate contain-ing many mast cells, eosinophils, neutrophils,and macrophages [32]. In chronic allergic dis-eases there is no clear-cut difference betweenthe allergen specific IgE responses and thenature and severity of the allergic responses;hence it is likely that non-IgE mechanisms arecontributory, with the involvement of cell medi-ated responses [32].

Most of the T cells in normal conjunctiva arenaïve, but in chronic allergic conditions 90% ofthe T cells are memory T cells [35]. Correspond-ing with this rise in activated T cells, there isalso upregulation of markers present on antigenpresenting cells.

CD4+ T cells can be further subdivided intotwo distinct subsets based on their pattern ofcytokine production. The first subset, Th1 cells,produce IL-2, IL-3, TNFb and interferon g(IFNg) and are more associated with classic de-layed type hypersensitivity. The second subset,Th2 cells, produce a range of cytokines encodedon chromosome 5, such as IL-4 and IL-5, whichpromote immediate hypersensitivity responsesthrough their ability to stimulate proliferation,B cell IgE production and eosinophil produc-tion, activation and survival [32]. It has beenshown that in AKC there is increased numbersof both Th1 and Th2 lymphocytes as opposed toin VKC where lymphocytes secreting cytokinestypical of the Th2 subset are found. This ob-servation suggests that VKC results from a mat-uration shift of CD4+ T cells towards a patternof secretion of cytokines which drives a mastcell and eosinophil mediated inflammatoryresponse [34].

14.3Clinical Syndromes of Allergic Eye Disease

Allergic diseases of the eye comprise a numberof different inflammatory conditions that sharecommon features such as seasonal variation,association with atopic disease and presumed involvement, to a greater or lesser extent, of thetype I hypersensitivity mechanism in theirpathophysiology. They are traditionally classi-fied, as outlined above, into five distinct entities:SAC, PAC, VKC, AKC, and GPC (Fig. 14.1).

As previously mentioned, recent evidencesuggests that the traditional type I hypersensi-tivity reaction may be less important in some ofthese diseases than others. However, these dis-eases share many symptoms in common and itis therefore reasonable to group them in thesame broad category of “allergic eye disease”.The cardinal feature of all allergic eye disease isitching – in the absence of this symptom oneshould be wary of making this diagnosis. Othersymptoms such as tearing, burning and foreignbody sensation may be present in variable de-grees in all of these conditions. Despite similar-ities in the symptoms, it is important to distin-guish, where possible, between the different


types of allergic eye disease as each of them hasa different visual prognosis. Accurate diagnosiswill allow appropriate counselling of patients.

The most common type of allergic eye dis-ease, seasonal allergic conjunctivitis (hay feverconjunctivitis), is also the least serious in termsof visual outcome. SAC and PAC together ac-count for 98% of allergic eye disease [41]. VKCand AKC, although much rarer, are more likely

to lead to visual impairment, with AKC beingthe most destructive disease and having theworst visual prognosis. The emergence of new-er treatments based on an increasing under-standing of the individual pathogenic mecha-nisms of each disease also underlines theimportance of accurate diagnosis. The differenttypes of allergic eye disease can usually be dis-tinguished by history and examination alone.

14.3 Clinical Syndromes of Allergic Eye Disease 213

Fig. 14.1. A Normal bulbar conjunctiva; B giant papillae in GPC; C typical appearance of superior tarsal con-junctiva in a severe case of SAC; D corneal ulcer in VKC; E early stages of corneal pannus in AKC; F Horner-Trantas dots seen in AKC. (Pictures courtesy of Dr. Mohammed Siddique, Institute of Ophthalmology, London)

A B

D

F

C

E

14.3.1Seasonal Allergic Conjunctivitis

Of the allergic eye diseases, SAC represents themost “pure” form of type I hypersensitivity. Asthe name suggests, the symptoms and signs areintermittent and occur rapidly following expo-sure to a specific allergen, with patients oftenhaving a personal or family history of atopy. Inthe absence of prolonged exposure to allergen,attacks are short lived. The commonest season-al allergen is pollen, with tree pollen predomi-nating in spring, grass pollen in summer andragweed pollen in autumn. Symptoms are typi-cally absent during winter. The severity of signsand symptoms varies from patient to patientdepending on the specific allergen and the ex-posure.

14.3.1.1Symptoms

Patients usually complain of intense itching ofthe eyes associated with a watery discharge.

14.3.1.2Signs

There may be eyelid oedema. Conjunctival ves-sels may be injected and conjunctival chemosismay give the conjunctiva a “milky” appearance.Symptoms and signs are usually bilateral al-though they may be asymmetrical. Young chil-dren can present with dramatic unilateral lidoedema and chemosis.

14.3.2Perennial Allergic Conjunctivitis

PAC is less common than SAC. Although thesymptoms and signs of these diseases are thesame, the distinction between them lies in thetiming of the symptoms.Whereas SAC sufferershave symptoms for a defined period of time,PAC sufferers are sensitive to allergens that arepresent year-round and so are perennially

symptomatic.“Household”allergens such as thedust mite or pet dander are the usual offendersin PAC. These patients may also be sensitive toseasonal allergens and so there may be a super-imposed seasonal element to their symptoms.

14.3.3Vernal Keratoconjunctivitis

A disease of childhood, VKC accounts for 0.5%of allergic eye disease [32]. Like AKC it has amale preponderance but onset is much earlier,typically late in the first decade. It is seen mostcommonly in temperate climates such as thoseof the Mediterranean, South Africa and NorthAmerica. However, genetic as well as environ-mental factors are important. Even in coolernorthern climates the disease is more common-ly seen in people of African or Asian descent[39]. There is frequently a personal or familyhistory of atopy but this association is not asstrong as in other types of allergic eye disease,with a large proportion of VKC patients havingno such history.

In the majority of cases the disease showsseasonal variation with symptoms typically ap-pearing in spring and lasting about 6 months.Additional recurrences in winter are common.In some cases the disease evolves over time intoa more chronic, perennial form of inflammationwith up to one-quarter of VKC patients having aperennial form of the disease from the outset[11]. Although serious visual complications mayoccur, VKC is a less destructive disease thanAKC and usually burns itself out by the earlytwenties [30].

14.3.3.1Symptoms

Symptoms are usually bilateral but may beasymmetrical and, like all allergic eye diseases,itching is a cardinal feature. Photophobia is alsoprominent and patients may complain of tear-ing and a mucoid discharge. Depending on theseverity of corneal involvement, they may alsocomplain of a foreign body sensation or pain.


14.3.3.2Signs

In contrast to AKC, the periorbital skin is usual-ly unaffected. The disease is further classifiedinto tarsal, limbal or mixed VKC depending onthe location of the conjunctival inflammatorysigns.

Tarsal. The inflammation is predominantly inthe superior tarsal conjunctiva although thebulbar conjunctiva may show non-specific signssuch as injection or chemosis. The superiortarsal conjunctiva develops a papillary reaction.Papillae are typically large (>1 mm) and diffuse,giving a “cobblestone” appearance. These tarsalpapillae tend to persist even when the disease isquiescent but become hyperaemic and oedema-tous during periods of disease activity. Thepresence of a thick, mucoid, white secretion as-sociated with these papillae is another indicatorof disease activity. Papillae may enlarge to sev-eral millimetres in diameter and may give riseto ptosis. In severe forms of the disease, linearsubepithelial scars (Arlt’s lines) may appearparallel to the lid margin.

