|Home | About | Journals | Submit | Contact Us | Français|
Cryptococcal meningitis is a leading cause of death in AIDS patients and contributes substantially to the high early mortality in antiretroviral treatment (ART) programs in low-resource settings. Screening for cryptococcal antigen (CRAG) in patients enrolling in ART programs may identify those at risk of cryptococcal meningitis and permit targeted use of pre-emptive therapy.
In this retrospective study, CRAG was measured in stored plasma samples obtained from patients as they enrolled in a well characterised ART cohort in South Africa. The predictive value of screening for CRAG prior to ART for development of microbiologically confirmed cryptococcal meningitis or death during the first year of follow-up was determined.
Of 707 participants with a baseline median CD4 count of 97 (IQR 46-157) cells/μL, 46 (7%) had a positive CRAG. Antigenaemia was 100% sensitive for predicting development of cryptococcal meningitis during the first year of ART and in multivariate analysis was an independent predictor of mortality (AHR 3.2, 95%CI 1.5-6.6). Most (92%) cases of cryptococcal meningitis developed in patients with a CD4 count ≤100 cells/μL. In this sub-set of patients, a CRAG titre ≥1 in 8 was 100% sensitive and 96% specific for predicting incident cryptococcal meningitis during the first year of ART in those with no previous history of the disease.
CRAG screening prior to commencing ART in patients with a CD4 count ≤100 cells/μL is highly effective at identifying those at risk of cryptococcal meningitis and death and might permit implementation of a targeted pre-emptive treatment strategy.
Cryptococcus neoformans is a major opportunistic pathogen and a leading cause of mortality in AIDS patients in much of the developing world [1-3]. It is the commonest cause of meningitis in central and Southern Africa [4-6], accounting for 40% of all cases in a recent study from Malawi . Introduction antiretroviral therapy (ART) has been associated with a decline in the incidence of cryptococcal meningitis across the developed world . However, in many low resource settings, most patients continue to present late to ART treatment programs, with low median CD4 counts, high risk of new AIDS events and high early mortality. Between 8 and 26% of patients die in the first year of ART, with most deaths occurring in the first few months . Cryptococcal meningitis is a leading contributor to this early mortality, accounting for up to 20% of all deaths [10-13]. This constitutes a heavy burden on healthcare facilities and accounted for 31% of all inpatient days in a study from South Africa . Treatment of cases presenting both prior to and after initiation of ART remains inadequate, with an acute mortality in unselected series of between 20 and 50%, even with the best current treatment [15-20].
Prevention of severe disease by routine screening for sub-clinical infection using cryptococcal antigen (CRAG) tests and administering pre-emptive treatment would therefore be an attractive strategy. The current cryptococcal antigen tests are highly sensitive and specific [21, 22] and are validated for use in patients with late stage HIV infection . Importantly in patients with cryptococcal disease, antigenaemia is detectable a median of 22 days before symptom onset and >100 days before symptom onset in 11% of patients .
The clinical course of patients with asymptomatic antigenaemia remains poorly defined since all such patients who are identified typically receive treatment. However, if left untreated, clinical disease may develop [24, 25] as fungal burden increases in the context of persisting immunodeficiency. Alternatively, if ART is commenced, rapid restoration of pathogen-specific immune responses may cause “unmasking” of sub-clinical disease. Finally it is possible that restoration of cell-mediated immunity during ART could lead to clearance of asymptomatic infection. To investigate this and to evaluate the potential utility of screening for CRAG prior to ART initiation, we conducted a retrospective study in a well-characterized ART cohort in Cape Town, South Africa. Using a large collection of plasma samples obtained from patients just prior to starting ART, we were able to determine the subsequent clinical course of patients with cryptococcal antigenaemia. We used this information to test the hypothesis that cryptococcal antigen screening prior to initiation of ART can be used to identify those individuals at risk of developing clinical cryptococcal disease.
The study was approved by the Research Ethics Committee of the University of Cape Town. All patients provided written informed consent to sample collection and research use. Samples were available from sequential patients entering the ART service based at a Cape Town community clinic from September 2002 to April 2005. The clinic serves a population of over 300,000 in which the ante-natal HIV seroprevalence was 28% in 2003. Patients from local primary care clinics are referred into the program and commenced on treatment based on World Health Organisation (WHO) 2002 recommendations (those with a prior AIDS diagnosis i.e. WHO stage 4 disease, or a blood CD4 count <200 cells/μl).
A mean delay of approximately one month from patient enrollment to initiation of ART permitted evaluation and preparation for treatment. Blood samples were drawn at a routine clinic visit during this period to measure CD4 cell counts (FACSCount™, Beckton Dickinson, Franklin Lakes, NJ, USA) and plasma viral load (Versant™, Bayer Healthcare, Leverkusen, Germany). Excess EDTA plasma was stored at −80°C. For a limited time period during recruitment of the cohort further blood samples were collected 16 weeks after initiation of ART for CD4 cell counts and plasma viral load testing as described above. Thus excess EDTA plasma was available at 16 weeks for a consecutive subset of patients.
