Semaglutide and cardiovascular risk reduction: a mediation analysis of the SELECT trial.

BACKGROUND AND AIMS: In SELECT, subcutaneous semaglutide (target dose 2.4 mg) weekly in adults with pre-existing cardiovascular disease and body mass index ≥27 kg/m2 without diabetes significantly reduced major adverse cardiovascular events (MACE) by 20% vs placebo. This analysis explored potential mediators of this cardiovascular benefit. METHODS: Changes from randomization to 24 months in body weight, waist circumference, plasma high-sensitivity C-reactive protein (hsCRP), glycated haemoglobin (HbA1c), lipids, blood pressure, estimated glomerular filtration rate, and urinary albumin-to-creatinine ratio were examined individually, and in combination, as potential mediators of MACE risk reduction, using the Vansteelandt repeated regression method (a counterfactual approach). RESULTS: Semaglutide significantly (P < .05) improved all potential mediators investigated. Point estimates for % mediation were largest for waist circumference (64.0% [95% confidence interval (CI) 27.5, 179.6]), hsCRP (42.1% [17.9, 110.5]), HbA1c (29.0% [-20.7, 120.5]), and body weight (19.5% [-33.0, 110.7]), examined separately, but all had wide CIs. Supplementary analyses suggested unreliability of estimates for body weight and waist circumference, potentially due to differing relationship of body weight and waist change to MACE between treatment arms. Multivariable analysis estimated joint mediation for all potential mediators combined to be 31.4% (95% CI -30.1, 143.6). When variables other than body weight and waist circumference were considered, % mediation was 46.0% (95% CI -6.0, 161.0). CONCLUSIONS: Mediation analyses could not fully ascribe the effects of semaglutide on MACE to known risk factors in SELECT with any certainty. Further investigation of potential additional biomarkers and mediators of semaglutide's cardiovascular protective effect are of interest.