Utility of Cryptococcal Antigen Screening and Evolution of Asymptomatic Cryptococcal Antigenemia among HIV-Infected Women Starting Antiretroviral Therapy in Thailand

Background

Cryptococcal meningitis (CM) continues to be an important HIV-associated opportunistic infection, with an estimated 957 900 cases annually worldwide resulting in 624 700 deaths. ¹ HIV-associated CM has a high mortality rate in resource-limited settings, despite the availability of antiretroviral therapy (ART). ^{2 –4} A prospective, observational study in Uganda found that, despite access to ART, the 6-month survival rate among HIV-infected patients diagnosed with CM was 41%. ³ Diagnosing HIV infection and initiating ART at an early stage (ie, before CD4 count [CD4] declines below 350 cells/mm³) is the most effective intervention to prevent cryptococcal disease in HIV-infected persons. ⁵ In persons with CD4 <100 cells/mm³, ART initiation can precipitate CM-immune reconstitution inflammatory syndrome (IRIS) by either unmasking previously undiagnosed infection or exacerbating partially treated cryptococcal disease. ⁶ To identify persons at high risk of developing cryptococcal disease and to prevent the development of CM and CM-associated deaths before and after ART initiation, the World Health Organization (WHO) recommends considering a “screen and treat” strategy in settings with a high prevalence (eg, ≥3%) of cryptococcal antigenemia, whereby ART-naive adults with CD4 <100 cells/mm³ undergo serum testing for cryptococcal antigen (CrAg) and receive preemptive antifungal therapy if CrAg positive. ⁵ Using data from an observational cohort study of ART-naive HIV-infected Thai women initiating ART, we assessed the utility of CrAg screening to identify the risk for developing CM and examined the effect of ART on asymptomatic cryptococcal antigenemia.

Methods

The patients studied were part of the multisite observational “Non-Nucleoside Reverse Transcriptase Inhibitor (NNRTI) Response Study” conducted from 2005 to 2007 in Thailand, Kenya, and Zambia. ⁷ The study protocol was approved prior to study initiation by the institutional review boards at the Thai Ministry of Public Health, the Mahidol University, Siriraj Hospital, and the US Centers for Disease Control and Prevention; all participants provided written informed consent. Eligible participants were ART-naive, nonpregnant women initiating NNRTI-based ART due to advanced HIV disease (in Thailand, CD4 <200 cells/mm³ or clinical AIDS). In Thailand, 217 women were enrolled in the NNRTI Response Study and were followed for 48 weeks.

All participants initiated ART in an outpatient setting approximately 1 to 2 weeks after a baseline visit. All patients had periodic clinical assessments and blood drawn at baseline and at weeks 2, 4, 8, 16, 24, and 48 while prescribed ART. Clinical variables collected included symptoms and diagnoses used in the WHO HIV Clinical Staging system, including cryptococcal disease and receipt of treatment for opportunistic infections. In accordance with Thailand National Antiretroviral Treatment Guidelines, fluconazole primary prophylaxis of 400 mg taken orally weekly was prescribed and continued at physicians’ discretion for women with CD4 <100 cells/mm³. ^{8 –10} Because CrAg testing was not available at the time of study enrollment, primary prophylaxis was defined as the use of fluconazole to prevent fungal disease in HIV-infected women without a prior history of CM and whose CrAg status was unknown.

We performed retrospective CrAg testing of stored blood plasma from 211 Thai enrollees who had consented to additional testing of their specimens; 2 women had a history of CM at baseline and were excluded from the analysis. Of the 209 women included in the analysis, 8 were lost to follow-up at week 24, and a total of 12 were lost to follow-up at week 48. We used a latex agglutination CrAg assay (IMMY, Norman, Oklahoma), validated for use with ethylenediaminetetraacetic acid plasma in our laboratory, to retrospectively test participants’ baseline plasma specimens. For women who were antigenemic (reactive titer > 1:1) at baseline, we additionally assayed CrAg titers from all available serial plasma specimens collected during follow-up visits. For all women with CD4 <100 cells/mm³ whose baseline plasma specimen was CrAg negative, we assessed for incident antigenemia by testing plasma specimens collected at weeks 24 and 48 after initiating ART. Study end points included CrAg titers over time and the development of CM during the follow-up period. Data were analyzed using Statistical Package for the Social Sciences software, version 18.0 (SPSS Inc, Chicago, Illinois). Chi-square and Fisher exact tests were used to assess significance of statistical associations.

