988

significantly different from those in controls. We think thatconjugates formed between polymerisation, degradation, ortransformation products and carrier proteins or cellmembranes 5 are the antigens responsible for thesensitisation. The rate at which penicillin-G degradationproducts are formed over 24 h suggests that penicillin-Gsolutions which are not given time to form these

derivatives-i.e., when freshly prepared solutions are usedand given as a bolus-do not seem to be immunogenic whengiven intravenously. The offending antigens seem unlikely tobe related to penicilloylated high molecular weightimpurities in the penicillin preparation since both group Aand B patients received the same products. It is generally saidthat penicilloylated protein impurities already present in thepenicillin preparations are responsible for most cases ofsensitisation against penicillins, 6,7 but our study indicates theimportance of freshly produced degradation products, suchas penicillenic acid, in the sensitising process.

The value of preventing titres of penicillin-specific IgGantibodies from rising after high-dose intravenous therapyremains to be seen. It is known, however, that IgG antibodiesare involved in several adverse reactions to penicillins.8,9 Thebest documented reaction-i.e., the immune haemolyticanaemia-occurs only when high anti-penicillin IgG titres areproduced, as after high dose therapy of sufficient duration;9such is also the case with severe penicillin-inducedneutropenia.’ None of our subjects treated with penicillinhad anti-penicilloyl IgE and immediate type allergy,probably because high doses of allergen tend to favour IgGrather than IgE production. However, another possiblereason for avoiding rises in IgG production could be thatpriming of the immune system by immunogenic penicillinderivatives might favour the later development of IgE-mediated penicillin hypersensitivity upon renewed therapy.

In conclusion, we suggest that penicillin should not beallowed to age, either during storage at 4°C or in infusionbags at room temperature. The use of vials containing 10million IU or more ofpenicillin-G should be discontinued forroutine therapy since such doses are usually administeredover several hours. Only smaller, single dose vials should beavailable.

We thank Ms M. Gugler, Ms J. Tilch, and Ms B. Grassi for technicalassistance and Prof. H. Biirgi, Biirgerspital Solothurn, for his interest andvaluable advice.

Correspondence should be addressed to K. A. N., Department of InternalMedicine, Medical Clinic, University Hospital, 8091 Zurich, Switzerland.

REFERENCES

1. Neftel KA, Wälti M, Spengler H, von Felten A, Weitzman SA, Bürgi H, de Weck AL.Neutropenia after penicillins: Toxic or immune-mediated? Klin Wochenschr 1981;59: 877-88.

2. Schneider CH. Immunochemistry of penicillin. In: Stewart GT, McGovern JP, eds.Penicillin allergy. Springfield: Charles C. Thomas, 1970: 23-58.

3. Weck AL de, Schneider CH, Gutersohn J. The role of penicilloylated protein impurities,penicillin polymers and dimers in penicillin allergy. Int Arch Allergy 1968; 33:535-67.

4. Schellekens PTA, Eijsvogel VP. Lymphocyte transformation in vitro. Clin Exp Immunol1968; 3: 571-84

5. Weck AL de. Molecular mechanisms of T and B lymphocyte triggering. Int Arch Allergy1975, 49: 247-54.

6. Batchelor FR, Dewdney JM, Feinberg JG, Weston RD. A penicilloylated proteinimpurity as a source of allergy to benzylpenicillin and 6-aminopenicillanic acid.Lancet 1967; i: 1175-77.

7. Ahlstedt S, Ekström B, Svärd PO, Sjöberg B, Kristofierson A, Ortengren B. Newaspects on antigens in penicillin allergy. CRC Crit Rev Toxicol 1980; 1: 219-77

8. Colvin RB, Burton JR, Hyslop NE, Spitz L, Lichtenstein NS. Penicillin-associatedinterstitial nephritis. Ann Intern Med 1974; 81: 404-05.

9. Petz LD, Garratty G. Acquired immune hemolytic anemias. Edinburgh: ChurchillLivingstone, 1980.

DOES CAPTOPRIL CAUSE RENAL DAMAGE INHYPERTENSIVE PATIENTS?

Report from the Captopril Collaborative StudyGroup*

Summary Renal biopsy material was studied from 38hypertensive patients, 19 of whom had

received captopril for at least six months. Comparableimmunofluorescence findings and prominent arteriolar andmesangial deposits of C3 and IgM were noted equally in allpatients. Within the non-captopril group one biopsy showedmesangial IgA glomerulonephritis and, in another,micropscopy revealed typical membranous glomerulo-nephritis. The biopsy specimens from 4 patients in the

captopril group showed evidence of glomerular deposits and2 had typical membranous glomerulonephritis. Small,spherical electron-dense deposits were seen with equalfrequency along the outer aspect of the glomerular basementmembrane in both groups, but were not thought to beimmune deposits. Renal biopsy material from hypertensivepatients may have ultrastructural and immunohistologicalabnormalities which cannot be ascribed to captopril.Subclinical glomerulonephritis appears to exist in the

hypertensive population. The membranous glomerulo-nephritis could be drug-related, but this cannot bedetermined with certainty from the small sample studied.Caution must be shown in attributing glomerulonephritis toany drug without information about the kidneys beforetreatment.

