aDesmond Tutu HSouth Africa, bDivLondon, UK, cInfeSouth Africa, eFacHospital, Ubon Rof Oxford, OxfordNijmegen, The NeMedicine, London
Correspondence tSt. George’s Hosp
Tel: +44 208 725Received: 24 Sept
DOI:10.1097/QAD
ISS
Objectives: To assess impact of serial lumbar punctures on association betweencerebrospinal fluid (CSF) opening pressure and prognosis in HIV-associatedcryptococcal meningitis; to explore time course and relationship of opening pres-sure with neurological findings, CSF fungal burden, immune response, and CD4 cellcount.
Design: Evaluation of 163 HIV-positive ART-naive patients enrolled in three trials ofamphotericin B-based therapy for cryptococcal meningitis in Thailand and SouthAfrica.
Methods: Study protocols required four lumbar punctures with measurements ofopening pressure over the first 2 weeks of treatment and additional lumbar puncturesif opening pressure raised. Fungal burden and clearance, CSF immune parameters, CD4cell count, neurological symptoms and signs, and outcome at 2 and 10 weeks werecompared between groups categorized by opening pressure at cryptococcal meningitisdiagnosis.
Results: Patients with higher baseline fungal burden had higher baseline openingpressure. High fungal burden appeared necessary but not sufficient for developmentof high pressure. Baseline opening pressure was not associated with CD4 cell count,CSF pro-inflammatory cytokines, or altered mental status. Day 14 opening pressure wasassociated with day 14 fungal burden. Overall mortality was 12% (20/162) at 2 weeksand 26% (42/160) at 10 weeks, with no significant differences between openingpressure groups.
ippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
IV Centre, Institute of Infectious Disease and Molecular Medicine, University of Cape Town, Cape Town,ision of Infectious Diseases, Department of Cellular and Molecular Medicine, St. George’s University ofctious Diseases Unit, GF Jooste Hospital, dDepartment of Medicine, University of Cape Town, Cape Town,ulty of Tropical Medicine, Mahidol University, Bangkok, fDepartment of Medicine, Sappasithiprasongatchathani, Thailand, gCentre for Tropical Medicine, Nuffield Department of Clinical Medicine, University, UK, hDepartment of Internal Medicine and Infectious Diseases, Radboud University Medical Centre,therlands, and iDepartment of Epidemiology and Population, London School of Hygiene and Tropical, UK.
o Dr Tihana Bicanic, Division of Infectious Diseases, Department of Cellular and Molecular Medicine,ital Medical School, Cranmer Terrace, London SW 17 ORE, UK.
5828; fax: +44 208 725 3487; e-mail: [email protected] 2008; revised: 24 November 2008; accepted: 5 December 2008.
.0b013e32832605fe
N 0269-9370 Q 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins 701
Conclusion: Studies are needed to define factors, in addition to fungal burden,associated with raised opening pressure. Aggressive management of raised openingpressure through repeated CSF drainage appeared to prevent any adverse impact ofraised opening pressure on outcome in patients with cryptococcal meningitis. Theresults support increasing access to manometers in resource-poor settings and routinemanagement of opening pressure in patients with cryptococcal meningitis.
� 2009 Wolters Kluwer Health | Lippincott Williams & Wilkins
The mortality of HIV-associated cryptococcal meningitisremains unacceptably high, ranging from 20 to 40% at10 weeks despite amphotericin B (AmB)-based therapyand access to antiretroviral therapy (ART) [1–3]. Inaddition to insufficiently active antifungal therapy, raisedcerebrospinal fluid (CSF) opening pressure may be thesecond contributing factor. The pathophysiology is notclearly defined: obstruction of CSF outflow by organismor polysaccharide capsule is thought to occur at thearachnoid villi [4,5], producing a communicating hydro-cephalus without ventricular dilatation. Manifestationsinclude headache, vomiting, visual and hearing loss,cranial neuropathies and depressed conscious level,though raised opening pressure may be clinically silent[6]. Supporting the concept of CSF outflow obstruction,high volume lumbar puncture and other CSF diversionprocedures often produce immediate symptom relief aswell as reversing neurological morbidity, most notablyblindness [7,8].
In a large Mycoses Study Group (MSG) trial [6], raisedopening pressure was common and associated with higherbaseline CSF cryptococcal antigen titre, more frequentheadache, meningismus, papilloedema and hearing loss,and poorer short-term survival. US guidelines [9] thusrecommend measurement of baseline opening pressure inall patients with cryptococcal meningitis and repeatedlumbar drainage for those with baseline opening pressureof more than 25 cmH2O until normalization of pressure,with consideration of lumbar drain [10] or ventriculo-peritoneal shunt for refractory cases. Departure fromthese guidelines was neurologically detrimental in a smallstudy [11].