Limbal. Limbal VKC is characterized by singleor multiple gelatinous, pale infiltrates in thelimbal conjunctiva. The extent of limbal in-volvement is variable. Infrequently, there maybe 360° limbal inflammation. There is usuallyinjection of the surrounding bulbar conjuncti-val vessels. Aggregates of degenerating eosino-phils at the apex of the infiltrates are seen assmall white spots (Horner-Trantas dots) – boththe limbal infiltrate and the Horner-Trantasdots are transient.

In mixed VKC both limbal and tarsal signsmay be observed. Although limbal and tarsalVKC are believed to be variants of the same dis-ease, certain differences have been observed intheir demographics and natural history. LimbalVKC is particularly common in people ofAfrican or Asian descent. There is mixed evi-dence as to which, if either, of the variants ismore responsive to treatment [11, 52]. Patientswith tarsal disease are certainly more likely todevelop sight-threatening corneal ulceration[52].

Cornea. Sight-threatening complications occurless frequently in the cornea than in AKC. How-ever, both non-specific and pathognomoniccorneal signs are seen. In a follow-up series of195 patients with VKC, 9.7% developed cornealulcers and 6% developed a permanent decreasein visual acuity [11]. Abnormalities of the cen-tral and superior cornea are most commonlyseen in tarsal disease. In its earliest form theremay be only punctuate epithelial erosions.These may, with time, coalesce to form largererosions that may in turn evolve into the charac-teristic “shield” ulcer of VKC. Shield ulcers arenon-infectious and occur in the central/superi-or cornea. At first they are shallow with a trans-parent base. Over time the ulcer becomes filledwith inflammatory debris and the base opaci-fies. Further accumulation of inflammatory de-bris leads to plaque formation. The pathogene-sis of these ulcers is incompletely understood.Mechanical abrasion of the epithelium by largepapillae on the superior tarsal conjunctiva isthought to play a role, as is epithelial corrosionby toxic granule proteins released from eosino-phils in the tarsal conjunctiva and tear film. Inpersistent or recurrent limbal disease, peripher-al corneal signs such as pannus or opacification(pseudogerontoxon) may develop. Limbal le-sions may also cause significant astigmatism.

14.3.4Atopic Keratoconjunctivitis

First described in 1952 [22], AKC constitutes amore relentless form of conjunctival inflamma-tion than either SAC or VKC. Atopic dermatitis(eczema), a pruritic skin condition that affects3% of the population, is present in 95% of pa-tients with AKC [7]. Conversely, 25–40% ofatopic dermatitis patients have AKC [18]. Typi-cally patients have had atopic dermatitis sincechildhood with ocular symptoms developing ata later stage. Symptoms may begin in the lateteens or early twenties but the peak incidence isbetween the ages of 30 and 50. Males are morecommonly affected than females and there is of-ten a personal or family history of other atopicdiseases.Unlike SAC,and most cases of VKC,thesymptoms are perennial. It differs from PAC in

14.3 Clinical Syndromes of Allergic Eye Disease 215

that the symptoms are less intermittent. Al-though there may be periods of relative quies-cence, signs of disease activity are usually pres-ent to some degree.

14.3.4.1Symptoms

Bilateral itching of the eyelids and periorbitalskin is the most frequent symptom. Patientsalso complain of tearing, photophobia, burningand blurred vision. Increased mucus and in-flammatory debris may thicken the tear filmand contribute to a stringy discharge. Depend-ing on the severity of corneal involvement, pa-tients may complain of a foreign body sensationand pain.

14.3.4.2Signs

Invariably there are signs of disease on the eye-lids and periorbital skin. Ocular surface inflam-mation in AKC may, as the name suggests, affectthe conjunctiva and cornea. In many cases thedisease is mild and corneal signs may actuallybe absent or minimal. Such cases have beentermed atopic blepharoconjunctivitis (ABC) [53].

Eyelids. The periorbital skin typically has thedry, indurated and scaly appearance of eczema.Eyelid swelling may contribute to the general-ized wrinkling of the skin and the developmentof a fold in the lower lid skin (Dennie-Morganfold). In severe cases there may be fissures at thelateral canthus and/or absence of the lateralpart of the eyebrow (Herthoge’s sign). The lattersigns may be induced or aggravated by vigorouseyelid rubbing. Lid margins may be thickened(tylosis) and may develop meibomian glanddysfunction. Colonization of the lid marginwith staphylococcus with resultant staphylo-coccal blepharitis is common [54].

Conjunctiva. There is typically a papillary re-action on the tarsal conjunctiva, which, in con-trast to VKC, is usually more prominent on theinferior, rather than the superior, tarsal con-junctiva. The bulbar conjunctiva may shownon-specific signs of inflammation such as

hyperaemia or chemosis. Rarely, papillaryhyperplasia of the limbal conjunctiva occurs,resulting in a gelatinous limbal nodule similarto those seen in limbal VKC.Associated Horner-Trantas dots have been seen. Prolonged orsevere inflammation may result in conjunctivalcicatrization.This is most commonly seen in thelower fornix and may result in shallowing of thefornix and symblepharon. Activation of fibro-blasts by mast cells has been proposed as amechanism for conjunctival scarring in allergicdisease [47]. Several cases of squamous cell car-cinoma/CIN have been reported in patientswith atopic dermatitis or AKC [20, 24] althoughthe mechanism of tumourigenesis remainsunclear.

Cornea. Visual deterioration in AKC is mostcommonly caused by corneal complications.Corneal scarring in AKC may result from vascu-larization, infection or keratoconus. A broadspectrum of corneal disease may be seen de-pending on the severity and chronicity ofinflammation. Punctate epithelial erosions areseen early in the course of the disease. Theseverity of the corneal erosions correlates withthe number of inflammatory cells (especiallyeosinophils) in brush cytology samples fromthe superior tarsal conjunctiva [50]. Peripheralcorneal vascularization, which may be associat-ed with opacification, is common. Thesechanges may occur as a result of limbal stem celldeficiency. Rarely, corneal vascularization mayencroach on the visual axis and cause visual im-pairment. Epithelial erosion may coalesce toform non-infectious corneal ulcers. Toxic gran-ule proteins derived from conjunctival eosino-phils have been implicated in the pathogenesisof these ulcers [33]. Staphylococcal colonizationof the lid margins coupled with a decrease inbarrier function [56] also puts AKC patients atincreased risk of developing bacterial infectiouscorneal ulcers. They are particularly vulnerableto herpes simplex keratitis [16]. Chronic eyerubbing may be an important factor in the asso-ciation between AKC and keratoconus [6].