ART consisting of stavudine, lamivudine plus efavirenz or nevirapine was supplied free of charge. Treatment adherence was good with rates of viral load suppression <400 copies/ml in the cohort both exceeding 90% at the 16-week follow up time point . Clinic visits were scheduled at 4, 8 and 16 weeks, and then 16-weekly, with additional open access appointments available. Principal causes of death in the cohort have been described in detail . Structured clinical records for all patients were prospectively maintained, and this information transferred weekly into a computer database.
Cryptococcal antigen (CRAG) testing was carried out using the Meridian Cryptococcal Latex Agglutination System™ (Meridian Bioscience Europe), a simple latex test capable of detecting the capsular polysaccharide of Cryptococcus neoformans in cerebrospinous fluid and blood. Plasma samples stored at the time of study enrollment (prior to initiation of ART), and after 16 weeks of ART in a subset of patients, were tested. The assay was carefully validated for use on EDTA plasma by running paired analysis of serum and EDTA plasma samples from both known positive and control patients. Concordance in antigen titre between serum samples and EDTA plasma samples was 100% across a wide range of dilutions. Samples were first incubated with pronase at 56°C for 15 minutes to minimize false positive results, then analysed according to the manufacturers instructions. All samples were analysed at a 1 in 2 dilution, then positive samples titred down to dilutions of 1 in 8192.
The main outcomes were detection of cryptococcal antigenaemia (defined as positive at a dilution of 1 in 2 or greater) at baseline and microbiologically confirmed cases of cryptococcal meningitis during the first year of follow-up. In patients with a history of cryptococcal disease prior to ART, symptomatic relapse after starting ART was defined as recurrence of typical symptoms, CSF antigen or culture positivity for C. neoformans, and no alternative diagnosis. Mortality data and ART treatment response were recorded as additional outcome measures. Information regarding causes of death of patients in the ART program was obtained from the local secondary and tertiary care hospitals, hospital mortality review meetings and postmortem examinations. The most likely attributable cause of death of each patient was assigned based on all available information after detailed review by two specialists in infectious diseases and HIV medicine .
Data were analysed using STATA version 9.0 (Stata statistical software, Stata Corporation, College Station, Texas, USA). Variables were compared across groups using Student's t-test, the Mann-Whitney U-test, the χ2 or Fisher's exact tests, and Cuzick's nonparametric test for trend as appropriate. Statistical significance was defined as a p-value <0.05. Hazard ratios (HR) with 95% confidence intervals (CI) were calculated using Cox proportional hazard modeling.
A total of 707 patients were eligible for inclusion in the study. The mean age of patients was 33 years and 26% were male. The median CD4 count of the cohort was 97 (IQR 46-157) cells/μL and the median viral load 76,803 (IQR 33,167-191,030) copies/ml. 8% of patients had WHO stage I disease, 12% stage II, 52% stage III and 28% stage IV. There was no clinical suspicion of meningitis in any of the patients at the time of blood sampling.
At baseline 46 patients (7%) had a positive CRAG assay. Patients who were antigen positive had lower baseline CD4 cell counts, were more likely to have a history of previous cryptococcal disease and were more likely to develop incident cryptococcal meningitis during ART (table 1).
Of those with complete follow-up data, 34% (n=14) of the CRAG-positive patients died in the first year of ART compared to 11% (n=64) of antigen negative patients (p<0.001). In a sensitivity analysis where all those lost to follow-up were assumed to have died, the relationship between antigen screening status and death remained highly significant. Causes of death are listed in table 1.
CRAG-positive patients were at far higher risk of mortality than antigen negative patients during the one year follow-up period (HR = 4.75, 95% CI 2.6-8.8, p<0.001). After adjustment for CD4 cell count, viral load, age and sex, baseline cryptococcal antigenaemia remained a strong independent risk factor for death (adjusted HR = 3.2; 95% CI 1.5-6.6, p<0.001). This relationship was also found when the analysis was restricted to patients with no prior history of cryptococcal disease (adjusted HR = 3.1; 95% CI 1.04-9.15, p<0.001).
Mortality risk was strongly associated with antigen titre. Of those with CRAG titres of ≤1 in 8, 1 in 16–64, 1 in 128–512, 1 in 1024–2048 or ≥1 in 4096, deaths occurred in 0% (n=0), 25% (n=2), 31% (n=4), 44% (n=4) and 67% (n=4), respectively (p trend 0.02). Mortality was 19% and 53% in those with an antigen titre of ≤1 in 512 and >1 in 512 respectively (p=0.04).
