Integrase strand transfer inhibitor (INSTI)-based regimens have been implicated in greater weight gain in antiretroviral therapy (ART)-naïve HIV+ persons starting ART, though metabolic consequences are unclear. We examined the impact of initial ART regimen class on incident diabetes mellitus (DM) and potential mediation of this effect by weight change in a large North American HIV cohort.
We included treatment-naïve adults (≥18 years) initiating INSTI-, protease inhibitor (PI)-, or non-nucleoside reverse transcriptase inhibitor (NNRTI)-based ART from 01/2007-12/2016 with 12-month (±6 months) weights in the North American AIDS Cohort Collaboration on Research and Design (NA-ACCORD). We followed individuals until incident DM (HgA1c >6.5%, initiation of diabetes-specific medication, or DM diagnosis along with diabetes-related medication, precluding prevalent DM or pre-diabetes), virologic failure (≥400 copies/mL), regimen core switch, administrative close, death, or loss to follow-up (≥12 months with no visit or lab before cohort close). We excluded those with incident DM before 12-month weight measure, and we multiply imputed missing baseline data. Cox regression stratified by clinic site and adjusting for age, sex, race, HIV transmission risk, year of ART initiation, and baseline weight, CD4+ cell count, and HIV-1 RNA yielded adjusted hazard ratios (HR) and 95% confidence intervals (CI) for incident DM by ART class. We conducted mediation analysis including 12-month weights along with all covariates from the primary analysis.
Among 16,305 eligible ART initiators, 8,082 (50%) started NNRTIs, 5,152 (32%) PIs, and 3,071 (19%) INSTIs, with median follow-up of 3.3, 2.8, and 2.1 years, respectively. Among INSTI initiators, 18% started dolutegravir (DTG), 30% raltegravir (RAL), and 52% elvitegravir (EVG). Overall, 333 (2%) developed DM. Tenofovir alafenamide (TAF) was part of <1% of regimens. Those starting INSTIs vs. NNRTIs had elevated incident DM risk (HR=1.30; CI: 0.89-1.90), greater than PI- vs. NNRTI-initiators (HR=1.07; CI: 0.83-1.38). Mediation analysis revealed an INSTI-DM association attenuated 5% (HR=1.24; CI: 0.85-1.81) by including 12-month weight in the full model (Figure).
Initiating ART with INSTI- vs. NNRTI-based regimens may confer greater risk of incident DM, and this risk is likely only partially due to 12-month weight gain. Research to elucidate metabolic changes after INSTI initiation and identify interventions to mitigate them continues.