Incidence of cardiometabolic outcomes among people living with HIV‐1 initiated on integrase strand transfer inhibitor versus non‐integrase strand transfer inhibitor antiretroviral therapies: a retrospective analysis of insurance claims in the United States

Introduction Integrase strand transfer inhibitor (INSTI)‐containing antiretroviral therapy (ART) has been associated with weight gain, though there is limited information on associations between ART‐related weight gain and cardiometabolic outcomes among people living with HIV‐1 (PLWH). We, therefore, evaluated risks of incident cardiometabolic outcomes following INSTI versus non‐INSTI‐based ART initiation in the United States. Methods We conducted a retrospective study using IBM MarketScan Research Databases (12 August 2012−31 January 2021). Treatment‐naïve PLWH initiating ART (index date) on/after 12 August 2013 (approval date of the first second‐generation INSTI, dolutegravir) were included and censored at regimen switch/discontinuation, end of insurance eligibility or end of data availability. We used inverse probability of treatment weights constructed with baseline (12 months pre‐index) characteristics to account for differences between INSTI‐ and non‐INSTI‐initiating cohorts. Doubly robust hazard ratios (HRs) obtained from weighted multivariable Cox regression were used to compare time to incident cardiometabolic outcomes (congestive heart failure [CHF], coronary artery disease, myocardial infarction, stroke/transient ischemic attack, hypertension, type II diabetes, lipid disorders, lipodystrophy and metabolic syndrome) by INSTI‐initiation status. Results Weighted INSTI (mean age = 39 years, 23% female, 70% commercially insured, 30% Medicaid insured) and non‐INSTI (mean age = 39 years, 24% female, 71% commercially insured, 29% Medicaid insured) cohorts included 7059 and 7017 PLWH, respectively. The most common INSTI‐containing regimens were elvitegravir‐based (43.4%), dolutegravir‐based (33.3%) and bictegravir‐based (18.4%); the most common non‐INSTI‐containing regimens were darunavir‐based (31.5%), rilpivirine‐based (30.4%) and efavirenz‐based (28.3%). Mean±standard deviation follow‐up periods were 1.5±1.5 and 1.1±1.2 years in INSTI‐ and non‐INSTI‐initiating cohorts, respectively. INSTI initiators were at a clinically and significantly increased risk of experiencing incident CHF (HR = 2.12, 95% confidence interval [CI] = 1.08−4.05; p = 0.036), myocardial infarction (HR = 1.79, 95% CI = 1.03−5.65; p = 0.036) and lipid disorders (HR = 1.26, 95% CI = 1.04−1.58; p = 0.020); there was no evidence of an increased risk for other individual or composite outcomes. Conclusions Over a short average follow‐up period of