Other Causes of Visual Deterioration in AKC.AKC is associated with the development of pre-mature bilateral cataracts. Typically the lens


opacity develops in the anterior subcapsularregion and has well defined margins. It is oftenreferred to as a “shield” cataract. A rarer cause of visual impairment in AKC is that of retinaldetachment [57]. The reasons for this associa-tion are not well understood. Finally, chronicuse of topical steroids in the treatment of AKCmay result in posterior subcapsular cataractsand glaucoma (see below).

14.3.5Giant Papillary Conjunctivitis

The term giant papillary conjunctivitis de-scribes the advanced stages of the conjunctivalresponse to the prolonged presence of a foreignbody on the ocular surface. It was first observedand characterized in contact lens wearers [3]and was later reported in patients with ocularprostheses and exposed suture ends. Nowadays,it is seen commonly in contact lens wearers, andmost of the knowledge of this condition arisesfrom experience with these patients. Wearers ofsoft contact lenses are most likely to developgiant papillary conjunctivitis, but it has beenestimated that 1–5% of rigid gas-permeable lenswearers may also be affected [26, 27]. The condi-tion shows no age or gender preference andthere does not appear to be a strong associationwith allergy [28].

14.3.5.1Symptoms

Earliest symptoms are of mucus discharge inthe morning and itching on removal of the lens-es. As the disease progresses these symptomsbecome more marked and may be associatedwith a foreign body sensation. Patients com-plain of blurred vision as a result of coating ofthe lens with mucus and increasing lens mobil-ity and instability. As the disease advances pa-tients become increasingly intolerant of theircontact lenses.

14.3.5.2Signs

Giant papillary conjunctivitis is characterized,in the late stages, by the presence of abnormallylarge (>0.3 mm) papillae on the superior tarsalconjunctiva. In the earliest stage, however, whenthe patient first becomes symptomatic, the con-junctiva may appear normal.As the disease pro-gresses the superior tarsal conjunctiva becomesthickened and hyperaemic. Small papillae de-velop first which increase in size and numberover time. The distribution of giant papillaevaries according to the type of lens worn. Inwearers of soft lenses papillae emerge first atthe superior edge of the tarsal plate. Wearers ofhard lenses, which are smaller, develop papillaecloser to the superior lid margin [23]. The bul-bar conjunctiva and inferior fornix are usuallynormal.

The symptoms and signs of this disease mayresemble those of VKC. Important factors in thehistory, which could help to distinguish theseconditions, include contact lens history andpatient age since VKC is seldom seen after theearly twenties.

14.4Treatment of Allergic Eye Disease

The mainstays of treatment for the majority ofallergic eye disease symptoms are topical eyedrops, and for this purpose a wide range of top-ically administered agents have been developedto treat the milder disease varieties. Theseinclude antihistamines, mast cell stabilizingagents and anti-inflammatory agents.Addition-ally, topical nasal decongestants are also avail-able. Of the topical eye drops, it is antihista-mines and mast cell stabilizers that have beenextensively studied to assess their therapeuticvalue in a large number of comparative clinicaltrials over the years. Furthermore, as the chem-ical and cellular infiltrates in both acute andchronic allergic eye disease become better char-acterized, there are significant implications fortreatment of these conditions. Efficacy of all ofthese agents varies from patient to patient andthe choice of agent used depends on a number

14.4 Treatment of Allergic Eye Disease 217

of variables, such as the underlying state ofhealth of the eye being treated, drug costs andavailability, contact lens wear, and the potentialfor compliance [8].

The preferred treatment modality in milddiseases such as SAC and PAC is topical therapy,since neither is sight threatening, and theirpathogenesis involves mast cell degranulationand the release of histamine. Topical treatmentoffers several advantages: the ease of applica-tion directly to the site affected by the diseaseprocess, the general lack of systemic side effects,and the washout effect of the drops themselvesaiding the removal of the inflammatory media-tors.

14.4.1Antihistamines

The first line of treatment of ocular allergy in-cludes the avoidance of allergens, the use of coldcompresses for symptom relief (especially itch-ing), and regular lubrication of the eye to washout tear histamine and other inflammatory me-diators, thus diluting their effects and aiding thepatient’s comfort. Topical therapy may startwith the use of antihistamines or mast cell sta-bilizers.Considering the former, the stimulationof H1 receptors in the conjunctiva mediates the symptom of itching whereas H2 receptoractivation results in vasodilation. Second gener-ation H1 receptor antagonists are used for thetopical treatment of the benign forms of allergicconjunctivitis, and these include levocabastine,azelastine and emedastine. They all bind selec-tively to H1 receptors in the conjunctiva andhave little or no effect on dopaminergic, adren-ergic or sertotoninergic receptors [46]. Of thisnew generation H1 receptor antagonists, topicalazelastine has been shown to be a powerfultopical antihistamine, decreasing eosinophiland T lymphocyte activation, having an in-hibitory effect on a broad array of other media-tors, and being a potent suppressor of itchingand conjunctival hyperaemia after conjunctivalprovocation with an allergen, with an onset of action seen within 3 min and a duration ofeffect of at least 8–10 h [32, 46].Although topicalantihistamines can be used alone to treat aller-

gic conjunctivitis, combining an antihistaminewith a vasoconstrictor is more effective thaneither agent alone. The vasoconstrictors com-monly used in combination with topical anti-histamines are phenylephrine or naphazoline[8].

14.4.2Mast Cell Stabilizing Agents

The most common topical drugs invariablyused by ophthalmologists for all forms of aller-gic conjunctivitis are the mast cell stabilizingagents. These include sodium cromoglygate,lodoxamide, ketotifen, nedocromil sodium andthe newly introduced olopatadine. Mast cell sta-bilizers are effective in the milder forms of aller-gic eye disease and have very few side effects,either locally or systemically, but for patients toreceive long-term benefit from them such thatexpected exposure to allergen reduces thetryptase and inflammatory cells after allergenchallenge, treatment is needed for many years[46].

Sodium cromoglygate is the prototypic mastcell secretion inhibitor. It is the oldest and mostwidely used agent of this family of drugs. How-ever, despite its extensive use, the mechanismsof its action are still unclear. The efficacy of themedication appears to be dependent on the con-centration of the solution used [9]. Nedocromilsodium has been shown to be able to inhibitchloride ion flux in mast cells, epithelial cellsand neurons. This feature may explain how itcan prevent responses such as mast cell degran-ulation. Others have suggested the inhibition of IgE production by B cells as an alternativemechanism [46]. Newer agents such as lodox-amide have become available, which are fasteracting and approximately 2,500 times morepotent than sodium cromoglycate in the pre-vention of histamine release, that also act to re-duce tear tryptase and inflammatory cells afterallergen challenge [8]. In a comparative trialwith sodium cromoglygate and lodoxamide insubjects with the more severe forms of allergiceye disease (VKC, AKC and GPC), lodaxamidewas found to be superior for symptom relief. Itwas also found to be effective in the long-term


treatment of VKC especially in cases with anepitheliopathy [17, 43].

