Virological outcomes in ARV-naïve patients switching or not from a first successful boosted PI-regimen to efavirenz, nevirapine or abacavir regimens

Methods Using the French Hospital Database on HIV (FHDHANRS Co4), 439 antiretroviral (ARV)-naive patients with undetectable viral load (VL) who switched from a first boosted PI-containing cART to a combination including EFV (n=196), NVP (n=123) or ABC (n=120) were selected. Each patient was matched with 3 patients who did not change their cART on the basis of sex, age, CD4 cell count, VL and date of the first cART initiation and duration of undetectability. Time to virological failure (VF) was analysed using Kaplan-Meier curves and Cox models. Potential confounding variables considered for the analyses were HIV transmission group, at the date of first PI-cART initiation: NRTI backbone, PI drug; at the index date: AIDS status, CD4 cell count, NRTIs background, calendar period, time since inclusion in the database, time since first PI-cART initiation, time between undetectability and switch. Each variable associated with VF in the univariate model (p<0.20) was included in a multivariable model designed to evaluate the impact of the sole switch first, then the impact of the switch regimen, on the risk of VF. Each model was stratified by the matched groups (exposed/matched nonexposed patients).