Abstract
BACKGROUND: The effect of initial antiretroviral therapy (ART) class on cancer risk in people with HIV (PWH) remains unclear. SETTING: Cohort study of 36,322 PWH enrolled (1996-2014) in the North American AIDS Cohort Collaboration on Research and Design. METHODS: We followed individuals from ART initiation (protease inhibitor [PI]-, non-nucleoside reverse transcriptase inhibitor [NNRTI]-, or integrase strand transfer inhibitor [INSTI]-based) until incident cancer, death, loss-to-follow-up, 12/31/2014, 85 months (intention-to-treat analyses [ITT]), or 30 months (per-protocol [PP] analyses). Cancers were grouped (non-mutually exclusive) as: any cancer, AIDS-defining cancers (ADC), non-AIDS-defining cancers (NADC), any infection-related cancer, and common individual cancer types. We estimated adjusted hazard ratios (aHR) comparing cancer risk by ART class using marginal structural models emulating ITT and PP trials. RESULTS: We observed 17,004 PWH (954 cancers) with PI-based (median 6 years follow-up), 17,536 (770 cancers) with NNRTI-based (median 5 years follow-up) and 1,782 (29 cancers) with INSTI-based ART (median 2 years follow-up). Analyses with 85 months follow-up indicated no cancer risk differences. In truncated analyses, risk of ADCs (aHR 1.33; 95% CI 1.00, 1.77 [PP-analysis]) and NADCs (aHR 1.23; 95% CI 1.00, 1.51[ITT-analysis]) were higher comparing PIs vs. NNRTIs. CONCLUSIONS: Results with longer-term follow-up suggest being on a PI- versus NNRTI-based ART regimen does not affect cancer risk. We observed shorter-term associations that should be interpreted cautiously and warrant further study. Further research with longer duration of follow-up that can evaluate INSTIs, the current first-line recommended therapy, is needed to comprehensively characterize the association between ART class and cancer risk.