In sub-Saharan Africa, cryptococcal meningitis causes 720 000 deaths annually . The mortality of cryptococcal meningitis in much of Africa is 40–66% [2,3][2,3]. This is due largely to factors including the unavailability of, or difficulty administering, gold-standard intravenous antifungal treatment with amphotericin B and flucytosine [4–6][4–6][4–6], challenges monitoring for drug toxicity, and shortages of physicians able to perform serial lumbar punctures. For these reasons, preventing the disease altogether may most effectively reduce mortality. Screening asymptomatic patients for cryptococcal antigenemia provides a window for such preemptive intervention. Cryptococcal antigen (CrAg) becomes detectable in serum and urine an average of 22 days before symptom onset in patients who ultimately develop cryptococcal meningitis . Asymptomatic cryptococcal antigenemia independently predicted mortality in patients initiating antiretroviral therapy (ART) , and asymptomatic CrAg-positive HIV-infected adults have a 20% higher mortality than CrAg-negative adults [2,7,8][2,7,8][2,7,8].
Retrospective studies suggest that fluconazole at a variety of doses decreases mortality in CrAg-positive patients , and the World Health Organization (WHO) recommends fluconazole for HIV-infected adults with asymptomatic cryptococcal antigenemia . The WHO has provisionally recommended that fluconazole be dosed at 800 mg daily for 2 weeks, then 400 mg daily for 8 weeks, then 200 mg indefinitely, but admits that ‘the optimal antifungal regimen in this population remains to be determined’ . In resource-limited settings, such as our hospitals in Tanzania, fluconazole supply is limited and 5–10% of patients initiating ART are CrAg-positive [2,7][2,7]. Our hospital and many others in the region have therefore determined that implementing the current WHO-recommended fluconazole regimen is not feasible.
Therefore, two Tanzanian hospitals implemented a serum CrAg screening and shorter-course fluconazole treatment programme for patients initiating ART. We were asked to determine outcomes and predictors of mortality after the programme's first 15 months. We hypothesized that the 6-month mortality of CrAg-positive patients who received short-course intensive fluconazole would be comparable to that of CrAg-negative patients. This would be a remarkable achievement for patients whose mortality would otherwise be expected to exceed that of CrAg-negative patients by 15–20%.
This study was conducted at the outpatient HIV clinics of Bugando Medical Centre, the referral hospital for 13 million people in northwest Tanzania, and at Sekou-Toure Regional Hospital, both in Mwanza City.
The hospital screening programme for asymptomatic cryptococcal antigenemia, which began in September 2012, was provided to all HIV-infected adults (≥18 years) initiating ART who had CD4+ T-cell counts (CD4+ cell counts) less than 200 cells/μl. Patients were first screened for signs/symptoms of meningitis, and those who screened positive were referred for further workup. The remaining patients were tested for asymptomatic cryptococcal antigenemia using the serum CrAg using the lateral flow assay (Immuno-Mycologics, Inc., Norman, Oklahoma, USA), as previously described . After CrAg-positive patients had been confirmed to have no signs or symptoms of meningitis, the programme provided them with 800 mg fluconazole orally for 2 weeks, followed by 400 mg orally for 2 weeks, given at the same time that ART was initiated. This intense, short-course regimen was in accordance with the WHO's conditional recommendation that CrAg-positive patients should be treated, as well as with hospital policy and feasibility based on fluconazole supply.
For our study, we identified all patients who had screened positive for CrAg during the first 15 months of the hospital programme. We reviewed patients’ files and excluded from our analysis any who had a history of cryptococcal meningitis, prior fluconazole use, or who were subsequently transferred to other health facilities in the 6 months following their CrAg screening. For every CrAg-positive patient, we used the ART initiation record book to identify the subsequent three CrAg-negative patients who started ART. We obtained written informed consent from these CrAg-positive and CrAg-negative patients or their family members to review their charts and to re-test those who had been CrAg-positive for serum CrAg at least 6 months after they had been treated with short-course fluconazole.
