CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

789 DECLINING INCIDENCE OF HIV-ASSOCIATED CRYPTOCOCCOSIS IN SOUTH AFRICA, 2005-2015 Nelesh P. Govender 1 , Cheryl Cohen 1 , Phelly Matlapeng 1 , Penny Crowther- Gibson 1 , Susan Meiring 1 , Alan Karstaedt 2 , Halima Dawood 3 , Kerrigan McCarthy 1 , Graeme Meintjes 4 , Vanessa Quan 1 1 National Institute for Communicable Diseases, Johannesburg, South Africa, 2 University of the Witwatersrand, Johannesburg, South Africa, 3 University of KwaZulu-Natal, Durban, South Africa, 4 University of Cape Town, Cape Town, South Africa Background: The annual incidence rate of cryptococcosis was 139 cases per 100000 HIV-infected persons in Gauteng province during South Africa’s pre- antiretroviral treatment (ART) era, 2002-2004. By 2015, an estimated 3.3 million of 6.9 million HIV-infected persons were on ART vs. approximately 100000 in 2005. We hypothesised that increased ART coverage would be associated with a decline in the incidence rate of cryptococcosis. Methods: We conducted national active population-based laboratory surveillance for cryptococcosis, 2005-2015. A case was defined as a person diagnosed with a first episode at any public-or private-sector laboratory with a positive cerebrospinal fluid India ink/ cryptococcal antigen (CrAg) test or culture of Cryptococcus from any specimen. We excluded patients with isolated cryptococcal antigenemia across the period because routine CrAg screening was initiated in 2012, initially in a few health districts. Annual incidence rates were calculated using mid-year HIV population denominators from the Thembisa model. We collected data on in-hospital management and outcome from a subset of cases at sentinel urban hospitals. Results: Over 11 years, 85969 incident cases were detected, the vast majority (81912; 95%) with cryptococcal meningitis. The national annual incidence rate peaked at 162 cases per 100000 HIV-infected persons (95%CI, 158-165) in 2006, declined to 127 cases per 100000 (95%CI, 125-130) in 2011 and to 90 cases per 100000 (95%CI, 88-92) in 2015. A reduction from the peak annual incidence rate was documented across all provinces in 2015 (range, 45% to 67%). Annual incidence was higher among males across all adult age groups in 2015; the most marked reduction in annual incidence occurred among women aged 20-44 years (Fig. 1). While a significantly higher proportion of patients was prescribed amphotericin B-based induction treatment (vs. fluconazole) each year (2005: 371/1204 [31%] vs. 2015: 355/391 [91%]; p<0.001), there was no corresponding reduction in the annual in-hospital case-fatality ratio (2005: 418/807 [34%]; 2015: 165/437 [38%]; p=0.06). More than half were ART-experienced in 2015 (204/379 [54%]). Conclusion: We demonstrated a 44% reduction in the national annual incidence rate of cryptococcosis from a peak in 2006 to below pre-ART levels in 2015, temporally associated with ART program expansion. Focused interventions, including HIV test and treat, could reduce the incidence among adult males and national CrAg screening may reduce cryptococcosis-related mortality.