Introduction

IN a study to assess the effects of the angiotensin convertingenzyme inhibitor, captopril, on patients with systemichypertension, nephrotic syndrome developed in one patientand a serum sickness-like condition in another. 1,2 Because ofthe possibility of subclinical glomerular lesions developing inthis population, renal biopsy specimens from hypertensive,non-proteinuric patients were studied.3 Electron microscopyrevealed dense spherical deposits along the epithelial side ofthe glomerular capillary basement membrane. The possiblerelation of these bodies to immune deposits was considered.Deposits of immunoglobulins and complement (C3) werealso demonstrated in the glomeruli of these patients. Thesefeatures raised several questions in relation to membranousglomerulonephritis, hypertension, and captopril treatment.,We therefore decided to study renal biopsy specimens from agroup of hypertensive patients who had not received

captopril (but were under consideration for treatment withthis drug) and to compare the results with those in patientswho had been treated with captopril for at least six months.

Patients, Materials, and MethodsRenal biopsy material taken from patients during a separate study

at the State University of Groningen was used in this study. Theconsent of the patients and of the university ethical committee wasobtained.3 All the patients in the Groningen study were about to be

*Organising committee: Prof. ROBERT H. HEPTINSTALL (Johns HopkinsUniversity School of Medicine): Prof. PHILIP J. HOEDEMAEKER (StateUniversity of Groningen, Netherlands); Prof. EDMUND J. LEWIS (RushMedical College); and Prof. ROBERT W. SCHRIER (University of ColoradoSchool of Medicine).Other participants: Prof. Robert T. McCluskey (Massachusetts General

Hospital, Harvard Medical School); Dr Jan J. Weening, Ms Susan van derLaan, Dr A.J.M. Donker, Dr S. J. Hoorntje, Dr C. G. M. Kallenberg, andProf. J. K. van der Hem (State University of Groningen): Prof. Jacob Churg(Mount Sinai School of Medicine); and Prof. Giuseppe Andres (StateUniversity of New York at Buffalo).

989

Fig I-Small spherical dense body (centre) along outer aspect ofglomerular basement membrane in captopril patient.

Ep=epithelial cell; En = endothelial cell (>30 000).

treated with captopril for its antihypertensive properties, but onlyafter other antihypertensive regimens had been tried

unsuccessfully. There was concern about the possible side-effects ofcaptopril, particularly glomerulonephritis, and for this reason

patients were asked to undergo renal biopsy in order to monitor anyglomerular reactions in the future. Epon-embedded material wasexamined from 20 of these patients. Specimens had first beenprocessed and interpreted in the university ultrastructuralpathology laboratory in Groningen. Coded, embedded specimenswere then sent to 2 independent pathologists (J. C. and G. A.). Tento twenty capillary loops were photographed at a magnification’ofx8000. The presence, amount, and location of immune depositsand spherical dense bodies was recorded. The spherical bodies weredenser, smaller, and more discrete than the usual electron-densedeposit associated with immune aggregates (fig. 1). The numbers ofcapillary loops studied and of loops with deposits, the total numberof deposits, and the number of deposits according to their location inthe capillary wall, were recorded.After interpretation at the department of pathology in Groningen,

frozen renal biopsy specimens (unstained cryostat sections) takenfrom 14 hypertensive patients before treatment with captopril, andfrom 11 patients in the post-captopril group, were coded anddistributed to two immunopathologists (E. J. L. and R. T. McC).Specimens were stained with standardised fluorescein-labelledantibody. Monospecific antisera to human IgG, IgA, IgM, C3, andproperdin used in this study were those currently standardised forthe International Study of Kidney Disease in Children. Thesestained sections were interpreted separately and photographed byeach laboratory. If discrepancies occurred, slides were restained anda consensus was reached.Clinical records and light microscopy sections from all patients

were reviewed. Student’s t-test was used for analysis of theultrastructural findings and X2 test for the immunofluorescenceresults.

Results

Spherical dense bodies were found as frequently and insimilar numbers in the untreated hypertensive patients as inthose treated with captopril. Typically, 1-2 sphericalbodies/16 capillary loops/biopsy were seen. These findingsagreed with those from the Groningen laboratory. Moreover,

IMMUNOFLUORESCENCE STUDIES

Fig. 2-C3 in hilar arteriole and granular deposits throughout theglomerular mesangium.

