CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Results: The subjects’ demographics were: mean age 41.9 years, 54%male, 77% reported income < $2000/month, 61% GED or less education. Duration of observation in days was Mean (SD) 97.1 (44.8), DOT 95 (30.1), WOT 98.1 (50.8). Table 1 shows the percentage of confirmed/prescribed doses for DOT and WOT analyzed at the group and individual subject level. WOT associated adverse events were skin rash and pruritus (10%). Conclusion: WOT remotely confirmed a significantly greater percentage of prescribed doses than DOT over both 7 and 5 day analysis periods within highly funded US TB DOT programs. WOT is highly accurate and was a superior alternative to DOT for confirming adherence to TB medication in the continuation phase of treatment.

784 CRAG STATUS AND EFFECT ON BENEFITS FROM ENHANCED PROPHYLAXIS IN THE REALITY TRIAL Sarah Pett 1 , Lewis Haddow 2 , Ruth Nhema 3 , Moira J. Spyer 1 , Laura Benjamin 2 , Grace Najjuka 4 , Ibrahim Daud 5 , Jay Berkley 6 , Juliet Kitabalwa 4 , James G. Hakim 3 , Rob Heyderman 2 , Sarah Walker 1 , Diana Gibb 1 1 MRC Clinical Trials Unit at UCL, London, UK, 2 University College London, London, UK, 3 University of Zimbabwe, Harare, Zimbabwe, 4 Joint Clinical Research Centre, Kampala, Uganda, 5 Moi University, Eldoret, Kenya, 6 KEMRI–Wellcome Trust Research Programme, Kilifi, Kenya Background: In HIV-infected adults/children with CD4<100 cells/ul starting ART in sub-Saharan Africa, the REALITY trial (ISRCTN43622374) showed that an enhanced prophylaxis (Px+) package (minimum cost $5.6) including 12 weeks’ fluconazole 100mg OD at ART initiation significantly reduced all-cause mortality, mortality from unknown causes and cryptococcus, and incidence of new cryptococcal disease vs cotrimoxazole alone (CTX). We assessed the impact of enrolment cryptococcal antigen (CRAG) status on these outcomes. Methods: Stored enrolment plasma was tested using the IMMY CrAg lateral flow assay. Logistic regression with backwards elimination (p>0.1) identified independent predictors of baseline CRAG status, and proportional hazards models estimated the impact of Px+ vs CTX on all, cryptococcal and unknown deaths, and new cryptococcal disease, through 24 weeks by baseline CRAG. Results: 1550(86%) adults from Kenya, Uganda and Zimbabwe (Malawi results pending) with median baseline CD4 36 cells/ul (IQR 16-63) and VL 275,700 c/ ml, were randomized to CTX (n=771) vs Px+ (n=779). Excluding 23 (1.5%) with active cryptococcal disease at enrolment, 55(7.3%) vs 51(6.7%) were CRAG+ in CTX vs Px+ respectively. CRAG+ patients had lower CD4 (OR=0.90 per 10 cells/ul higher (95% CI 0.83-0.97) p=0.006) and were less often female (OR vs male=0.68 (0.45-1.02) p=0.06). Over 24 weeks on ART, there were 96 CTX vs 68 Px+ deaths. 6 of 7 CTX deaths and 2 of 3 Px+ deaths due to cryptococcal disease were CRAG+, whereas there was only 1 CRAG+ among 44 deaths from unknown causes in CTX vs 1 among 24 on Px+. Over 24 weeks, there were 12 CTX vs 6 Px+ new cryptococcal meningitis cases; 10 vs 5 respectively were baseline CRAG+. Px+ reduced new cryptococcal disease equally in CRAG+ (HR=0.50) and CRAG- (HR=0.49) (interaction p=0.99) (figure); similarly for all deaths (interaction p=0.93). Of 49 patients treated for cryptococcal disease post-enrolment, only 12(24%) were CRAG+ at enrolment. Conclusion: ~ 7% patients were CRAG+ pre-ART, without overt cryptococcal disease. CRAG+ was rare among unknown deaths suggesting these were unlikely due to cryptococcus. The relative benefits of fluconazole-containing Px+ were similar among CRAG+ and CRAG- patients. These data support the use of this affordable fluconazole-containing enhanced Px+ bundle in this severely immunocompromized group, particularly where costs of tests are similar to costs of fluconazole ($2.6) or where availability is limited.

783 DIAGNOSTIC UTILITY OF MULTIPLEX PCR IN AN HIV-INFECTED POPULATION WITH MENINGITIS Sarah Bridge 1 , Kathy Huppler Hullsiek 2 , Richard Kwizera 3 , Edwin Nuwagira 1 , Anna Stadelman 2 , Tadeo K. Kiiza 3 , Liberica Ndyatunga 1 , Kenneth Ssebambulidde 3 , James Mwesigye 1 , Edward Mpoza 3 , Mahsa Abassi 2 , David Meya 3 , Conrad Muzoora 1 , 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: Meningitis remains a frequent cause of mortality among HIV- infected people in sub-Saharan Africa. We evaluated the utility of a commercial multiplex PCR assay (BioFire FilmArray® Meningitis/Encephalitis (ME) panel, Salt Lake City, Utah) in HIV-infected individuals with suspected meningitis in Uganda. Methods: We collected CSF samples from 211 HIV-infected Ugandan adults with suspected meningitis from January 2016 to May 2017. We collected CSF samples at diagnosis in first episode of meningitis (n=129), recurrent cryptococcosis (N=6), and during therapeutic LPs following cryptococcal diagnosis (N=76). Standard bacterial, mycobacterial, and fungal CSF diagnostics were performed on site. The prevalence of Cryptococcus, bacterial, and viral pathogens in CSF were determined using the FilmArray® ME panel, run on-site in real time. We assessed the diagnostic performance of the panel for Cryptococcus. Results: Cryptococcal meningitis was common in this population with 57% (74/129) of baseline specimens having a positive CSF cryptococcal antigen (CrAg). The sensitivity of the PCR panel for first cryptococcal episode was 85% (63/74) and the specificity was 98% (54/55) compared to CSF CrAg. All but one false negative baseline PCR result occurred when corresponding CSF cultures were either sterile (n=7) or had a quantitative cryptococcal count ≤ 100 CFU/mL (n=3). The single false positive result occurred in a subject with a positive serum CrAg but no other evidence of CNS infection. In those with a previous history of Cryptococcus (symptomatic relapse), multiplex PCR identified 3 of 5 cases of relapse. In follow-up samples obtained during therapeutic LPs, FilmArray® predicted conversion to culture sterility with 89% (34/38) negative predictive value (a negative FilmArray® = sterile culture). Other pathogens detected by PCR included CMV (n=6), HHV-6 (n=6), HSV-1 (n=5), S. pneumoniae (n=4), H. influenzae (n=3), HSV-2 (n=3), and VZV (n=3) (Table). Conclusion: The FilmArray® ME panel appears to be a useful platform for the rapid diagnosis of CNS infections in an HIV-infected population. Commercial PCR testing is sensitive for detecting Cryptococcus in CSF when fungal burden is high, though sensitivity may be decreased at lower fungal burdens. PCR may predict conversion to culture sterility after cryptococcal meningitis diagnosis.

Poster Abstracts

CROI 2018 294