We performed a single study visit approximately 6–10 months after patients’ initial CrAg screening and treatment. During this visit, we performed a physical examination and repeated the CrAg test. We also reviewed data from patients’ 1, 3, and 6-month visits to the CTC. Causes of death were determined from hospital records and verbal autopsy. Patients who were found to be CrAg-positive at the 6 to 10–month visit were given high-dose long-term fluconazole.
Data was analysed using Stata13 (College Station, Texas, USA). The primary endpoint was all-cause mortality rate in each group at 6 months. We conducted a noninferiority analysis with an inferiority margin of 20%, based on the ∼20% population-attributable mortality in CrAg-positive versus CrAg-negative patients [2,4,5][2,4,5][2,4,5]. Secondary endpoints included rates of cryptococcal meningitis and of nonmeningeal cryptococcal disease in both groups. Medians and proportions were compared by Wilcoxon rank-sum and Fisher's exact test, respectively. Factors associated with mortality were evaluated by univariable and subsequent multivariable analysis. Kaplan–Meier curves and the Cox proportional hazards ratio were used to compare survival differences.
Ethical clearance was obtained from the CUHAS/BMC/Sekou-Toure Ethics Committee and the IRB at Weill Cornell Medical College.
From September 2012 to November 2013, 377 adults with CD4+ cell counts less than 200 cells/μl initiated ART at our HIV clinics. Of these, 70 had had signs/symptoms that were possibly consistent with cryptococcal disease and had been referred for further workup rather than screening for asymptomatic cryptococcal antigenemia. In addition, 22 had used fluconazole previously, 59 were transferred to an outside facility in the 6 months following ART initiation, and 10 did not have serum CrAg testing. Therefore, 216 patients underwent serum CrAg screening in the first 15 months of the hospital programme, and 18 of these (8.3%) tested positive. Among the remaining 198 CrAg-negative patients, three were selected serially after every positive patient was identified, yielding 54 CrAg-negative patients for comparison.
Study participants were 56% female, with a median age [interquartile range] of 36.5 [29.5–45] years, BMI of 21.1 [18.8–23.7] kg/m2, and CD4+ cell count of 96 [52–149.5] cells/μl. Over 65% had either WHO Stage 2 or 3 clinical disease. No statistically significant differences were identified between CrAg-positive and CrAg-negative groups with respect to any of these characteristics (Table 1).
All CrAg-positive patients or their families retrospectively reported complete adherence to the 4-week fluconazole regimen. In the CrAg-positive group, 2/18 (11.1%) died, compared with 8/54 (14.8%) in the CrAg-negative group by 6 months. In the CrAg-positive group, mortality incidence was 3.7% less (one-sided 90% confidence interval (CI), no more than 10.8% higher; two-sided 95% CI, [−3.4% – +25.6%]; Cox proportional hazards ratio = 0.73 [0.16–3.45], P = 0.70, Fig. 1). In the positive cohort neither death was attributable to cryptococcal disease, whereas in the negative cohort, one patient died of cryptococcal meningitis confirmed by positive cerebrospinal fluid India ink and CrAg (Table 2). At the time of starting ART 3 months prior to his death, the patient's CD4+ cell count had been 66 cells/μl.
Significant predictors of mortality on univariable analysis included male sex (odds ratio (OR) 6.3 [1.2–32.4], P = 0.027), older age (OR 1.10 [1.02–1.18] per year older, P = 0.012), lower BMI (OR 0.73 [0.57–0.94] per unit increase, P = 0.013), lower CD4+ cell count (OR 0.98 [0.97–0.99)] per cell/μl increase, P = 0.031), and advanced WHO clinical stage (OR 13.0 [3.1–60.6] per advancing stage, P = 0.001). Of note, serum CrAg did not significantly predict mortality in either univariable or multivariable analysis (OR 0.72 [0.14–3.72], P = 0.70 and OR 0.19 [0.005–7.6], P = 0.38, respectively). Factors remaining significantly predictive of mortality on multivariable analysis were age, CD4+ cell count, and WHO stage.