A few of these granules could be paramesangial or peripheral in position,and some can be seen along Bowman’s capsule adjacent to the arteriole andwithin some tubular basement membranes.

there was no consistent relation between the presence of

spherical bodies and the immunofluorescence findings,indicating that they were not immune aggregates.The commonest finding on immunofluorescence was

deposits of C3 in the arterioles of the glomerular hilus (seetable and fig. 2). These were present in both groups ofpatients. In some patients other "vessels also contained thisprotein. Granules of C3 were seen in the mesangium in mostpatients in both groups. Granular C3 deposits were

sometimes found in the paramesangial and peripheralportions of the capillary loops. Properdin was present lessoften, but in a similar distribution to C3. Small amounts ofIgM were also found in the blood vessels and glomerularmesangium in most cases. IgM was found more frequently inthe non-captopril group, but not significantly so (p=0.1).The only disagreement between laboratories involved theinterpretation of minimal staining for trace amounts of

immunoglobulins in the mesangium. There were no

significant differences between glomerular or arteriolar C3,or the immunoglobulin deposits found in the two groups.4 patients in the captopril group had ultrastructural or

immunofluorescence findings suggestive of significantglomerular damage. The presence of electron-dense depositsof the type found in membranous glomerulonephritis wasconfirmed in cases 1 and 2 of the original study,3 whichcorrelated well with immunoglobulins and complementfound in the one case in which these studies were available

(no. 1 from ref. 3). Unfortunately, frozen tissue was notavailable from case no. 2, but the microscopy findingssupported the original Groningen diagnosis of membranousglomerulonephritis. Another biopsy specimen from a patienton captopril, not included in the original paper, showed

990

mesangial and subepithelial deposits. Immunoglobulins andcomplement were also seen at these sites. However, thesefindings were not typical of uncomplicated membranousglomerulonephritis. The biopsy specimen from a fourthpatient on captopril showed mesangial sclerosis withscattered mesangial, epimembranous, and intramembranouselectron-dense deposits. Material for immunofluorescencewas not available. Two biopsies taken before giving captoprilhad abnormalities on electron microscopy in addition to thepresence of spherical bodies. One showed mesangial,intramembranous, and subendothelial electron-dense

deposits, and immunofluorescence findings compatible withmesangial IgA nephropathy, but no abnormal urinaryfindings. One biopsy showed subepithelial deposits with agranular immunofluorescence pattern for immunoglobulinsand complement characteristic of membranous

glomerulonephritis. This hypertensive patient had provenproteinuria and was therefore not put on captopril.

Discussion

We have shown that many of the glomerular abnormalitiesseen in patients on captopril are also present in a hypertensivepopulation not receiving the drug. The spherical densebodies reported previously were found in both groups ofpatients in similar numbers. Pre-treatment and post-treatment biopsy specimens from a group of hypertensivepatients on captopril have revealed that the spherical densebodies do not increase in size or number after 6 months’treatment (Hoedemaeker P. and van der Laan S.,unpublished). There was no correlation between the findingof spherical dense bodies and the detection ofimmunoglobulins and complement by immunofluorescence.There is therefore no evidence that the spherical dense bodiesrepresent immune aggregates. Neither the nature nor the

significance of these structures was apparent from ourstudies. Further proof that they are probably unrelated toimmunological processes is that pathologists, including thoseinvolved in this study, have seen them in conditions such aslipoid nephrosis (minimal change idiopathic nephroticsyndrome), recurrent haematuria with negativeimmunofluorescence, and in 1-hour post-transplantationbiopsy specimens.4 However, their numbers were not

recorded. It is remarkable, in retrospect, that the dense bodieshave attracted so little attention although they have beenreported from time to time. Bariety et al. described "roundextracellular particles" in some detail and listed the manyconditions in which they had found them, and also quotedother authors who had seen them.We also describe the immunoglobulin and complement

component deposits that may be seen in the kidneys ofhypertensive patients. The results demonstrate the carewhich must be taken in the interpretation of the finding of aparticular protein of the immune system in the renal

glomerulus of such patients. The immunofluorescence

findings in the two groups reflect: (a) immunoglobulin andcomplement component deposits attributable to

hypertension; (b) immunoglobulin deposits associated withpre-existing glomerulonephritis-either occult or manifest;and (c) membranous glomerulonephritis.The significance of the presence of C3 and IgM in the

hypertensive patients is not known, but may represent areaction to capillary damage.