790 ETIOLOGIES OF SYMPTOMATIC RECURRENCE OF HIV-ASSOCIATED CRYPTOCOCCAL MENINGITIS Emily Evans 1 , Kathy Huppler Hullsiek 2 , Edwin Nuwagira 1 , Edward Mpoza 3 , Kenneth Ssebambulidde 3 , Lillian Tugume 3 , Bozena M. Morawski 2 , Sarah Lofgren 2 , Mahsa Abassi 2 , Darlisha A. Williams 3 , Conrad Muzoora 1 , David Meya 3 , David R. Boulware 2 , Joshua Rhein 2 1 Mbarara University of Science and Technology, Mbarara, Uganda, 2 University of Minnesota, Minneapolis, MN, USA, 3 Infectious Disease Institute, Kampala, Uganda Background: Recurrence of cryptococcal meningitis (CM) after initial therapy causes substantial mortality among HIV-infected patients. Differentiating cases of immune reconstitution inflammatory syndrome (IRIS) from relapse, persistent infection, or new opportunistic infections presents a diagnostic challenge. We assessed clinical characteristics and etiologies of HIV-infected Ugandans with symptomatic CM recurrence. Methods: We enrolled a prospective cohort of 75 HIV-infected persons with recurrent symptoms of cryptococcal meningitis in Uganda from August 2013 to May 2017. All cases were reviewed by two experts, who provided a consensus diagnosis of relapse, persistence, CM-IRIS, or “other”. Statistical comparisons were made with Kruskal-Wallis, linear regression and chi-squared tests. An objective classification system based on CSF culture growth, timing of previous cryptococcal meningitis episodes and antiretroviral (ART) initiation was used to categorize cases, and the accuracy of this systemwas assessed. Results: CSF cultures were positive in 76% (57/75) of participants, reflecting frequent occurrence of relapse (40%; 30/75) or persistence of prior infection (36%; 27/75). Those with positive cultures had lower CD4 counts than those with negative cultures (18 (IQR 7-55) vs 76 (IQR 35-148) cells/µL; p=0.01). Compared to those with persistence, participants with relapse presented later (8.0 (IQR 5.2-13.3) vs 2.1 (IQR 1.4-3.6) months from initial diagnosis; p<0.001), were more likely to have been previously treated with amphotericin B (90% vs 26%; p<0.001), and were less likely to be currently receiving fluconazole (33% vs 52%; p=0.16). CSF cultures were negative in 24% (18/75) of participants, of which 39% (7/18) were thought to represent paradoxical CM-IRIS. Individuals with CM-IRIS presented a median of 3.0 (IQR 1.1-3.7) months after initial cryptococcal meningitis diagnosis. A classification system utilizing culture results and the timing of previous cryptococcal meningitis and ART initiation optimally differentiated etiology in 85% (64/75) of the cases. Conclusion: The causes of recurrent symptoms after cryptococcal meningitis include relapse following early discontinuation of fluconazole prophylaxis, persistence with substandard induction therapy, and CM-IRIS. Culture, together with a detailed patient history, remains central to determining the etiology of symptomatic cryptococcal meningitis recurrence. Additional studies aimed at improving diagnostics in this diverse population are needed. 791 HIGH MORTALITY ASSOCIATED WITH UNMASKING CRYPTOCOCCAL MENINGITIS Joshua Rhein 1 , Kathy Huppler Hullsiek 1 , Nathan C. Bahr 1 , Lillian Tugume 2 , Edwin Nuwagira 3 , Kenneth Ssebambulidde 2 , Reuben Kiggundu 2 , Emily Evans 3 , Edward Mpoza 2 , Darlisha A. Williams 2 , Mahsa Abassi 1 , Abdu Musubire 2 , Conrad Muzoora 3 , David Meya 2 , David R. Boulware 1 1 University of Minnesota, Minneapolis, MN, USA, 2 Infectious Disease Institute, Kampala, Uganda, 3 Mbarara University of Science and Technology, Mbarara, Uganda Background: Increased antiretroviral therapy (ART) availability in Africa has led to more patients developing cryptococcosis after ART initiation. Despite this changing epidemiology, data regarding cryptococcal meningitis (CM) in those already receiving ART are lacking. Preliminary analyses (2015; n=172) suggested poor outcomes of unmasking CMwith recent ART initiation. We sought to confirm and further characterize this observation by comparing clinical presentation and outcomes in a large cohort of ART-naïve and ART-experienced adults. Methods: We prospectively enrolled 626 HIV-infected persons with CM in Uganda from August 2013 to May 2017. Participants were classified by ART status and the timing of ART initiation. Statistical comparisons were made with Kruskal-Wallis or Fisher’s Exact tests, with a primary endpoint of 2-week survival. Results: Overall, 48% (300/626) of participants were receiving ART at presentation, having initiated ART a median of 122 (IQR, 28-760) days prior to CM diagnosis. Compared with those not receiving ART, participants receiving ART had higher CD4 counts (median 30 (IQR, 10-79) vs 12 (IQR, 6-46) cells/µL; p =.02) and lower CSF fungal burdens (median 4.1 (IQR, 2.1-5.2) vs 5.0 (IQR,

Poster Abstracts

CROI 2018 297