14.4.3Dual-Acting Agents

Dual-acting agents are named for their antihis-tamine effects and their inhibition of mediatorrelease. They are the newest generation of an-tiallergic agents. The advantages of these drugslie in the rapidity of symptomatic relief given byimmediate histamine receptor antagonism cou-pled with the long-term disease modifying ben-efit of mast cell stabilization. Not all of theseagents are equivalent and in selecting a dual-ac-tion agent, one should look for a potent andlong-lasting agent that relieves the signs andsymptoms of allergy, including itching, redness,lid swelling and chemosis [41].

Clinical studies have demonstrated the effi-cacy and tolerance of olopatadine for the man-agement of allergic conjunctivitis or in a con-junctival allergen model [1, 4, 13]. This agentboth acts as a mast cell stabilizer and has anti-histamine activity. This dual mode of action hasbeen shown to be advantageous for the manage-ment of allergic conjunctivitis, and as a topicalpreparation has been subjectively preferred bypatients [4, 44]. Furthermore, a direct anti-inflammatory property for this drug has beensuggested by a study which showed thatolopatadine inhibited the anti-IgE antibody-mediated release of TNFa from human con-junctival mast cells [14].

14.4.4Non-steroidal Anti-inflammatory Drugs(NSAIDs)

Prostaglandins, especially PGE2 and PGI2, lowerthe threshold of the human skin and conjuncti-va to histamine-induced itching. NSAIDs, by in-hibiting the production of prostaglandins, helpto alleviate this itching but also reduce pain andinflammation of the eye associated with allergicreactions [9, 46]. NSAIDs used in the topicaltreatment of allergic ocular conditions in-clude ketorolac, diclofenac, fluribrofen and

indomethacin. These agents, unlike corticos-teroids, do not mask ocular infections, affectwound healing, increase intraocular pressure,or contribute to cataract formation [8]. How-ever, of these agents, only ketorolac trometh-amine (Acular) has been approved by the Foodand Drug Administration for the managementof acute SAC [15]. It acts to significantly reducetear tryptase levels and the number of eosino-phils and lymphocytes in tear specimens afterconjunctival provocation [29].

Ocular NSAIDs have been associated with alow-to-moderate incidence of burning andstinging [9]. The concern of NSAID-inducedasthma does not appear to be a problem exceptin patients who have the triad of asthma, nasalpolyposis and aspirin sensitivity [45].

14.4.5Topical Corticosteroids

Topical steroid preparations are the most effec-tive therapy for moderate to severe forms ofVKC, but their use should be strictly limited forsevere cases and carefully monitored since theirlong-term use is associated with an increasedrisk for the development of cataracts and glau-coma and can potentiate ocular herpetic infec-tions. In fact, topical steroids are responsible forthe 2% incidence of glaucoma in VKC patients[12]. In T cell dependent AKC and VKC, sodiumcromoglycate has been used either prophylacti-cally or as maintenance therapy to control mildsymptoms only, but is ineffective in acute exac-erbations. In acute exacerbations, even the new-er class of mast cell stabilizers may not beenough, and under these circumstances steroids(fluoromethalone or dexamethasone) tend to beused in doses of up to one drop hourly toreverse corneal epitheliopathy caused by therelease of epithelial toxic mediators fromeosinophils and neutrophils [32]. Once controlof the acute phase of the disease has beenachieved, steroids should be discontinued andalternative topical treatment, as outlined previ-ously, should be started [12].

Two modified corticosteroids have recentlybeen investigated for their efficacy in allergicconjunctivitis: rimexoline (a derivative of pred-

14.4 Treatment of Allergic Eye Disease 219

nisolone) that is quickly inactivated in the ante-rior chamber of the eye, thus improving efficacyand decreasing the safety concerns, e.g. raisedintraocular pressure; and both low-dose andhigh-dose loteprednol etabonate are highly ef-fective as prophylaxis against, and in the acutephase of, allergic conjunctivitis [8].

14.4.6Calcineurin Inhibitors

Two calcineurin inhibitors are currently in clin-ical use:1. Cyclosporin A (CsA) is a fungal antimetabo-

lite and anti-CD4+ agent that decreases theclinical signs and symptoms of the chronicforms of VKC and AKC. It acts to controlocular inflammation by blocking Th2 lym-phocyte proliferation and IL-2 production,by inhibiting histamine release from mastcells and basophils, and by reducing the pro-duction of IL-5, thereby reducing the recruit-ment and effects of eosinophils on the con-junctiva [12]. Although systemic CsA hasbeen used for the treatment of severe AKCand keratoconjunctivitis sicca, topical cyclo-sporin causes ocular irritation with burning,tearing, erythema and itching. This is due tothe fact that since the drug is lipophilic, it hasto be dissolved in an alcohol base whichcauses the ocular irritation [8]. However, thetopical form of this drug is not yet generallyavailable.CsA has been evaluated in patients withsteroid dependent AKC. In one study, 12 pa-tients were randomized to treatment withCsA and 9 patients to a vehicle treatmentgroup. The results showed that in the CsAgroup, 9 out of 12 patients were able to ceasesteroid therapy as compared to 1 out of 9 inthe vehicle group [21]. Furthermore, the finalsteroid use was significantly lower in the CsAgroup versus the vehicle group. This studyconcluded that CsA is an effective and safesteroid sparing agent in AKC and is also ca-pable of improving the symptoms and signsof AKC. In another randomized trial theshort-term efficacy and safety of topical CsA0.05% was evaluated in the treatment of pa-

tients with severe, steroid resistant AKC [2].Patients were randomly assigned to treat-ment with topical CsA 0.05% or placebo fora period of 28 days with the symptoms andsigns of AKC recorded on the day of enroll-ment and at the end of the treatment period.The results, recorded by a composite scorecomputed by summing the severity grade ofall five symptoms and six signs of AKC,showed a greater improvement in the CsAgroup relative to the placebo group at the endof the treatment period. It was hence con-cluded that topical CsA 0.05% is safe, andmay actually have some effect in alleviatingthe signs and symptoms, in severe AKC thatis resistant to topical steroid treatment.

2. Tacrolimus (FK-506) is a macrolide antibiot-ic with potent immunomodulatory proper-ties which has already been used to treat theimmune mediated problems encounteredwith corneal graft rejection, ocular pem-phigoid and uveitis. It acts on T lymphocytesto block the production of lymphokines,such as IL-2, IL-2, IL-5, TNFa and interferon-g a. It also blocks the degranulation of mastcells and several mast cell cytokines, such asIL-3 and IL-5 [8].

14.4.7Future Drug Developments

The aims of future drug development will focuson steroid-sparing agents that control theimmune response. These may be administeredalone, or in combination with newer drugs thathave already demonstrated their efficacy in themanagement of these conditions, such as anti-histamines and mast cell stabilizers.