’

The interpretation of immunofluorescence findings whichappear characteristic of immune-mediated glomerulo-

nephritis is perplexing. The lesion of either occult or manifestpre-existing glomerulonephritis may be present in the

hypertensive population. One patient in the pre-captoprilgroup had deposits on immunofluorescence microscopy andelectron microscopy which appeared compatible with

mesangial IgA nephropathy. A second patient in the samegroup had manifest membranous glomerulonephritis, andhad proteinuria at the time of biopsy and therefore presentedno clinical problem in differentiation. However, Noel et al.6have pointed out that patients with membranous

glomerulonephritis can have resolution of all urinaryabnormalities while continuing to be hypertensive. Thefrequency of this phenomenon in the hypertensivepopulation is unknown. One of the biopsies which showedsubepithelial deposits in the non-proteinuric captopril grouphad intramembranous and mesangial deposits with mesangialsclerosis, strongly suggesting that the lesion was not theresult of, and could have antedated, treatment.Membranous glomerulonephritis causing symptoms has

been reported in several patients on captopril. 7,8 Scrutiny ofour captopril-treated patients confirmed that 2 of the biopsyspecimens of the 13 cases originally reported had thepathological features of pure membranous glomerulo-nephritis.3 In 2 cases both subepithelial and mesangialelectron-dense deposits were seen-an abnormality which hasneither been reported in the membranous glomerulo-nephritis associated with drugs such as penicillamine orgold,9-13 nor in other cases of membranous glomerulo.nephritis in patients receiving captopril.3,8 The relation totreatment of combined subepithelial and mesangial depositsis unclear.Renal biopsy material from patients with hypertension may

reveal ultrastructural and immunohistological abnormalitieswhich cannot be ascribed to captopril or any other drugs.Caution should be exercised in attributing glomerular diseaseto drugs when no pre-existing biopsy specimens are available.Correspondence should be addressed to Prof. Edmund J. Lewis, Section of

Nephrology, Rush-Presbyterian-St. Luke’s Medical Center, 1753 West

Congress Parkway, Chicago, Illinois 60612, U.S.A.

REFERENCES

1. Prins EJL, Hoorntje SJ, Weening JJ, Donker AJM. Nephrotic syndrome in a patient oncaptopril. Lancet 1979; ii: 306.

2. Hoorntje SJ, Weening JJ, Kallenberg CGM, Donker AJM. Serum-sickness-likesyndrome with membranous glomerulopathy in a patient on captopril. Lancet 1979;ii: 1297.

3. Hoorntje SJ, Kallenberg CGM, Weening JJ, Donker AJM, The TH, Hoedemaeker PJ.Immune complex glomerulopathy in patients treated with captopril. Lancet 1980; i;1212.

4. Kincaid-Smith P, Whitworth JA, Walter NMA, Dowling JP. Immune complexglomerulopathy and captopril. Lancet 1980; ii: 37.

5. Bariéty J, Callard P, Appay MD, Grossetete J, Mandet CH. Ultrastructural study ofsome frequent and poorly known intraglomerular structures. In: Hamburger J,Crosnier J, and Maxwell MH, eds. Advances in nephrology from the NeckerHospital. Chicago: Year Book Medical Publishers Inc, 1974: vol 3, 153

6. Noel LH, Sanetti M, Droz D, Barbanel C. Long-term prognosis of idiopathicmembranous glomerulonephritis. Study of 116 untreated patients. Am J Med 1979,66: 82.

7. Rosendorff C, Milne FJ, Levy H, et al. Nephrotic syndrome during captropril therapyS Afr Med J 1980; 11 586.

8. Case DB, Atlas SA, Mouradian JA, Fishman RA, Sherman RL, Laragh JH Proteinuriaduring long-term captopril therapy. JAMA 1980; 346: 244.

9. Samuels B, Lee JC, Engleman EP, Hopper J, Jr. Membranous nephropathy in patientswith rhuematoid arthritis: Relationship to gold therapy. Medicine 1978; 57: 319.

10. Törnroth T, Skrifvars B. Gold nephropathy: Prototype of membranous

glomerulonephritis. Am J Pathol 1974, 75: 573.11. Watanabe I, Whittier FC, Jr, Moore J, Cuppage FE. Gold nephropathy

Ultrastructural, fluorescence, and microanalytic studies of two patients. Arch PatholLab Med 1976; 100: 632.

12. Bacon PA, Tribe CR, Mackenzie JC, Jones JV, Cumming RH, Amer B. Penicillaminenephropathy in rheumatoid arthritis: A clinical, pathological and immunologicalstudy. Quart J Med 1976; 45: 661.

13. Gärtner H-V. Membranous (peri-epi-extra-membranous) glomerulonephrito-prototype of an immune complex glomerulonephritis. A clinical morphologicalstudy. Veroeff Morph Pathol 1980; 113: 1.