CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

942 Impact of Each Month Delay in Initiation on the Effect of 1 Year of ART on CD4 Count Shu Yang 1 ; Susan J. Little 2 ; Davey M. Smith 2 ;Victor De Gruttola 1 ; Judith Lok 1 1 Harvard Sch of PH, Boston, MA, USA; 2 Univ of California San Diego, San Diego, CA, USA

Background: START results demonstrate that earlier antiretroviral treatment (ART) benefits HIV-infected individuals. We assess the impact of time from infection to starting ART on one-year ART effects using data from the observational prospective Acute Infection Early Disease Research Program (AIEDRP) . Challenges arise from confounding by indication (e.g., time from infection to initiation of ART may depend on disease state). Methods: AIEDRP data, which provide well-characterized estimated dates of infection (EDI), permit estimation of the dependence of change in CD4 counts after one year of ART on the time from EDI to initiation. We considered treatment initiation times within one year of EDI and investigated the effect of a year of ART in two ways: ignoring ART interruptions, and censoring patients who interrupted ART for more than 21 days. We used doubly robust estimation of Coarse Structural Nested Mean Models to adjust for baseline and time- varying confounders to eliminate selection bias likely to be present in a longitudinal observational study. Results: We included 1696 patients; 36.7% and 24.5% initiated ART during the acute and early phases respectively, and 38.8% did not initiate ART during follow-up. We modeled the expected CD4 count increase due to one-year ART as a linear function of time from EDI to ART initiation, given pre-treatment covariates. The results were presented in Table 1. These findings show a benefit of earlier ART initiation—particularly if there is no interruption in the first year of therapy; for example, starting 6 months earlier would lead to an expected added improvement in CD4 counts of 45.6 cells/μl after a year of therapy. These findings support those from Le et al (NEJM 2013), based on the same data but analyzed using different methods. Conclusions: Earlier initiation of ART during acute and early infection improves the CD4 count gain associated with one year of uninterrupted ART, emphasizing the importance of early detection of HIV infection and subsequent therapy. The new causal methods permit estimation of quantities that may be clinically relevant—the cost of each month of delay in ART initiation on CD4 response to therapy—that require sophisticated methods for adjustment for confounding by indication.

Poster Abstracts

943 Outcomes on CART in France According to Geographic Origin, Sex, and Transmission Group Laure-Amélie de Monteynard 1 ; Jacques Gilquin 2 ; Juliette Pavie 3 ; Jean-Luc Meynard 4 ; Marie-Aude Khuong 5 ; Anne Simon 6 ; Aba Mahamat 7 ; David Rey 8 ; Dominique Costagliola 9 ; Sophie Abgrall 10 1 Inst Pierre Louis d’Epidémiologie et de Santé Publique (IPLESP) UMR-S 1136, Paris, France; 2 Hosp Hôtel Dieu, Paris, France; 3 Hosp Européen Georges Pompidou, Paris, France; 4 APHP, Hosp Saint Antoine, Paris, France; 5 Hosp Delafontaine, Paris, France; 6 La Pitié Salpêtrière Hosp, Paris, France; 7 Cntr Hospier Andrée Rosemon, Cayenne, France; 8 Hôpitaux Universitaires de Strasbourg, Strasbourg, France; 9 Sorbonne Univs, Paris, France; 10 Univ Hosp Antoine-Béclère, Clamart, France Background: Cohort studies on outcomes after cART initiation according to geographic origin (GO) show conflicting results. We thus aimed to compare biological and clinical outcomes on cART in HIV-1-infected individuals in France, according to GO, sex and transmission group. Methods: Antiretroviral-naïve HIV-1-infected adults enrolled in the FHDH-ANRS CO4 cohort, originating either from France (FRA) or from sub-Saharan Africa or non-French West Indies (SSA/NFW), and who started cART between 2006 and 2011 were included. Women initiating cART because of pregnancy were excluded. We assessed 2-year Kaplan-Meier (KM) probabilities and adjusted hazard ratios (aHRs) for undetectability (2 consecutive plasma viral loads (pVL) <50 copies/mL) and CD4 cell recovery (2 consecutive CD4 >500/ µL), and 5-year cumulative incidences (CI) and adjusted subdistribution-hazard ratios (aSHRs) for clinical outcome (AIDS-event, serious non-AIDS-event (SNAE) or death) using competing risk models taking into account loss to follow-up, according to GO, sex and transmission group. Models were adjusted for demographic, immunovirological data and therapeutics at cART initiation (table), plus time-updated undetectability in immunological analysis or plus time-updated undetectability and CD4 increase >50/µL in clinical analyses. Results: Among 9746 individuals, 7297 (74.9%) were native from FRA and 2449 (25.1%) were migrants from SSA/NFW, of whom 1552 (21.3%) and 1350 (55.1%) were women respectively. A higher proportion of patients from SSA/NFW (38.1%) than from FRA (27.5%) initiated cART with CD4 <200 cells/µL (p<0.0001). After cART initiation and compared to FRA men who have sex with men (MSM), aHRs for undetectability were lower in men whatever their origin and slightly lower in FRA women; aHRs for CD4 cell recovery were lower in all groups of patients except for FRA non homosexual men; aSHRs of clinical outcomes (359 individuals with a new-AIDS event mainly in the first 6 months of cART, 1366 with NSAE, 49 deaths) were higher in men whatever their origin and borderline significant higher in SSA/NFWwomen. AIDS status, older age and lower CD4 cell count at cART initiation had the highest impact on the change between the crude and adjusted SHRs of clinical outcomes. Conclusions: Although migrants initiated cART at lower CD4 count and had the lowest likelihood of CD4 cell recovery compared to FRA MSM, male sex in non homosexuals whatever the geographic origin had a negative impact on undetectability and clinical outcome.

399

CROI 2016