In three prospective studies of antifungal therapy forcryptococcal meningitis in Thailand and South Africa[1,3,12], serial lumbar punctures over the first 2 weeks oftreatment were done, with prospective recording ofopening pressure at each lumbar puncture. Patients withsignificantly raised pressure or attributable symptoms orboth received additional lumbar punctures. The aim of
incott Williams & Wilkins. Unauthor
this retrospective analysis was to establish the impact ofthis intervention on prognosis of patients with raisedopening pressure. Secondary aims were to explore thetime course of opening pressure and any associationwith neurological findings, CSF fungal burden, fungalclearance and immune response.
Methods
Study participantsThis cohort comprises 163 patients from Thailand andSouth Africa enrolled in three published studies [1,3,12].Studies were approved by ethics committees in Thailandand South Africa as well as in London, and participants, ornext of kin, gave written informed consent. Participantsreceived AmB-based therapy for 1–2 weeks, thereafterswitching to fluconazole, except for three patients inCape Town treated with fluconazole from the outset.AmB infusion reactions were not treated with corticos-teroids. Patients were followed up for up to 1 year. At thetime of the Thai study, ART was not routinely available,and no patients received ART before 10 weeks fromcryptococcal meningitis diagnosis. In South Africa, allART-naive patients were commenced on ART no earlierthan 4 weeks from cryptococcal meningitis diagnosis.
Given the association of cryptococcal immune recon-stitution syndrome with raised opening pressure [13],only patients who were ART-naive at study entry wereincluded in this analysis.
Management of raised cerebrospinal fluidpressureLumbar punctures were performed on days 1, 3, 7, and14 of treatment for quantitative CSF cultures, using22 G or 20 G needles. Additional therapeutic lumbarpunctures for patients with opening pressure of more than35 cmH2O and/or headache or other symptoms attribu-table to raised pressure were performed until openingpressure and symptoms were stable. At lumbar punc-ture, opening pressure was measured (Rocket spinal
ized reproduction of this article is prohibited.
C
CSF pressure in cryptococcal meningitis Bicanic et al. 703
manometers; Rocket Medical, Washington, England),and if raised (>20 cmH2O), sufficient CSF was drained,up to 30 ml, to achieve a closing pressure of 20 cmH2O orless, or less than 50% of opening pressure. Computedtomography (CT) or MRI brain imaging was notavailable at either centre during the studies and was notroutinely performed. No patients received steroids forraised opening pressure.
EvaluationBaseline clinical and laboratory parameters were recorded,as well as mortality at 2 and 10 weeks (Table 1). At lumbarpunctures, CSF opening pressure, closing pressure, andvolume removed were recorded. CSF was analysed for cellcount and differential, protein, glucose and cryptococcalantigen titre. Quantitative fungal culture, rate of clearanceand CSF cytokines were determined as previouslydescribed [12,14]. In the Thai study, CSF osmolalitywas determined.
Statistical analysisPatients were categorized into groups according toopening pressure (<20, 20–29.5, �30 cmH2O) atcryptococcal meningitis diagnosis, with a further groupcomprising those with no baseline opening pressuremeasurement. Log transformation was used for quanti-tative culture, cryptococcal antigen titres (base 2) and CSF
Values shown are median (IQR) or number of patients (%), except where indifungicidal activity; QCC, quantitative cryptococcal culture; WBC, white baDenominator shows number of patients in whom parameter was recordebAbnormal mental status indicates GCS less than 15.cOne patient was lost to follow-up by 2 weeks and three by 10 weeks.
cytokines. The groups with opening pressure measure-ments were compared using linear regression. Mann–Whitney, t-tests and x2(1 DF) tests were used for two-group comparisons between those with and withoutopening pressure measurement. Correlation was analysedusing Pearson’s test. Logistic regression and Coxregression were used to assess association of openingpressure with mortality at and survival to 2 and 10 weeks,adjusted for other prognostic indicators. Analysis wasperformed using Stata version 8 (Stata Corp., CollegeStation, Texas, USA).
Results
One hundred and eighty-one patients were enrolled inthree studies. Eighteen patients on ART at cryptococcalmeningitis diagnosis were excluded from further analysis.Of the remaining 163 patients, 138 had baseline openingpressure measured. Reasons for not measuring openingpressure were diagnostic lumbar puncture performedby emergency doctor and inability to perform lumbarpuncture in the lateral decubitus position. Median[interquartile range (IQR)] opening pressure was 23(15–32) cmH20. Eighty-seven (63%) had raised openingpressure of �20 cmH2O, and 26 (19%) had openingpressure of �35 cmH2O.
orized reproduction of this article is prohibited.
