CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

the PANGEA-HIV initiative. The vast majority of participants were on ART. Amplification and near full-length HIV genome sequencing were performed from proviral DNA. Duration of HIV infection was dichotomized, as < or > 1 year for 1153 participants based on longitudinal follow-up before and after diagnosis. We calculated viral diversity at each nucleotide site and for gag, pol and env based on nucleotide frequency files. We optimized a logistic regression model to predict recency <1 year and assessed model performance by cross- validation. Results: In our dataset, 140 individuals had been infected for less than a year at diagnosis. Most patients (954/1143, 83.5%) were on ART at the time of sampling. We split our data randomly into training (70%) and testing (30%) subsets. Our best predictive model included genetic diversity for pol, gag and env, viral load, age, gender and ART-status. Prediction accuracy in the test dataset was 94.4% (91.3%-96.6%) compared to a null model accuracy of 87.8% (p<0.0001). Model sensitivity was 74.4%, specificity 97.2% and positive predictive value was 78.4%. Accuracy was consistent across 1000 cross- validation tests. The model including only genetic diversity measures, without any demographic or clinical variables, was not significantly better than the null model (p>0.05) because of the confounding effect of ART-status. Conclusion: In contrast to previous analyses, most of our sequences came from proviral DNA from individuals on suppressive ART. Among treated patients, genetic diversity measures (e.g. entropy) displayed overlap between recent and chronic infections (Figure 1) but including clinical and demographic data improved prediction of recency. We are currently evaluating whether machine learning can incorporate additional information (e.g. sites under selection) to further distinguish between recent and chronic infections in treated individuals.

HIV RNA, ARCHITECT Ag/Ab, Bio-Rad HIV-1/2 Ab (IgM sensitive), and Geenius HIV-1/2 lateral flow antibody assay. Results: From January 2017 through August 2019, 174 were enrolled and completed centralized confirmatory testing. Their median age was 27 (interquartile range 23-38) years and 29 (17%) were female. ART was started by 154 (89%) on the day of enrollment and 20 (11%) the next day, mostly with study-provided elvitegravir/cobicistat/tenofovir alafenamide/emtricitabine (n=136, 78%). AHI was confirmed in 167 (96%) participants after centralized testing and varied in Fiebig stage according to the AHI criteria used (Figure Panel A). Three participants with locally detectable HIV RNA had no evidence of infection on centralized testing, discontinued ART and were withdrawn. Four others were in Fiebig VI, not AHI, at enrollment. Centralized ARCHITECT S/CO ≥10 combined with nonreactive or indeterminate HIV antibody on the Geenius assay correctly identified 87 of 106 (82%) Fiebig II-IV AHI cases (Figure Panel B). Conclusion: Novel efficient AHI criteria incorporating ARCHITECT S/CO into diagnostic algorithms facilitated rapid ART initiation pending confirmation. False-positive diagnoses of AHI were rare. These new criteria may facilitate AHI diagnosis, staging, and immediate ART initiation in research studies and clinical practice.

Poster Abstracts

974 NOVEL CRITERIA FOR DIAGNOSING ACUTE HIV IN A MULTINATIONAL ART-INITIATION STUDY Trevor A. Crowell 1 , Justin Ritz 2 , Robert Coombs 3 , Lu Zheng 2 , Joseph J. Eron 4 , John W. Mellors 5 , Gert U. van Zyl 6 , Javier R. Lama 7 , Kiat Ruxrungtham 8 , Beatriz Grinsztejn 9 , Roberto Arduino 10 , Lawrence Fox 11 , Jintanat Ananworanich 1 , Eric Daar 12 , for the AIDS Clinical Trials Group (ACTG) A5354 Study Team 1 US Military HIV Research Program, Silver Spring, MD, USA, 2 Harvard T.H. Chan School of Public Health, Boston, MA, USA, 3 University of Washington, Seattle, WA, USA, 4 University of North Carolina at Chapel Hill, Chapel Hill, NC, USA, 5 University of Pittsburgh, Pittsburgh, PA, USA, 6 Stellenbosch University, Cape Town, South Africa, 7 Asociacion Civil Impacta Salud y Educacion, Lima, Peru, 8 Thai Red Cross AIDS Research Center, Bangkok, Thailand, 9 Institute de Pesquisa Clinica Evandro Chagas, Rio de Janeiro, Brazil, 10 University of Texas at Houston, Houston, TX, USA, 11 DAIDS, NIAID, Bethesda, MD, USA, 12 Los Angeles Biomedical Research Institute at Harbor– UCLA Medical Center, Torrance, CA, USA Background: Antiretroviral therapy (ART) initiation during acute HIV infection (AHI) limits HIV reservoirs, enhances reservoir decay, restricts viral genetic diversification and may facilitate post-ART control. Identifying and treating persons with AHI is highly desirable but logistically challenging. We describe the performance of new AHI diagnostic criteria for an ongoing multi-national study of ART initiation during AHI. Methods: ACTG 5354 enrolls adults during AHI at 29 sites in the Americas, Africa, and Southeast Asia. Participants must meet one of the following criteria: (A) detectable HIV RNA and non-reactive HIV antibody; (B) detectable HIV RNA or reactive antibody and negative/indeterminate Western blot (WB) or Geenius; (C) negative HIV RNA or antibody within 90 days and reactive antibody, WB (p31-), or Geenius (p31-) within 7 days; (D) ARCHITECT or GSCOMBO antigen/ antibody (Ag/Ab) combo signal-to-cutoff ratio (S/CO) ≥10 and non-reactive HIV antibody. Participants start ART at enrollment. HIV infection and Fiebig stage at ART initiation are subsequently confirmed by centralized testing that includes

975 IMPROVING CLASSIFICATION FOR RECENT HIV INFECTION USING TOP SCORING PAIRS Athena Chen 1 , Oliver Laeyendecker 2 , Charles S. Morrison 3 , Susan H. Eshleman 2 , Ingo Ruczinski 1 1 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 2 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 3 FHI 360, Durham, NC, USA Background: HIV cross-sectional incidence assays often measure characteristics of the antibody (Ab) response, such as titer, class, and avidity. These responses are affected by various factors including viral suppression. Since older guidelines recommended ART initiation later in infection, incidence algorithms often use viral suppression as a surrogate for non-recent infection. As new guidelines recommend ART initiation early in infection, viral suppression no longer indicates non-recent infection, and new approaches are needed to identify recent infections. Methods: We analyzed Ab profiles for 258 samples from 57 individuals with HIV subtype C infection and known duration of infection (2 mo. - 8.7 yrs.) using phage immunoprecipitation sequencing (PhIP-Seq). PhIP-Seq quantifies Ab binding to 3384 peptides spanning the HIV genome. Our novel classifier for recent (2-6 mo.) and non-recent (18+mo.) infection is based on the k-top scoring pairs (TSP) classifier. For each peptide pair, relative Ab abundances

CROI 2020 366