CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

Conclusions: The sensitivity of HIV-RTDs showed significant differences. The range included state-of-the-art HIV antibody detection, low sensitivity for the identification of recent infection and false-negative results including in chronic infection. The sensitivity with oral fluid in this study was less suitable. The use of rapid tests with reasonable sensitivity therefore requires careful selection: In a situation following recent exposure to HIV and if individuals have received either early antiretroviral therapy, or haved failed on post- or pre-exposure prophylaxis, a negative result with oral fluid should be viewed with caution, even after a 3 months period has passed.

520 Comparison of Methods for HIV Incidence Estimation in a Cohort With Subtype C HIV

Allison R. Kirkpatrick 1 ; Charles Morrison 2 ;Tsungai Chipato 3 ; Katherine Schlusser 4 ; Pai-Lien Chen 2 ; Marshall Munjoma 5 ;Thomas C. Quinn 6 ; Susan H. Eshleman 6 ; Oliver Laeyendecker 7 1 NIAID, NIH, Baltimore, MD, USA; 2 FHI 360, Durham, NC, USA; 3 Univ of Zimbabwe, Harare, Zimbabwe; 4 Johns Hopkins Univ, Baltimore, MD, USA; 5 Univ of California San Francisco Prog, Univ of Zimbabwe, Belgravia, Zimbabwe; 6 Johns Hopkins Univ Sch of Med, Baltimore, MD, USA; 7 NIH, Bethesda, MD, USA Background: Accurate methods for cross-sectional incidence estimation are needed for HIV surveillance and HIV prevention research. This is particularly true for southern Africa, an HIV-1 subtype C endemic region, where the burden of HIV is highest worldwide. Methods: Samples were tested from a cohort of 176 seroconverters (1600 samples) from Zimbabwe (subtype C endemic area) followed for up to 9.9 years after seroconversion (median 6.9 years, interquartile range: 5.9 to 7.8) from the FHI 360 Hormonal Contraception and HIV trial. Three HIV avidity assays were analyzed alone at predetermined assay cutoffs, in combination with a viral load test (cutoff for recent infection >1000 copies/ml), or as part of a multi-assay algorithm (MAA). The assays evaluated were: the Limiting Antigen Avidity assay (LAg-Avidity; Sedia); a modified 3 rd generation Genetic Systems (GS) HIV 1 /2 +O EIA (JHU-3 rd BioRad); and a modified 4 th generation GS HIV Combo Ag/Ab EIA (JHU-4 th BioRad). The MAAs evaluated were previously optimized for incidence estimation in subtype B epidemics; these included JHU-3 rd BioRad and LAg-Avidity alone or JHU-3 rd BioRad, LAg-Avidity, CD4 count, and a viral load test (cutoff >400copies/ml). Performance characteristics evaluated were: the mean duration of recent infection (MDRI, the average time individuals appeared recently infected using a time window of 2 years); and the false recent rate (FRR, the frequency of being identified as recently infected for samples from individuals infected >2 years). Results: The MDRI and FRR for the different assays and algorithms are presented in the table below. A number of testing methods had MDRI >120 days and FRR <2%, which are the minimum performance characteristics needed for a cross-sectional incidence assay. Conclusions: All algorithms that included a viral load test met minimum performance guidelines. Both JHU-3 rd BioRad <40% AI and JHU-4 th BioRad <80% AI had an MDRI >140 days and an FRR <1%without the need of an additional viral load test. All two-assay algorithms that included an avidity assay plus viral load met minimum performance characteristics with MDRIs ranging from 133 to 214 days and FRRs ranging from 0 to 1%. The MAA that included LAg-Avidity + JHU-3 rd BioRad + CD4 count + viral load had an MDRI of 7.5 months and an FRR of 1%, well within the minimum performance characteristics specified by the WHO/UNAIDS working group for cross-sectional incidence assays.

Poster Abstracts

202

CROI 2016