Increased mortality among HIV infected patients with cryptococcal antigenemia in Guinea-Bissau

Cryptococcal antigenemia may precede development of cryptococcal meningitis and death among patients with advanced HIV infection. Among 200 retrospectively and randomly selected ART-naïve patients with CD4 counts < 100 cells/μl from Guinea-Bissau, 20 (10%) had a positive cryptococcal antigen test. Self-reported headache and fever were predictors of a positive test, while cryptococcal antigenemia was a strong predictor of death within the first year of follow-up, MRR 2.22 (95% CI: 1.15-4.30). Screening for cryptococcal antigenemia should be implemented for patients with advanced HIV in Guinea-Bissau. Pre-emptive anti-fungal therapy should be initiated prior to ART-initiation if the screening is positive.


Introduction
Cryptococcal meningitis (CM) is a severe opportunistic infection among HIV-infected patients and is estimated to cause almost 625.000 deaths each year globally. The highest prevalence is found in sub-Saharan Africa, where 13-44% of HIV/AIDS-related deaths are due to CM [1]. Cryptococcal antigen (CrAg) is detectable in blood several weeks before meningitis symptoms [2] and thereby precedes the development of CM. The prevalence of cryptococcal antigenemia among HIV-infected patients has been described in several African and Southeast Asian countries, with a prevalence ranging from 1-19%, the highest prevalence found among patients with low CD4 counts [3]. Data from West Africa are sparse. One study has described the prevalence of cryptococcal antigenemia among a smaller group of HIV-infected individuals in Guinea-Bissau and found a prevalence of 1% among patients with a median CD4 count of 185 cells/µl [4]. A high prevalence of cryptococcal antigenemia could be expected in Guinea-Bissau, since late presentation is common [5]. The WHO recommends the use of CrAg screening in antiretroviral therapy (ART)-naïve patients with a CD4 count < 100 cells/µl in populations with a high prevalence of cryptococcal antigenemia (prevalence > 3%). If the CrAg test is positive, pre-emptive anti-fungal therapy prior to ART is recommended. The aim of this study was to investigate the prevalence of cryptococcal antigenemia, risk factors for antigenemia and examine cryptococcal antigenemia as a risk factor for mortality among patients with advanced HIV-infection in Guinea-Bissau, to provide data to guide decision on CrAg screening.

Methods
Setting and study design: All data, except CrAg status, were prospectively collected from patients enrolled in the Bissau HIV Cohort, located at the HIV clinic at Hospital Nacional Simão Mendes, Bissau, Guinea-Bissau [5]. CrAg was measured retrospectively in stored samples. Patients testing HIV-positive at the clinic are invited to be enrolled in the cohort. ART-naïve patients are recorded as lost to follow-up (LTFU) when they do not report to the clinic for 7 months, whereas patients on ART are recorded as LTFU when they are more than 3 months late for their last scheduled visit [5].
Patients are registered as dead if reported by family members or a contact person or confirmed by telephone calls from the clinic.
Whenever a CD4 count is performed, surplus plasma is stored in a biorepository in Aarhus, Denmark. Retrospectively, we randomly selected 200 ART-naïve patients with a CD4 count < 100 cells/µl and a corresponding available plasma sample.

Discussion
In these 200 patients with CD4 counts < 100 cells/µl, 10% had a positive plasma CrAg test. CrAg positive patients had a higher mortality compared with CrAg negative patients within one year from CrAg detection. The prevalence found in our study is comparable to the prevalence found in the few previous studies from West Africa [4,6]. Except self-reported headache and fever, we found no difference in socio-demographics or clinical presentation between CrAg positive and negative patients. Previous studies have found low CD4 counts, high WHO clinical stage, headache and fever to be associated with cryptococcal antigenemia [2,7]. Including only patients with CD4 counts below 100 cells/µl may be the reason we did not see the same associations between CD4 counts and cryptococcal antigenemia in this study. Presence of plasma CrAg was associated with higher risk of death in our study. Similar associations have been found in previous studies from Sub-Saharan Africa. In Tanzania, cryptococcal antigenemia was an independent predictor of death or being LTFU among ART-naïve HIV-infected adults [8] and in South Africa, cryptococcal antigenemia was an independent predictor of mortality among ART-naïve patients [9]. A study from Kenya found no significant associations between CrAg positivity and increased risk of death [10].
The difference in mortality outcomes between these studies may be affected by the level of CD4 cell counts, clinical status and ART status of the patients. This is the first study in Guinea-Bissau to investigate the prevalence of cryptococcal antigenemia among a group of patients to whom WHO recommends CrAg screening if the prevalence is higher than 3%. Furthermore, it is one of few studies