Fig. 1. (a) Scatter plot of baseline cryptococcal CSF CFUcount versus baseline opening pressure. (b) Relationshipbetween volume of CSF drained at lumbar puncture andthe difference between opening and closing CSF pressures,for 235 lumbar punctures. CFU, colony-forming units; CSF,cerebrospinal fluid.
Associations with baseline cerebrospinal fluidopening pressurePatients with higher baseline opening pressure had highercryptococcal CSF colony-forming units (CFU) counts(P¼ 0.02, Table 1, Fig. 1a) and were more likely tohave papilloedema (P¼ 0.001). There were no significantdifferences between pressure groups in baseline CSFparameters or cryptococcal antigen, fungal clearance(EFA), day 14 CSF CFU counts, CD4 cell count,presence of headache, or altered mental status. In the Thaistudy, there was no association between opening pressureand CSF osmolality (r¼�0.3, P¼ 0.15).
Overall mortality was 12% (20/162) at 2 weeks and 26%(42/160) at 10 weeks, with no significant differencesbetween opening pressure groups (Table 1). The oddsratios for death at 2 and 10 weeks for each increment inopening pressure group were 1.2 [95% confidenceinterval (CI) 0.7–2.3, P¼ 0.5] and 0.8 (95% CI 0.5–1.2, P¼ 0.2). There was still no association of openingpressure OP with mortality at, or survival over, 2 or10 weeks, when logistic and Cox regression were used toadjust for altered mental status and baseline quantitativeculture. The results were consistent whether opening
pressure was analysed in categories, as defined, or as acontinuous variable and whether three patients lost tofollow-up by 10 weeks were censored or classified asdeaths.
Twenty-five patients with no opening pressure did notdiffer significantly from patients with opening pres-sure measured in most baseline clinical and laboratoryparameters (Table 1) or outcome at 2 and 10 weeks.However, none in this group had altered mental status(P¼ 0.01) or cranial nerve palsies (P¼ 0.01), and theyreceived fewer lumbar punctures (mean 3.3 vs. 4,P¼ 0.02). To test the possibility of bias, we repeatedthe analyses, including those with no opening pressuremeasured in the lowest opening pressure group (0–19.5 cmH2O). There remained no significant differencein outcome at 2 and 10 weeks between opening pressuregroups (P¼ 0.3 and 0.5, respectively).
Volumes of cerebrospinal fluid drainedSix hundred and thirty-one lumbar punctures in totalwere performed in the first 14 days of treatment, a meanof 4 (range 1–8) lumbar punctures per patient. Data onopening pressure, closing pressure and volume of CSFdrained were complete for 235 lumbar punctures atwhich CSF was drained: mean opening pressure was31 cmH2O, mean closing pressure 13 cmH2O and meanvolume drained 14 ml. The relationship between volumeof CSF drained and fall in CSF pressure is shown in theFig. 1 (panel b).
Time course of opening pressureFor patients with opening pressure readings in the secondweek of treatment (n¼ 75), we analysed the time courseof opening pressure over the first 2 weeks. Median (IQR)opening pressure was 27 (16–36) cmH2O on day 1, 19(10–31) cmH2O on day 3, 19 (11–27) cmH2O on day 7and 21 (15–31) cmH2O on day 14.
Four patients had persistent symptoms and raised openingpressure requiring additional lumbar punctures beyondday 14, with successful control of pressure in threepatients. One patient required a temporary lumbar drainfor 5 days.
Day 14 cerebrospinal fluid opening pressureThe time course and our clinical experience suggestedthat high opening pressure might develop in the secondweek of treatment. We therefore analysed the associationbetween baseline and day 14 (d14) opening pressure; andbetween baseline and d14 fungal burden, rate of clearanceof infection, and d14 opening pressure, in the subset withd14 opening pressure recorded.
Interestingly, there was no significant correlation betweenbaseline and d14 opening pressure (P¼ 0.5). The mostsignificant associations were between d14 openingpressure and d14 CSF CFU count (r¼ 0.4, P< 0.001)
ized reproduction of this article is prohibited.