Among patients with a negative CRAG screen (n=661), no cases of cryptococcal meningitis developed during the 1-year period of follow-up. In contrast, among patients with a positive CRAG screen (n=46), 13 (28%) developed new or relapsed clinically apparent cryptococcal meningitis during follow-up.
A previous history of cryptococcal meningitis was recorded among 21 (46%) of the CRAG-positive patients, with disease onset occurring a median of 140 (IQR 99-230) days prior to enrolment in the ART program. Six of these 21 patients (29%) developed symptomatic relapse of their cryptococcal disease a median of 33 days (Range 5-70) after starting ART. All had been prescribed fluconazole maintenance therapy, and CSF cultures were negative in all cases. All these patients had baseline antigen titres of ≥1 in 512 (Figure 1A). However, there were no significant relationships observed between risk of relapse and time interval between prior cryptococcal disease and ART initiation, baseline CD4 cell count and viral load or immunological and virological response to ART at 16 weeks (data not shown).
Of CRAG-positive patients who did not have a previous history of cryptococcal meningitis (n=25), 7 (28%) developed clinically apparent cryptococcal meningitis a median of 35 days (Range 8-188) after starting ART. In these patients, risk of developing cryptococcal meningitis was significantly associated with CRAG titre. The proportions of patients with antigen titres of ≤1 in 8, 1 in 16-64, 1 in 128-512, 1 in 1024-2048, and ≥1 in 4096 who developed symptomatic cryptococcal disease during follow-up were 12.5% (n=1), 17% (n=1), 29% (n=2), 67% (n=2), and 100% (n=1), respectively (p trend 0.03, Figure 1B).
If CRAG screening were to be introduced into routine care in ART programs, those with a prior history of cryptococcal meningitis would be excluded on the assumption that they were receiving consolidation treatment or adequate secondary prophylaxis. Thus of the 683 patients who had no history of cryptococcal disease, 7 cases could have been detected and potentially averted through use of CRAG screening (i.e. 98 patients screened to identify 1 case).
Nine CRAG-positive patients with no history of cryptococcal meningitis and who did not develop clinically apparent cryptococcal disease during follow-up had a second sample after 16-weeks of ART available for antigen testing. Of these, the antigen titre decreased in 7 (78%), and increased in two (from 1 in 4 to 1 in 8, and 1 in 32 to 1 in 128). The median change from baseline was a fall of 1 dilution. Eight patients with a history of cryptococcal meningitis prior to starting ART who did not relapse (all of whom were on fluconazole maintenance treatment) had paired baseline and week 16 samples available for testing. Titres fell in 7 (88%) and remained unchanged at a titre of 1 in 16 in one patient. The median change from baseline was a fall of 2 dilutions.
When analysis was restricted to patients with a baseline CD4 cell count ≤100 (n=336), baseline characteristics of patients (including CD4 cell counts) were similar in sub-groups of patients with and without antigenaemia (Table 2). 42 (13%) patients had a positive CRAG screen. 33% of CRAG-positive patients died within the first year of treatment, compared to only 13% of CRAG-negative patients (p = 0.002).
Half (n=21) of the CRAG-positive patients did not have a history of cryptococcal meningitis. Of these 29% (n=6) developed cryptococcal meningitis and these may potentially have been averted by screening and appropriate treatment. Thus, 6 cases would have been detected by screening the total of 312 patients who did not have a history of cryptococcal meningitis (i.e. 52 patients screened to identify 1 case).
Using a cut-off titre of 1 in 8, CRAG screening had 100% sensitivity and 96% specificity for predicting incident cryptococcal disease in patients with baseline CD4 cell counts ≤100 cells/μL and no previous history of the disease. Thus, the positive predictive value (PPV) was 35% and the negative predictive value (NPV) was 100% in this sub-group of patients. As would be expected, the sensitivity decreased with increasing antigen titre cut-offs, while specificity increased (Table 3).
We found a high prevalence of cryptococcal antigenaemia in patients enrolling in this ART service in South Africa. Thirteen percent of those with a CD4 count of less than or equal to 100 cells/μL were antigen positive at baseline, consistent with other studies from sub-Saharan Africa and South East Asia, where between 10% and 18% of HIV-infected first time clinic attendees have a positive serum CRAG on routine screening   . Antigenaemia was not only highly predictive of the development of cryptococcal meningitis but was also an independent predictor of mortality. In addition, both of these outcomes were strongly associated with higher CRAG titres. These data suggest that screening for CRAG may be useful in this patient group.