Our understanding of the pathophysiologyof allergic conjunctivitis has increased greatlyover the last 3 years. New areas of investigationto elucidate novel treatment strategies includethe study of the genetics of ocular allergy, sinceit has been known for some time that differentmouse strains are more or less responsive tospecific allergen challenge in the eye, and link-age analysis of these mice is being pursued todefine disease susceptibility genes for ocularallergy [41]. A few studies have addressed the


role of environmental factors in the pathogene-sis of ocular allergy. For example, it has beenshown that there is a positive association be-tween the dietary intake of n-6 polyunsaturatedfatty acids and seasonal allergic rhinoconjunc-tivitis [55]. Other studies have focused on thegenetics of allergic conjunctivitis. One of theearliest published studied approximately 117families with probands with allergic conjunc-tivitis [40]. Evidence was found, by analysis ofthe genomic DNA, for genetic linkage of allergicconjunctivitis for chromosomes 5, 16 and 17.This genetic linkage for allergic conjunctivitiswas shown to differ from that reported foratopic asthma, and hence it was concluded thatthere were likely to be organ specific diseasesusceptibility genes,which, together with gener-al atopy genes, target the allergic response tospecific mucosal tissues.

Resident dendritic cells in the conjunctivahave also been the focus of recent research sinceit has been shown that dendritic cell activationby an allergen is a very early step in diseasepathogenesis, with dermal allergy being used asthe prototype [41]. Other areas of interest lie inthe activation and mediator release from hu-man conjunctival mast cells on FceRI cross-linking. A recombinant humanized monoclonalanti-IgE antibody, omalizumab, was recentlydeveloped which binds specifically to the IgEbinding site on human FceRI and therebyblocks the binding of IgE to mast cells and baso-phils [5]. Studies have shown that this agentbenefits patients with moderate to severe aller-gic asthma who remain symptomatic despitetreatment with systemic or inhaled corticos-teroids [5]. Additionally, omalizumab has beenshown to be safe and well tolerated.

One of the most innovative treatment ad-vances has been in the use of immunostimula-tory DNA sequences that can inhibit the allergicresponse. Both bacterial DNA and syntheticoligodeoxynucleotides containing specific mo-tifs centered on a CpG dinucleotide have beenshown to be potent immunostimulatory agents[46]. It is likely that these sequences represent asignal to the immune system, resulting in apowerful Th1 response and this can be used toswitch an allergic response from a Th2 domi-nated immune profile towards a Th1 profile

[46]. Miyazaki et al. evaluated the therapeuticpotential of immunostimulatory sequenceoligodeoxynucleotide (ISS-ODN) administra-tion in ocular allergy using a mouse model ofragweed-specific conjunctivitis [36]. They con-cluded that ISS-ODN was an effective treatmentfor ocular allergy when administered systemi-cally or conjunctivally. Systemic treatmentmarkedly inhibited clinical parameters of SACand blocked conjunctival eosinophilia in thelate phase reaction. Additionally, it also effec-tively blocked neutrophilia, which is a hallmarkof the late phase reaction.

Other areas of potential therapeutic valuewhich require further research include the useof antagonists of the action of macrophage in-flammatory protein-1a (MIP-1a) and the use ofIL-1 receptor antagonists. Data have shown thatMIP-1a constitutes an important second signalfor mast cell degranulation in the conjunctiva invivo and consequently for acute phase disease[38]. Therefore, antagonizing the interaction ofMIP-1a with its receptor (CCR1) or signal trans-duction from this receptor may hold promisefor future treatment of both acute and latephase reactions. Similarly, in a mouse model ofallergic eye disease, IL-1 inhibition using an IL-1 receptor antagonist was found to downreg-ulate the recruitment of eosinophils and inflam-matory cells by decreasing the concentration ofattractant chemokines [25]. This research alsooffers a potential novel treatment for the pre-vention and treatment of allergic eye disease.

14.5Conclusion

Allergic eye disease represents a heterogeneousgroup of diseases that share a common sympto-mology but different pathogenesis. They arefurther distinguished by their long-term visualprognosis, with diseases such as SAC and PAChaving no long-term effects on sight whereasVKC and AKC, through corneal involvementand subsequent scarring reactions, can adverse-ly affect visual prognosis. Future work needs toincrease our understanding of the genetics andmechanisms of mast cell cytokine expressionand mediator release, the regulation of the

14.5 Conclusion 221

cellular inflammatory response and the B cellregulation of IgE secretion. Armed with thisknowledge, more ways of treating allergic eyedisease will be developed which will target morespecific components of the allergic response.Most novel therapies so far have been directedat controlling the allergic response in thebronchial airways and the nasal mucosa, but it ishoped that new strategies will begin to focustreatment on ocular disease to downregulatethe allergic response rather than to control itseffects.


∑ Allergic eye disease is a common problem.It is reported to affect about 20 % of thepopulation worldwide but this may be anunderestimate of the true prevalence of thecondition due to geographical variationsand the lack of any clear cut objective diagnostic criteria

∑ There are five main syndromes of allergiceye disease, two of which (vernal and atopickeratoconjunctivitis) have sight-threaten-ing complications; hence it is important tostrive to make an accurate diagnosis due tothe prognostic implications

∑ The majority of patients have an atopic ten-dency or a family history of atopy. There isa particularly strong association betweenatopic dermatitis and atopic keratoconjunc-tivitis

∑ The mainstays of treatment for the majorityof allergic eye disease symptoms are topicaleye drops, including antihistamines, mastcell stabilizers and anti-inflammatoryagents

∑ Topical steroid preparations are the mosteffective therapy for moderate to severeforms of allergic eye disease but their useshould be limited to these cases and the eyemonitored carefully for steroid related sideeffects such as cataracts and glaucoma

∑ Topical calcineurin inhibitors may be ofbenefit as steroid sparing agents or in thetreatment of allergic eye disease where the disease is failing to respond to steroidtreatment

References

1. Aguilar AJ (2000) Comparative study of clinicalefficacy and tolerance in seasonal allergic con-junctivitis management with 0.1% olopatadinehydrochloride versus 0.05% ketotifen fumarate.Acta Ophthalmol Scand Suppl 78:52–55

2. Akpek EK, Dart JK,Watson S, Christen W, DursunD, Yoo S, O’Brien TP, Schein OD, Gottsch JD(2004) A randomized trial of topical cyclosporin0.05% in topical steroid-resistant atopic kerato-conjunctivitis. Ophthalmology 111(3):476–482

3. Allansmith MR, Korb DR, Greiner JV, HenriquezAS, Simon MA, Finnemore VM (1977) Giant pap-illary conjunctivitis in contact lens wearers. Am JOphthalmol 83:697–708

4. Artal M, Luna J, Discepola M (2000) A forcedchoice comfort study of olopatadine hydrochlo-ride 0.1% versus ketotifen fumarate 0.05%. ActaOphthalmol Scand 78:64–65

5. Ashrafzadeh A, Raizman MB (2003) New modali-ties in the treatment of ocular allergy. Int Oph-thalmol Clin 43(1):105–110

6. Bawazeer AM, Hodge WG, Lorimer B (2000)Atopy and keratoconus: a multivariate analysis.Br J Ophthalmol 129:667–668

7. Bielory L (2000) Allergic and immunologic dis-orders of the eye, part II: ocular allergy. J AllergyClin Immunol 106:1019–1032

8. Bielory L (2002) Ocular allergy guidelines: a prac-tical treatment algorithm. Drugs 62(11):1611–1634

9. Bielory L, Kempuraj D, Theoharides T (2002)Topical immunopharmacology of ocular aller-gies. Curr Opin Allergy Clin Immunol 2:435–445