C
CSF pressure in cryptococcal meningitis Bicanic et al. 705
and between d14 opening pressure and baseline CSFcryptococcal antigen (r¼ 0.4, P< 0.001). Because d14fungal burden will be affected by rate of clearance ofinfection between d1 and d14, we hypothesized that d14opening pressure may be associated with rate of clearanceof infection. In fact, there was no significant correlationbetween rate of clearance of infection and d14 openingpressure (r¼ 0.2, P¼ 0.07), though the trend was forthose with more rapid clearance to have lower d14opening pressure. There was no significant relationshipbetween d14 opening pressure and mortality at 10 weeks(P¼ 0.5).
Discussion
Raised opening pressure was common in our cohort:two-thirds of the patients had opening pressure of atleast 20 cmH2O, and a fifth had opening pressure of atleast 35 cmH2O, comparable to three-quarters and one-quarter of patients, respectively, in the MSG trial [6]. Incontrast to the findings in the earlier trial, in this study,raised opening pressure did not impact adversely onoutcome at either 2 or 10 weeks.
This difference in findings may be explained by the factthat as per protocol, serial lumbar punctures wereroutinely performed, and additional therapeutic lumbarpunctures done for high opening pressure or symptomsthereof. At the time of the MSG trial, the significance ofraised opening pressure was less well established: only28 of 221 patients had repeated lumbar punctures toreduce pressure, and 41 received high-dose steroids.
Patients with raised opening pressure usually feel relieffrom headache within minutes of therapeutic lumbarpuncture. With appropriate monitoring of opening andclosing pressure, there were no adverse reactions recordedas a result of CSF drainage in any of the studies, despitelack of access to routine brain imaging. Drainage of anaverage of 14 ml CSF produced an average 18 cm reduc-tion in CSF pressure. However, occasionally, drainage ofmodest volumes produced large falls in pressure. Wewould suggest rechecking pressure after removal of every10 ml of CSF. Temporary lumbar drains may occasionallybe needed and can be used, with care, in resource-limitedsettings [7,15].
Headache was common and there were no neurologicalfindings, apart from papilloedema, that discriminatedbetween those with normal and raised pressure, under-lining the need to measure opening pressure in allpatients. Graybill et al. [6] found that patients with CSFopening pressure of greater or equal to 25 cmH2O hadhigher CSF cryptococcal antigen titres. We found a sig-nificant association between baseline opening pressureand baseline quantitative cryptococcal culture. The scatter
plot pattern (Fig. 1a) suggests that a high organism load isnecessary but not always sufficient for the development ofhigh opening pressure. There may be additional factors,such as phenotype of infecting Cryptococcus neoformansstrain, that contribute to development of raised pressure incryptococcal meningitis, as demonstrated in a rat model[16].
The association between fungal burden and d14 openingpressure suggests rapid clearance of infection may helpprevent development of raised pressure in the 2nd week.The lack of correlation between baseline and 2-weekopening pressure, and the suggestion of a biphasic timecourse for opening pressure, raises the possibility that thepathophysiological basis of raised pressure may not beprecisely the same at baseline and 2 weeks.
In conclusion, in a cohort of patients undergoing frequentrepeat lumbar punctures, raised opening pressure wascommon yet not associated with increased mortality. Theresults support the efficacy of repeat lumbar punctures inmanagement of patients with raised opening pressure.Efforts are needed to improve access to manometersin resource-limited settings and to implement routinemeasurement and management of pressure in patientswith cryptococcal meningitis.
Acknowledgements
This work was presented at the 7th International Con-ference on Cryptococcus & Cryptococcosis, Nagasaki,Japan, 11–14 Sept 2008, abstract no. 10068.
This article was sponsored by St George’s HospitalTrustees, British Infection Society (fellowship grant toT.B.), The Lancet (international fellowship to A.B.),Wellcome Trust (training fellowship to A.E.B.), MedicalResearch Council, UK. The Thailand study was part ofthe Wellcome Trust-Mahidol University-Oxford Uni-versity Tropical Medicine Research Programme, fundedby the Wellcome Trust of Great Britain.
We thank Supraphada Pinpraphaporn, Bina Maharjan,Anna Checkley, Vanaporn Wuthiekanun, PremjitAmornchai, Nongluk Getchalarat, Pissamai Manupan,Jintana Suwanpruek, Nick Day, Sharon Peacock, for helpwith the study in Sappasithiprasong Hospital. We thankNomqondiso Sidibana, Tom Crede, Vanessa Burch,Anthony Williams, Noxolo Mahlaza, Elma de Vriesfor help with the studies at GF Jooste hospital.
Contributors: T.B. and A.E.B. enrolled all the patients inthe study and collected the data. T.B. analysed the dataand wrote the article. A.E.B. helped write the article.G.M., K.R., D.L., W.C. and P.T. provided clinical
orized reproduction of this article is prohibited.