Two of the 16 living CrAg-positive patients had persistent antigenemia on serum retesting 6–10 months after their original screening and treatment with short-course high-dose fluconazole. The other 14 patients were retested and were CrAg-negative. These patients, both women, had CD4+ cell counts that had increased from 51 to 72 and 90 to 142 cells/μl on ART. Both remained asymptomatic and were restarted on long-term fluconazole.
Six-month mortality of HIV-infected patients initiating ART with asymptomatic cryptococcal antigenemia, who received an intensive 4-week course of fluconazole, was noninferior to CrAg-negative patients. This finding is important at a time when the optimal fluconazole dose for CrAg-positive patients remains unknown, leading to conditional recommendations for high-dose and subsequent indefinite maintenance fluconazole. Lifelong fluconazole dosing may place severe strain on the fluconazole supply in resource-limited settings in which nearly 10% of patients initiating ART are CrAg-positive. Fluconazole is also the only treatment available for patients with invasive cryptococcal disease. A randomized, controlled trial recently demonstrated that a 10-week course of high-dose fluconazole (without the lifelong maintenance therapy afterwards) was associated with an approximately 15% reduction in mortality among HIV-infected African adults beginning ART . Ensuring that CrAg screening and treatment programmes are feasible for wide implementation is critical to reducing AIDS-related mortality in the poorest settings.
Our study supports other data suggesting that fluconazole courses far shorter than the WHO recommendations may be beneficial in asymptomatic CrAg-positive patients. In Uganda, 71% of CrAg-positive patients treated with 200–400 mg of fluconazole for 2–4 weeks survived for 30 months, compared with zero of five who received no fluconazole . In our cohort, we would have expected at least 4–5 of 18 CrAg-positive patients to develop cryptococcal meningitis if insufficiently treated. Instead, none developed cryptococcal meningitis and all but two cleared their antigenemia within 6 months. Notably, these two patients both had experienced a CD4+ increase of less than 75 cells/μl, leading us to note that two of four patients with CD4+ increases less than 75 cells/μl, compared with 0/12 patients with CD4+ increases greater than 75 cells/μl, had persistent antigenemia. This data suggests that shorter course fluconazole could be possible in people with good CD4+ cell count recovery.
It is also possible that shorter, intense-course fluconazole could be less effective in patients with high fungal burdens. Due to limited resources, CrAg titers at the time of fluconazole initiation had not been measured in any patient in this study. Other possible causes of persistent antigenemia, denied by both of our patients, are poor adherence to fluconazole or use of concurrent medications (e.g. rifampin) that enhance fluconazole metabolism. Alternately, these patients, whose CD4+ cell counts remained low after 6 months, may have cleared their initial antigenemia but subsequently become re-infected. Both patients restarted high-dose fluconazole and now remain on lifelong therapy, with ongoing follow-up.
Two important limitations of this study are the retrospective design and the small number of study participants. Larger, prospective studies are underway to determine the relationship between CrAg titers and antigen clearance, as well as time to antigen clearance in asymptomatic patients given fluconazole.
Interestingly, one patient in our CrAg-negative cohort died of cryptococcal meningitis. A major strength of this work was our ability to trace causes of deaths for all 10 patients. Although we cannot determine whether this death represents a false-negative CrAg screen or a patient who acquired cryptococcal infection after starting ART, the 96–100% sensitivity of the serum CrAg assay  suggests the latter. The patient's extremely low CD4+ cell count (66 cells/μl at the time that he initiated ART 3 months prior to his death) suggests he may have remained susceptible to cryptococcal infection and acquired it shortly after beginning ART.