10. Bonini S, Bonini S (1998) Pathogenesis of allergicconjunctivitis. In: Denburg JA (ed) Allergy, aller-gic diseases: the new mechanisms and therapeu-tics. Humana Press, Totowa, NJ

11. Bonini S, Bonini S, Lambaste A, March S,Pasqualetti P, Zuccaro O, Rama P, Magrini L, JuhasT, Bucci MG (2000) Vernal keratoconjunctivitisrevisited: a case series of 195 patients with long-term follow up. Ophthalmology 107:1157–1163

12. Bonini S, Coassin M,Aronni S, Lambiase A (2004)Vernal keratoconjunctivitis. Eye 18:345–351

13. Butrus S, Greiner JV, Discepola M, Finegold I(2000) Comparison of the clinical efficacy andcomfort of olopatadine hydrochloride 0.1% oph-thalmic solution and nedocromil sodium 2%ophthalmic solution in the human conjunctivalallergen challenge model. Clin Ther 22:1462–1472

14. Cook EB, Stahl JL, Barney NP, Graziano FM(2000) Olopatadine inhibits TNFalpha releasefrom human conjunctival mast cell. Ann AllergyAsthma Immunol 84:504–508


15. Donshik PC, Pearlman D, Pinnas J, Raizman MB,Tauber J, Tinkelman D, Walters TR (2000) Effica-cy and safety of ketorolac tromethamine 0.5%and levocabastine 0.05%: a multicentre compari-son in patients with seasonal allergic conjunctivi-tis. Adv Ther 17(2):94–102

16. Easty DL, Entwistle C, Fund A, Witcher J (1975)Herpes simplex keratitis and keratoconus in theatopic patient: a clinical and immunologicalstudy. Transcripts Ophthalmol Soc UK 95:267–276

17. Fahy GT, Easty DL, Collum LM, Benedict-SmithA, Hillery M, Parsons DG (1992) Randomizeddouble-masked trial of lodoxamide and sodiumcromoglycate in allergic eye disease. A multicen-tre study. Eur J Ophthalmol 2(3):144–149

18. Foster CS, Calonge MD (1990) Atopic keratocon-junctivitis. Ophthalmology 97:992–1000

19. Fukagawa K, Nakajima T, Saito H, Tsubota K,Shimmura S, Natori M, Hirai K (2000) IL-4 in-duces eotaxin production in corneal keratocytesbut not in epithelial cells. Int Archiv Allergy Im-munol 121:144–150

20. Heinz C, Fanihagh F, Steuhl KP (2003) Squamouscell carcinoma of the conjunctiva in patients withatopic eczema. Cornea 22:135–137

21. Hingorani M, Moodaley L, Calder VL, Buckley RJ,Lightman S (1998) A randomized, placebo-con-trolled trial of topical cyclosporin A in steroid-dependent atopic keratoconjunctivitis. Ophthal-mology 105(9):1715–1720

22. Hogan MJ (1953) Atopic keratoconjunctivitis. AmJ Ophthalmol 36:937–947

23. John T (2000) Contact lens-induced allergic reac-tions. In: Abelson MB (ed) Allergic diseases of theeye. WB Saunders, Philadelphia

24. Kallen C, Reinhard T, Schilgen G, Cartsburg O,Bocking A,Auw-Hadrich C,Sundmacher R (2003)Atopic keratoconjunctivitis: probably a risk fac-tor for the development of conjunctival carcino-ma. Ophthalmologie 100:808–814

25. Keane-Myers AM, Miyazaki D, Liu G, Dekaris I,Ono SJ, Dana MR (1999) Prevention of allergiceye disease by treatment with IL-1 receptor antag-onist. Invest Ophthalmol Vis Sci 40(12):3041–3046

26. Korb DR, Allansmith MR, Greiner JV, HenriquezAS, Herman JP, Richmond PP, Finnemore VM(1981) Biomicroscopy of papillae associated withhard contact lens wearing. Ophthalmology 88:1132–1136

27. Korb DR, Allansmith MR, Greiner JV, HenriquezAS, Richmond PP, Finnemore VM (1980) Preva-lence of conjunctival changes in wearers of hardcontact lenses. Am J Ophthalmol 90:336–341

28. Larkin DFP (1999) Immunological, neoplasticand degenerative diseases of the conjunctiva. In:Easty D, Sparrow J (eds) The Oxford textbook ofophthalmology. Oxford University Press, Oxford

29. Leonardi A, Busato F, Fregona I, Plebani M, SecchiAG (2000) Anti-inflammatory and antiallergic ef-fects of ketorolac tromethamine in the conjuncti-val provocation model. Br J Ophthalmol 84(11):1228–1132

30. Leonardi A, Secchi AG (2003) Vernal keratocon-junctivitis. Int Ophthalmol Clin 43:41–58

31. McGill J (2000) Conjunctival cytokines in ocularallergy. Clin Exp Allergy 30:1355–1357

32. McGill JI, Holgate ST, Church MK, Anderson DF,Bacon A (1998) Allergic disease mechanisms. Br JOphthalmol 82:1203–1214

33. Messmer EM, May CA, Stefani FH,Welge-LuessenU, Kampik A (2002) Toxic eosinophil granule pro-tein deposition in corneal ulceration and scarsassociated with atopic keratoconjunctivitis. Am JOphthalmol 134:816–821

34. Metz DM, Bonini S, Lightman S (1993) mRNA ex-pression for interleukin-2 interleukin-5 in vernalkeratoconjunctivitis. Invest Ophthalmol Vis Sci34:772

35. Metz DP, Bacon AS, Holgate ST, Lightman S (1996)Phenotypic characterization of T-cells infiltrat-ing the conjunctiva in chronic allergic eye dis-ease. J Allergy Clin Immunol 98:686–696

36. Miyazaki D, Liu G, Clark L, Ono SJ (2000) Preven-tion of acute allergic conjunctivitis and late phaseinflammation with immunostimulatory DNA se-quences. Invest Ophthalmol Vis Sci 41(12):3850–3855

37. Miyazaki D, Nakamura T, Komatsu N, Nawata N,Ikeda Y, Inoue Y, Higashi H, Ono SJ (2004) Rolesof chemokines in ocular allergy and possibletherapeutic strategies. Cornea 23(Suppl 1):S48–54

38. Miyazaki D, Nakamura T, Toda M, Cheung-ChauKW, Richardson R, Ono SJ (2005) Macrophage in-flammatory protein-1alpha as a costimulatorysignal for mast cell mediated immediate hyper-sensitivity reactions. J Clin Invest 115:434–442

39. Montan PG, Ekstrom K, Hedlin G, van Hage-Hamsten M, Hjern A, Herrmann B (1999) Vernalkeratoconjunctivitis in a Stockholm ophthalmiccentre: epidemiological, functional and immuno-logical investigations. Acta Ophthalmol Scand77:559–563

40. Nishimura A, Campbell-Meltzer RS, Chute K, Or-rell J, Ono SJ (2001) Genetics of allergic disease:evidence for organ-specific susceptibility genes.Int Archiv Allergy Immunol 124:197–200