The high prevalence of cryptococcal antigenaemia in those patients without a prior history of cryptococcal disease suggests routine antigen screening may have a role in this setting [1, 23]. Evaluation of the possible benefits of such an intervention requires an understanding of the clinical course of patients with antigenaemia. Cryptococcal antigenaemia in the context of advanced HIV infection has been assumed to indicate disseminated disease with progression to severe symptomatic cryptococcosis inevitable unless appropriate anti-fungal treatment is given [24, 25]. However, the clinical course of untreated antigenaemia in patients commencing ART has not previously been described. Liechty et al. in Uganda reported that asymptomatic cryptococcal antigenaemia in 377 patients commencing ART independently predicted death during the first 12 weeks of treatment . The relative risk of mortality was 6.6 after controlling for CD4 count, viral load and other adverse prognostic markers. However, data on the development of incident cryptococcal disease or the causes of death were not reported .
Stored plasma samples from a large patient cohort with comprehensive follow-up data gave us a unique opportunity to investigate the clinical course of untreated cryptococcal antigenaemia in HIV-infected patients commencing ART. Of the patients with no history of cryptococcal meningitis who were CRAG-positive at baseline, 28% developed cryptococcal meningitis, 8% died of causes other than cryptococcosis and 8% were lost to follow up leaving 56% who were alive and had not developed cryptococcal disease at one year. We speculate that in many cases immune reconstitution during ART resulted in effective clearance of asymptomatic infection. This suggestion is supported by the observation that 78% of the antigen positive patients who did not develop disease had a fall in antigen titre, similar in magnitude to that seen with effective antifungal treatment   . Immune-mediated clearance however, becomes less likely as antigen titre increases. In addition, high antigen titres (≥ 1 in 512) in those with a prior history of cryptococcal disease were predictive of symptomatic relapse after starting ART. In these patients we suspect that clinical presentation was due to immune reconstitution disease .
In terms of the potential utility of CRAG screening in patients commencing ART, the majority of patients (91%) with positive cryptococcal antigenaemia had a CD4 ≤ 100 cells/μL, a finding replicated in Cambodia , and all but one case of incident cryptococcal disease (92%) occurred in this sub-group of patients. Limiting analysis to patients with a CD4 ≤ 100 cells/μL, and using an antigen titres of ≥1:8 to minimise false positive results, antigen screening was highly effective at identifying those at risk of developing cryptococcal disease. The number needed to screen to identify and potentially prevent one case of cryptococcal meningitis was 52 patients. At an approximate cost of $3.97 per test (CDC estimate ) the cost per case identified was $206.44.
Routine primary fluconazole prophylaxis in both developed and developing world settings has been shown to reduce the number of cases of cryptococcal meningitis, yet it has not shown a consistent survival benefit . Large numbers of patients require long-term medication and concerns exist around the development of drug resistance . In Thailand, national guidelines advocate fluconazole for all patients with a CD4 count below 100 cells/μL. Although such prophylaxis has not been shown to select for fluconazole resistance in cryptococcal isolates from those who subsequently developed cryptococcal disease , recent work has nevertheless demonstrated a significant association with development of fluconazole resistant candida .
Antigen screening in patients with a CD4 count ≤ 100 cells/μL could allow a targeted pre-emptive treatment strategy, reducing costs and drug resistance. However, important questions remain to be answered. What should be done about a positive CRAG result in an asymptomatic patient? Many experts would recommend lumbar puncture to rule out CNS involvement, and amphotericin B-based therapy if CNS involvement is found. However, this would entail lumbar punctures in a substantial number of patients prior to starting ART, placing an additional workload on overstretched ARV programs, and may not be acceptable to all asymptomatic patients. There is limited evidence that, for patients with isolated cryptococcal antigenaemia, fluconazole alone is sufficient to prevent clinical disease  . It is possible that below a certain antigen titre daily fluconazole treatment may be adequate in asymptomatic patients, avoiding the need for lumbar puncture. Of note, in a primary prophylaxis strategy, as opposed to pre-emptive treatment based on antigen screening, these patients would receive intermittent fluconazole. Prospective studies are urgently required to address these issues and test the benefits of antigen-based screening.
A further potential limitation of our study was that some of those lost to follow up might in fact have died without detection by the ART program. However non-death losses in this cohort have previously been characterized in detail and were not found to be associated with baseline immunodeficiency and were unlikely to represent unascertained deaths . Overall loss to follow-up and late mortality rates were low, reflecting excellent treatment response and cohort retention.
In conclusion, CRAG screening prior to commencing ART in patients with a CD4 count ≤100 cells/μL is highly effective for identification of patients at risk of cryptococcal meningitis and death and might permit implementation of a targeted pre-emptive treatment strategy.
JNJ and SDL are supported by the Wellcome Trust, London, UK.
RW is funded in part by the National Institutes of Health, USA, through a CIPRA grant 1U19AI53217-01 and RO1 grant (A1058736-01A1).