In conclusion, our work suggests that 4 weeks of high-dose fluconazole prevents death in CrAg-positive patients. Though small, the study's robustness is attested by the finding that other well known predictors (age, CD4+ cell count, WHO stage) were significantly associated with mortality. Our work offers new hope for cryptococcal meningitis prevention even in settings where fluconazole is not limitless, and even where laboratory facilities are scarce as this study was done using a point-of-care diagnostic test. We strongly urge support for routine CrAg screening and preemptive fluconazole treatment for patients initiating ART, and for additional studies to further define optimal yet practical fluconazole dosing regimens.
This study would not have been possible without the hard work and determination of the staff at the Care and Treatment Centres at Bugando and Sekou-Toure, as well as in the Department of Internal Medicine at the Catholic University of Health and Allied Sciences. We additionally deeply thank the patients for their willing participation in this study.
Authors’ contributions: S.W.K. designed the study, conducted data collection, performed data analysis, drafted the manuscript. K.A.M. assisted with study design, performed data collection, revised and approved the final manuscript. S.E.K. assisted with study design, assisted with data interpretation, approved the final manuscript. D.W.F. assisted with study design, assisted with analysis and interpretation, revised and approved the final manuscript. R.N.P. designed the study, assisted with data analysis and interpretation, revised and approved the final manuscript. J.A.D. designed the study, oversaw data collection, performed data analysis, drafted the manuscript.
Support: This study was supported in part by grants from the NIH (K24 AI098627; K23 AI 110238).
Conflicts of interest
No authors have any potential conflicts of interest to disclose.
1. Park BJ, Wannemuehler KA, Marston BJ, Govender N, Pappas PG, Chiller TM. Estimation of the current global burden of cryptococcal meningitis among persons living with HIV/AIDS
2. Rajasingham R, Meya D, Boulware D. Integrating cryptococcal antigen screening and pre-emptive treatment into routine HIV care
. J Acquir Immune Defic Syndr
3. Wajanga BM, Kalluvya S, Downs JA, Johnson WD, Fitzgerald DW, Peck RN. Universal screening of Tanzanian HIV-infected adult inpatients with the serum cryptococcal antigen to improve diagnosis and reduce mortality: an operational study
. J Int AIDS Soc
4. Brouwer AE, Rajanuwong A, Chierakul W, Griffin GE, Larsen RA, White NJ, et al. Combination antifungal therapies for HIV-associated cryptococcal meningitis: a randomised trial
5. Perfect JR, Dismukes WE, Dromer F, Goldman DL, Graybill JR, Hamill RJ, et al. Clinical practice guidelines for the management of cryptococcal disease: 2010 update by the Infectious Diseases Society of America
. Clin Infect Dis
6. Jarvis JN, Harrison TS. HIV-associated cryptococcal meningitis
7. Liechty CA, Solberg P, Were W, Ekwaru JP, Ransom RL, Weidle PJ, et al. Asymptomatic serum cryptococcal antigenemia and early mortality during antiretroviral therapy in rural Uganda
. Trop Med Int Health
8. Mfinanga S, Chanda D, Kivuyo SL, Guinness L, Bottomley C, Simms V, et al. Cryptococcal meningitis screening and community-based early adherence support in people with advanced HIV infection starting antiretroviral therapy in Tanzania and Zambia: an open-label, randomised controlled trial
9. World Health Organization. Rapid advice diagnosis, prevention and management of cryptococcal disease in HIV infected adults, adolescents and children.
Geneva, Switzerland: WHO. 2011.
10. Magambo KA, Kalluvya SE, Kapoor SW, Seni J, Chofle AA, Fitzgerald DW, et al. Utility of urine and serum lateral flow assays to determine the prevalence and predictors of cryptococcal antigenemia in HIV-positive outpatients beginning antiretroviral therapy in Mwanza, Tanzania
. J Int AIDS Soc
11. Meya DB, Manabe YC, Castelnuovo B, Cook BA, Elbireer AM, Kambugu A, et al. Cost-effectiveness of serum cryptococcal antigen screening to prevent deaths among HIV-infected persons with a CD4+ cell count < or = 100 cells/microL who start HIV therapy in resource-limited settings
. Clin Infect Dis