References 223

41. Ono SJ, Abelson MB (2005) Allergic conjunctivi-tis: update on pathophysiology and prospects forfuture treatment. J Allergy Clin Immunol 115:118–122

42. Raizman MB, Rothman JS, Maroun F, Rand WM(2000) Effect of rubbing on signs and symptomsof allergic conjunctivitis in cat-sensitive individ-uals. Ophthalmology 107(12):2158–2161

43. Santos CI, Huang AK, Abelson MB, Foster CS,Friedlaender M, McCulley JP (1994) Efficacy oflodoxamide 0.1% ophthalmic solution in resolv-ing corneal epitheliopathy associated with vernalkeratoconjunctivitis. Am J Ophthalmol 117(4):488–497

44. Sharif NA, Xu SX, Miller ST, Gamache DA, YanniJM (1996) Characterization of the ocular antial-lergic and antihistaminic effects of olopatadine(AL-4943A), a novel drug for treating ocularallergic diseases. J Pharmacol Exp Ther 278(3):1252–1261

45. Sitenga GL, Ing EB, Van Dellen RG, Younge BR,Leavitt JA (1996) Asthma caused by topical appli-cation of ketorolac. Ophthalmology 103(6):890–892

46. Solomon A, Pe’er J, Levi-Schaffer F (2001) Ad-vances in ocular allergy: basic mechanisms, clin-ical patterns and new therapies. Curr Opin Aller-gy Clin Immunol 1:477–482

47. Solomon A, Puxeddu I, Levi-Schaffer F (2003) Fi-brosis in ocular allergic inflammation: recentconcepts in pathogenesis of ocular allergy. CurrOpin Allergy Clin Immunol 3:389–393

48. Stahl JL, Barney NP (2004) Ocular allergic disease. Curr Opin Allergy Clin Immunol 4:455–459

49. Tabbara KF (2003) Immunopathogenesis ofchronic allergic conjunctivitis. Int OphthalmolClin 43(1):1–7

50. Takano Y, Fukagawa K, Dora M, Santo-Kato N,Tsubota K, Fujishima H (2004) Inflammatorycells in brush cytology samples correlate with theseverity of corneal lesions in atopic keratocon-junctivitis. Br J Ophthalmol 88:1504–1505

51. Trocme SD, Karan KS (2002) Spectrum of ocularallergy. Curr Opin Allergy Clin Immunol 2:423–427

52. Tuft SJ, Dart JK, Kemeny M (1989) Limbal vernalkeratoconjunctivitis: clinical characteristics andIgE expression compared with palpebral vernal.Eye 3:420–427

53. Tuft SJ, Kemeny DM, Dart JK, Buckley RJ (1991)Clinical features of atopic keratoconjunctivitis.Ophthalmology 98(2):150–158

54. Tuft SJ, Ramakrishnan M, Seal DV, Kemeny DM,Buckley RJ (1992) Role of Staphylococcus aureusin chronic allergic conjunctivitis. Ophthalmology99:180–184

55. Wakai K, Okamoto K, Tamakoshi A, Lin Y,Nakayama T, Ohno Y (2001) Seasonal allergicrhinoconjunctivitis and fatty acid intake: a cross-sectional study in Japan. Ann Epidemiol 11(1):59–64

56. Yokoi K, Yokoi N, Kinoshita S (1998) Impairmentof ocular surface epithelium barrier function inpatients with atopic dermatitis. Br J Ophthalmol82:797–800

57. Yoneda K, Okamoto H, Wada Y, Morita K, Taka-hashi M, Ogura Y, Imamura S (1995) Atopic retinaldetachment: report of 4 cases and a review of theliterature. Br J Dermatol 133(4):586–591


AABCG2 40ACAID (Anterior chamber associated immune

deviation) 110Acanthamoeba keratitis 153, 196Adenoviral keratoconjunctivitis

see keratoconjunctivitisAir drying 25Airdissection 67Albumin 17Allergen 214Allergic eye disease 209Allograft rejection 65Allorecognition 110– direct pathway 76, 110– indirect pathway 110Amnion 22– composition 22– structure 22Amniotic epithelium 26Amniotic membrane 21, 49, 51, 52, 57, 58Amphotericin B 0.15% 158Anecortave acetate 95Angiogenesis 83Angiogenic growth factor see growth factorAniridia 83Anterior chamber associated immune deviation

see ACAID 110Antiangiogenic therapy 83Anticoagulant 6Antigen presenting cells (APCs) 76, 105Antihistamine 217, 218Antilymphangiogenic therapy 83Antithymocyte globulins 117APCs (antigen presenting cells) 105Arlt’s lines 215Artificial anterior chamber 129Asthma 209Astigmatism 123, 136– irregular 123, 124Atopic keratoconjunctivitis see kerato-

conjunctivitisAuto-limbal transplantation 48, 50

Autologous serum 1– eyedrops 4, 52Autologous transplantation 59Avastin 95Azathioprine 112

BBandage contact lens 51Basal cells 184Basement membrane 24, 58Basiliximab 112, 117Bioengineering 57Biomicroscopy, slit-lamp 174Blood group antigen 106Blood product 6Bowman’s membrane 187Broad antigen 102Buccal mucosa graft 35Bullous keratopathy 30

CCalcineurin inhibitor 220Candida albicans 159Caustication 61CD4+ T cells 110, 212CD8+ T cells 110, 212Centration 123, 134Certican 116Chemical burns 83Chronic limbitis 41Cidofovir 163, 168– topical 167Collagen 22, 52Conjunctiva 177Conjunctival hyperreactivity– non-specific 211Conjunctivalisation 35Conjunctivitis– giant papillary 209, 217– perennial allergic 209, 214– seasonal allergic 209, 213, 214Connexin 43Contact lens 83, 90, 173, 192

Subject Index

Contamination 15Corneal angiogenic privilege 84Corneal basement membrane dystrophy 13Corneal dissection– laser 68– manual 68Corneal epithelial thickness 173Corneal epithelium 58, 180– basal cells 180– Bowman’s membrane 180– intermediate cells 180– stroma 180– superficial cells 180Corneal lymphatic vessel 94Corneal nerve 173, 185Corneal oedema 75Corneal opacity– nummular 165– persistence 168– subepithelial 165 Corneal perforation 28Corneal surgery, refractive 173, 200Corneal transplantation 88Corneal ulcer 144, 199Corneoscleral tissue 59Corticosteroid 52, 114– topical 219Cortisone 112Cryopreservation 26Cut angle 129Cyclophosphamide 112Cyclosporin A (CsA) 52, 109, 112, 114, 118, 163, 220– topical 169– eyedrops 98Cystic epithelial change 191Cytokeratin profile 39Cytotoxic T lymphocyte 77

DDaclizumab 112, 117Delayed type hypersensitivity 77Dendritic cells 173, 209Dennie-Morgan fold 216Descemet’s membrane 188Digital photography 175Donor 123Dry eye 5, 8, 10, 11, 190Dystrophy 144– stromal 144