Results

Of the 209 women included in this analysis, median age was 31 years (interquartile range [IQR] 28-37), median CD4 was 147 cells/mm³ (IQR 51-220), and 74 (36%) had WHO stage III or IV disease. Of the 125 (60%) women with CD4 ≥100 cells/mm³, none (0%, 95% confidence interval [CI]: 0%-3%) were antigenemic at baseline, none developed CM during 48 weeks of ART, and all were alive at the end of the study period. Of the 84 (40%) women with CD4 <100 cells/mm³, the median CD4 was 40 cells/mm³ (IQR 15-69), and 9 women (11%, 95% CI: 5%-19%) were antigenemic at baseline. There was no difference in the baseline prevalence of antigenemia among the 35 women prescribed fluconazole primary prophylaxis (n = 4, 11%) compared with the 49 women not prescribed fluconazole (n = 5, 10%, P > .9).

Of the 9 women who were antigenemic, the median baseline CD4 was 36 cells/mm³ (IQR 14-79), 8 (89%) achieved plasma HIV RNA <100 copies/mL and immune recovery (CD4 ≥ 100 cells/mm³) by week 24 of ART, and 8 (89%) had CrAg titers decline by ≥2-fold or declined to nonreactive by week 48 (Figure 1). Among women with CD4 <100 cells/mm³ at baseline, 2 (22%) of the 9 women with antigenemia developed CM (Figure 1), which was diagnosed by India ink stain of cerebral spinal fluid, compared with none of the 75 women without antigenemia (P < .05). One case of CM occurred during fluconazole primary prophylaxis at week 5 after initiating ART. The other case occurred at week 42 despite primary fluconazole prophylaxis, immune recovery (CD4 increased from 56 cells/mm³ at baseline to 181 cells/mm³ at week 24), and viral suppression (plasma HIV RNA <50 copies/mL) at 24 and 48 weeks. Both women who developed CM were treated successfully at a tertiary care hospital in Bangkok, where therapeutic lumbar punctures and amphotericin B were the standard of care. All 9 women with baseline antigenemia were alive at 48 weeks.

OPEN IN VIEWER

Figure 1.

Evolution of cryptococcal antigen titers after ART initiation among women with baseline antigenemia. A, Titers of women who did not develop cryptococcal meningitis (n = 7). B, Titers of women who developed cryptococcal meningitis (n = 2). Solid line, fluconazole primary prophylaxis (400 mg weekly) prescribed. Dotted line, fluconazole primary prophylaxis was not prescribed.

Of the 75 women who at baseline had CD4 <100 cells/mm³ and were not antigenemic, 1 woman developed incident antigenemia, none developed CM, and 4 died. The woman who developed incident antigenemia was not receiving primary fluconazole prophylaxis and had a titer of 1:2 at week 24. She did not develop CM during the study, and her titer declined to nonreactive at week 48. The 4 women who died were not antigenemic at the last study visit prior to death, and there was no evidence that the deaths were related to cryptococcal disease.

Discussion

Our study confirmed findings from previous studies in Thailand and South Africa and provided novel data regarding the course of cryptococcal antigenemia in patients initiating ART and the poor efficacy of low-dose fluconazole prophylaxis in preventing CM among patients with antigenemia.

Our study detected a high prevalence of cryptococcal antigenemia (11%, 95% CI: 5%-19%) among Thai women initiating ART with CD4 <100 cells/mm³, consistent with previous studies from Thailand. ¹¹ Consistent with previous studies from South Africa, ¹² we found that CrAg screening in this group was 100% sensitive in identifying women at risk for disease, with a 22% positive predictive value of baseline antigenemia for development of CM. These results suggest that routine CrAg screening of HIV-infected adults with CD4 <100 cells/mm³ who are initiating ART can be an important method to reduce CM-related morbidity and mortality in Thailand, by identifying candidates for preemptive treatment with high-dose fluconazole. ⁵ With the advent of the cryptococcal lateral flow assays (LFAs), which, compared with cryptococcal latex-agglutination assays, require less training and minimal laboratory equipment to perform and are significantly cheaper, ^13,14 routine CrAg screening will likely become a more effective strategy and yield greater cost savings in preventing CM and cryptococcal-related deaths.

