SELECT in the light of FAST-MI. Potential population-based benefit of semaglutide 2.4mg/week on mortality, an extrapolation from the SELECT trial results to the FAST-MI registries.
F Schiele, J Ferrieres, N Meneveau, R Cador, G Lemesle, E Puymirat, J Mesnier, V Aboyans, W Amara, T Simon, N DanchinAbstract
Abstract
Rationale: The SELECT trial showed that semaglutide 2.4 mg reduced major adverse cardiovascular events and was associated with lower mortality in overweight/obese patients with cardiovascular disease and no diabetes. Following myocardial infarction (MI), the proportion of real-world patients eligible for SELECT criteria, and the potential population-level impact on mortality of prescribing semaglutide remain unknown.
Aim
(1) determine the % of post-MI patients eligible to SELECT criteria in a large nationwide registry and (2) estimate the potential reduction in all-cause mortality if semaglutide was implemented in such a population, based on the treatment effect in SELECT.
Methods
Data from the nationwide French FAST-MI 2005 2010 and 2015 registries restricted to patients alive at 1 year, with BMI available and no diabetes. Two groups defined according to SELECT inclusion/exclusion criteria: SELECT-like (n=3,088; 40.1%) and Non-SELECT (n=4,060; 59.9%). Baseline characteristics and secondary prevention therapies were compared. Four-year mortality assessed, with Cox models adjusted on relevant variables. To estimate potential population impact, we applied the hazard ratio for all-cause mortality in SELECT (HR=0.78) to the observed mortality risk of the SELECT-like FAST-MI group at 4 years. We used a conservative estimate of incident MIs annually in France. Number needed to treat (NNT) was derived from the absolute risk reduction at 4 years, assuming comparable treatment effect in routine practice and stable baseline mortality risk.
Results
Compared with SELECT-like patients, non-SELECT patients were older (65 vs 62 years), had lower BMI (23 vs 30 kg/m²), smaller waist circumference (91 vs 108 cm), and less hypertension (44 vs 68%), but received similar secondary prevention therapies. Compared with SELECT participants, SELECT-like FAST-MI patients had similar age, sex distribution, and background secondary prevention, but lower BMI. Four-year mortality rates were 7.5% in the SELECT-like group and 9.0% in the non-SELECT group (not significant after Cox adjustment). Cumulative mortality at 48 months was comparable in the FAST-MI SELECT-like group and the placebo arm of SELECT, supporting external validity of the extrapolation (Figure). Assuming 50,000 new MIs per year in France, approximately 12,500 patients (24%) would meet SELECT eligibility criteria. Applying the mortality reduction observed in SELECT (HR=0.78), estimated absolute 4-year reduction in deaths would be 208, resulting NNT: 60 over 48 months.
Conclusions
Nearly one quarter of post-MI patients in routine practice fulfilled SELECT eligibility criteria, with a mortality risk comparable to that observed in the placebo arm of SELECT, supporting the plausibility of effect extrapolation. Use of semaglutide 2.4 mg in stable post-MI patients eligible to SELECT criteria would result in a substantial improvement in survival on a population level.4-year mortality in SELECT and FAST-MI