EEmbryonic stem cells 37Endothelial cells 188– density 66Epidemic keratoconjunctivitis

see keratoconjunctivitis

Epitheliotrophic factor 1, 2, 3, 5Epithelium 42, 57, 181– conjunctival 58– corneal see corneal epithelium Erbium: YAG Laser 147Everolimus 109, 116Excimer Laser 193-nm 144Extracapsular cataract extraction 138Extracellular matrix 37

FFibrin 52Fibroblast feeder layer 58Fibronectin 2, 22FK506 112, 114, 220Flap lift 153, 155Flieringa ring 133Fluoroquinolone 158FTY 720 116Fuchs’s endothelial dystrophy 111, 129, 145Functional imaging 173, 203Fungal infection 153

GGas-permeable lens 217Gatifloxacin 158Giant papillary conjunctivitis

see conjunctivitisGlutaraldehyde 25Goblet cell 58Graft failure 144Graft rejection 105Graft size 123, 129, 133Graft survival 102Gram-negative organism 153Growth factor 5, 23, 24– angiogenic 86– TGF-b1 26Guided trephine system see trephine

HHA-3 105Hand-held trephine see trephineHanna trephine see trephineHay fever (see also conjunctivits,

seasonal allergic) 213Herpes simplex keratitis see keratitisHertoge’s sign 216HLA 102– class I molecules 102– class II molecules 102– matching 52, 61, 78, 102– typing 52, 102HLAMatchmaker 102– algorithm 104, 107Horizontal torsion 127

226 Subject Index

Horner-Trantas dot 215H-Y 105Hypersensitivity type I 210

IIL (Interleukin)-1 114– inhibition 221IL-2– antagonist 112– inhibitor 109IL-6 114Immune privilege 76Immune rejection 94Immunoglobulin E 210Immunomodulatory cytokine 94Immunosuppressive molecule 76Immunosuppressive strategy 109Impression cytology 39, 60Infiltrates 165Inflammation 29Inflammatory cells 173Instrastromal ring segment 67Interface haze 65, 68Interferon 163, 169Interleukin (see also IL) 210Intermediate cells 184Intraoperative adjustment 123Irradiation 25

KKeratectomy, phototherapeutic (PTK) 67, 129Keratinisation 43Keratitis 83, 158, 159– acanthamoeba 196– diffuse lamellar (sands of the Sahara) 154– fungal 158– herpes simplex 69, 91, 118, 159, 216– infective 153Keratoconjunctivis– adenoviral 163– atopic 209, 215– epidemic 195– limbal 215– superior limbic 13– vernal 209, 214Keratoconus 67, 69, 111, 123, 136, 144, 216Keratocytes– viral persistence in 195Keratometric refractive power 124Keratometry 125Keratopathy 129– aphakic/pseudophakic bullous 129, 144Keratoplasty 64, 129– deep anterior lamellar 65, 69– high-risk 90, 109– indications 144– low-risk 94, 104

– penetrating 65, 69 – posterior lamellar 129, 148Keratoprothesis 64Keratotomy– radial 138

LLaminin 22Langerhans cells 52, 105, 177, 184Laser 123– femtosecond 123, 148LASIK 123, 153Leflunomide 112Limbal stem cells 35– culture 57– deficiency 35, 111– causes 41– ex vivo expansion 52, 57Limbal epithelial crypt 40Limbal palisades of Vogt 38, 44Limbal region 188Limbal transplantation 45Lymphangiogenesis 83Lymphocytes– CD4+ T cells 110, 212– CD8+ T cells 110, 212– Cytotoxic T 77Lyophilisation 25

MMacrolide antibiotic 220Macrophage inflammatory protein-1a

(MIP-1a) 221Mapping 125Mast cell 210– degranulation 212– stabilizer 217– agents 218Mechnical barrier 76Meesmann’s dystrophy 190Metalloprotease 23– inhibitors 23Methotrexate 112MHC (major histocompatibility

complex) 102, 110– class II molecule 76, 210Microbiological examination 155– scraping of material 155Microenvironment 37Microscopy 173– confocal 173, 175– laser-scanning 176– specular 173Minor H mismatch 105Minor matching 102MIP-1a (Macrophage inflammatory protein-1a) 221Mismatches, class I and II 78

Subject Index 227

MMF see mycophenolate mofetilMolecular typing 107Moxifloxacin 158Muromonab-CD3 117Mycobacteria 158– non-tuberculous 153Mycophenolate mofetil 52, 109, 112, 114

NNatamycin 5% 158Neovascularisation 29, 88Neurotrophic keratitis 8, 9, 41– diabetic 9– postherpetic 9NMSO3 (sulphated sialyl lipid) 163Non-steroidal anti-inflammatory drug (NSAID) 219NSAID (Non-steroidal anti-inflammatory drug) 219Nummular subepithelial infiltrates 75

OOcular surface burn 35Ocular surface reconstruction 57

Pp63 (transcriptional factor) 39Pachymetry 194PCIOL (Posterior chamber intraocular lens) 138– toric 138Pemphigoid 41Perennial allergic conjunctivitis

see conjunctivitisPeritomy 48Plasma 3, 17Posterior chamber intraocular lens (PCIOL) 138Postoperative treatment 65Progenitor cell 36Prograf see tacrolismusProinflammatory cytokine 25Prostaglandin 25, 219Proteoglycan 22Pseudogerontoxon 215Pterygium surgery 30Punctual occlusion 11

RRAD (Everolimus, Certican) 116Rapamune 115Rapamycin 52, 109, 115, 118RCM (Rostock cornea module) 178Recurrent erosion syndrome 13Regularity 123Rejection 73– acute 110– chronic 110– endothelial 73, 75– epithelial 75 – stromal 75Rheumatoid arthritis 8

Rubbing 212, 216– degranulation of the mast cells 212

SSands of the Sahara 154Scarring 29Seasonal allergic conjunctivitis see conjunctivitisSequential sector conjunctival epitheliectomy 46Severe atopic dermatitis 111Shield cataract 217Shield ulcer 215Sirolimus 112, 115, 116Sjögren’s syndrome 11Slit-scanning technique 175– confocal microscopy 175Stem cells, embryonic 37Stem cell niche 37, 40Sodium cromoglygate 218Split antigen 102Steroid 109Steroid, topical 111, 166, 167Stevens-Johnson syndrome 41Superficial cells 183Suture removal 125Symblepharon 28Systemic immunosuppression 61

TTacrolimus 52, 109, 112, 114, 118, 220T-cell activation 114Tear film 42, 182– dynamics 173Tear substitute 2Topical steroid 77Topography 125– regularity of 123TOR inhibitor 109Toxicity, local 168Trachoma 41Transcriptional factor p63 39Transplant 53– allo-limbal 53– auto-limbal 53Trephination 129– nonmechanical 129– techniques 130– elliptical 137– nonmechanical laser 143– host 123Trephine 138– guided system 138, 142– Hanna 143– hand-held 141– motor 141– suction 142Triple procedure 138Triplet 104Tylosis 216

228 Subject Index

Date post:	03-Oct-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

Cornea and External Eye Disease - Beximco Pharma · Anterior lamellar keratoplasty is presented by...

Documents