The numbers of women who developed incident cryptococcal antigenemia or CM were low in our study, and we did not find an association between fluconazole primary prophylaxis and the prevention of either cryptococcal antigenemia or CM. Fluconazole prophylaxis was not associated with a lower incidence of CM; of the 2 women with baseline antigenemia who developed CM, both had been prescribed fluconazole primary prophylaxis in accordance with Thai HIV Treatment Guidelines. This finding supports recent WHO guidelines that do not recommend primary prophylaxis when prompt ART initiation is likely. ⁵ For asymptomatic cryptococcal antigenemia, the optimal antifungal regimen remains an area for further research, although the WHO recommends preemptive treatment with high-dose fluconazole at 800 mg daily for 2 weeks, then 400 mg daily for 8 weeks, followed by 200 mg for daily maintenance treatment. ⁵ We estimate that the cost of treating the 9 women with asymptomatic cryptococcal antigenemia in our study with the WHO-recommended preemptive treatment regimen would have been less than half of the cost of providing fluconazole primary prophylaxis for 1 year to all 84 women with CD4 <100 cells/mm³ (data not shown). The use of the newer LFA CrAg point-of-care test, which is estimated to cost approximately US$2 per test, would further reduce the cost of the WHO recommended “screen and treat” strategy.

Although there was no control group in our study, our results support the importance of early ART for the prevention of cryptococcal disease and suggest that initiating ART may reduce the risk for incident CM in persons with CD4 <100 cells/mm³ and asymptomatic cryptococcal antigenemia. Among women initiating ART at CD4 ≥100 cells/mm³, cryptococcal antigenemia was not present and CM did not develop; this result suggests that CM can be prevented by starting ART before CD4 declines to <100 cells/mm³. Among women with CD4 <100 cells/mm³ who initiated ART without cryptococcal antigenemia at baseline, only 1 woman developed incident antigenemia, and none developed cryptococcal disease. In the small number of women with asymptomatic cryptococcal antigenemia and CD4 <100 cells/mm³, CrAg titers decreased over time with ART, regardless of fluconazole prophylaxis use. Among 78% (7 of 9) of women with baseline antigenemia, ART resulted in a significant decline in CrAg titers without evidence of CM or CM-IRIS, of which 71% (5 of 7) were not on primary prophylaxis. This finding supports the importance of ART in preventing CM and is consistent with other reports, which have described that 60% to 75% of persons initiating ART with untreated cryptococcal antigenemia did not develop CM. ^11,12,15

The optimal timing of ART initiation among HIV-infected persons with asymptomatic antigenemia has yet to be determined. Of the 9 women with baseline antigenemia, only 1 developed CM-IRIS at week 5 after initiating ART. It is not known whether a short delay in ART combined with preemptive treatment with high-dose fluconazole would have prevented the case of CM-IRIS. Similarly, among persons with active CM, timing of ART initiation remains an area of ongoing research, with a recent randomized trial stopped early due to apparent decreased survival among persons receiving immediate ART following CM diagnosis compared with persons receiving delayed (5-6 weeks) ART (“Cryptococcal Optimal ART Timing [COAT] study”). ¹⁶

Our analysis was limited by the small number of women who developed CM or died, which restricted our ability to assess the risks for these outcomes. We did not have a comparator arm to evaluate outcomes among asymptomatic women with plasma cryptococcal antigenemia who did not initiate ART. Additionally, fluconazole primary prophylaxis was prescribed and continued at physicians’ discretion, which could have introduced channeling bias. Since adherence to prophylaxis was not assessed, and because Thai clinics and hospitals use locally produced fluconazole with potentially different pharmacokinetics and pharmacodynamics, the dosage of fluconazole primary prophylaxis at 400 mg weekly may have been insufficient. Finally, we had inadequate microbiological capacity in this study to ascertain whether fluconazole resistance (although rare) contributed to the development of CM in women receiving fluconazole primary prophylaxis.

In settings with a high prevalence of cryptococcal antigenemia, CrAg screening among HIV-infected persons with CD4 <100 cells/mm³ can identify patients with antigenemia who may benefit from preemptive treatment. ⁵ Further research into the optimal timing of ART and antifungal treatment for persons with antigenemia is needed. Our results support the WHO 2011 recommendation for ART initiation as an important intervention for the prevention of cryptococcal disease in HIV-infected persons. ⁵