Co
706 AIDS 2009, Vol 23 No 6
supervision and support for the studies in Cape Town andThailand and helped write the paper. A.L. performedclinical follow-up on patients from Cape Town. N.J.W.and R.W. were involved in the studies’ design and helpedwrite the article. S.J. helped with the statistical analysis.T.H. designed the studies, was principal investigator andwrote the article.
There are no conflicts of interests.
References
1. Bicanic T, Meintjes G, Wood R, Hayes M, Rebe K, Bekker LG,et al. Fungal burden, early fungicidal activity, and outcome incryptococcal meningitis in antiretroviral-naive or antiretrovir-al-experienced patients treated with amphotericin B or fluco-nazole. Clin Infect Dis 2007; 45:76–80.
2. Lortholary O, Poizat G, Zeller V, Neuville S, Boibieux A,Alvarez M, et al. Long-term outcome of AIDS-associated cryp-tococcosis in the era of combination antiretroviral therapy.AIDS 2006; 20:2183–2191.
3. Bicanic T, Wood R, Meintjes G, Rebe K, Brouwer A, Loyse A,et al. High-dose amphotericin B with flucytosine for the treat-ment of cryptococcal meningitis in HIV-infected patients: arandomized trial. Clin Infect Dis 2008; 47:123–130.
4. Denning DW, Armstrong RW, Lewis BH, Stevens DA. Elevatedcerebrospinal fluid pressures in patients with cryptococcalmeningitis and acquired immunodeficiency syndrome. Am JMed 1991; 91:267–272.
5. Hussey F, Schanzer B, Katzman R. A simple constant-infusionmanometric test for measurement of CSF absorption. II: Clin-ical studies. Neurology 1970; 20:665–680.
6. Graybill JR, Sobel J, Saag M, van Der Horst C, Powderly W,Cloud G, et al. Diagnosis and management of increased in-tracranial pressure in patients with AIDS and cryptococcalmeningitis. The NIAID Mycoses Study Group and AIDS Co-operative Treatment Groups. Clin Infect Dis 2000; 30:47–54.
7. Macsween KF, Bicanic T, Brouwer AE, Marsh H, Macallan DC,Harrison TS. Lumbar drainage for control of raised cerebrosp-inal fluid pressure in cryptococcal meningitis: case report andreview. J Infect 2005; 51:e221–e224.
8. Rex JH, Larsen RA, Dismukes WE, Cloud GA, Bennett JE.Catastrophic visual loss due to Cryptococcus neoformans me-ningitis. Medicine (Baltimore) 1993; 72:207–224.
9. Saag MS, Graybill RJ, Larsen RA, Pappas PG, Perfect JR, Pow-derly WG, et al. Practice guidelines for the management ofcryptococcal disease. Infectious Diseases Society of America.Clin Infect Dis 2000; 30:710–718.
11. Shoham S, Cover C, Donegan N, Fulnecky E, Kumar P. Cryp-tococcus neoformans meningitis at 2 hospitals in Washington,D.C.: adherence of healthcare providers to published practiceguidelines for the management of cryptococcal disease. ClinInfect Dis 2005; 40:477–479.
12. Brouwer AE, Rajanuwong A, Chierakul W, Griffin GE, LarsenRA, White NJ, et al. Combination antifungal therapies for HIV-associated cryptococcal meningitis: a randomised trial. Lancet2004; 363:1764–1767.
13. Shelburne SA III, Darcourt J, White AC Jr, Greenberg SB, HamillRJ, Atmar RL, et al. The role of immune reconstitution inflam-matory syndrome in AIDS-related Cryptococcus neoformansdisease in the era of highly active antiretroviral therapy. ClinInfect Dis 2005; 40:1049–1052.
14. Siddiqui A, Brouwer AE, Suprapada P, Wuthiekanun V, Jaffar S,Shattock R, et al. Interferon-gamma at the site of infectiondetermines rate of clearance of infection in cryptococcalmeningitis. J Immunol 2005; 174:1746–1750.
15. Manosuthi W, Sungkanuparph S, Chottanapund S, Tansuphas-wadikul S, Chimsuntorn S, Limpanadusadee P, et al. Temporaryexternal lumbar drainage for reducing elevated intracranialpressure in HIV-infected patients with cryptococcal meningi-tis. Int J Std AIDS 2008; 19:268–271.
16. Fries BC, Lee SC, Kennan R, Zhao W, Casadevall A, GoldmanDL. Phenotypic switching of Cryptococcus neoformans canproduce variants that elicit increased intracranial pressure in arat model of cryptococcal meningoencephalitis. Infect Immun2005; 73:1779–1787.
