Long-term weight loss effects of semaglutide in obesity without diabetes in the SELECT trial

The worldwide obesity prevalence (BMI ≥30 kg m−2) has nearly tripled since 1975. Although BMI correlates with body fat at a population level, it may not accurately indicate amount or distribution of body fat in individuals; visceral and ectopic adiposity drive cardiovascular disease and other chronic conditions. Reducing excess abnormal body fat through weight loss is a public health priority; modest weight loss (~5%) improves cardiovascular risk factors, glycemia and quality-of-life measures, with greater losses yielding more benefit. Producing and sustaining durable, clinically significant weight loss with lifestyle alone is challenging, but anti-obesity medications—particularly GLP-1 receptor agonists such as once-weekly subcutaneous semaglutide 2.4 mg—enable greater and more sustainable weight loss. Semaglutide has also reduced cardiovascular events in patients with type 2 diabetes at lower doses. SELECT evaluated adults with established CVD and overweight/obesity without diabetes to assess semaglutide’s effects on cardiovascular outcomes and, in this prespecified analysis, weight, anthropometrics (BMI, waist circumference, waist-to-height ratio), safety, and tolerability across baseline BMI categories.

Design and participants: Prespecified analysis of the randomized, double-blind, placebo-controlled SELECT trial (NCT03574597), conducted in 41 countries. Adults ≥45 years with BMI ≥27 kg/m2, established CVD (prior MI, stroke, or symptomatic PAD), and no diabetes were eligible. All participants provided written informed consent; protocols were approved by national and institutional authorities. Intervention: Participants were randomized 1:1 to once-weekly subcutaneous semaglutide 2.4 mg or matching placebo. Dosing started at 0.24 mg weekly with 4-week escalations (0.5, 1.0, 1.7, to 2.4 mg by week 16). Dose escalation could be slowed, paused, or maintained below target for tolerability; dose reductions were allowed. All patients received standard-of-care management for CVD risk factors and healthy lifestyle counseling, not specifically targeted at weight loss. Measurements: Sex and race were self-reported. Body weight was measured on calibrated digital scales (light clothing, no shoes). Height was measured without shoes. Waist circumference (WC) was measured midway between the lower rib margin and iliac crest with a nonstretch tape. BMI was calculated at screening. Waist-to-height ratio (WHtR) was derived from WC and height. Endpoints: Assessed at baseline and at years 2, 3, and 4: percentage change in body weight; proportions achieving ≥5%, ≥10%, ≥15%, and ≥20% weight loss; change in WC (cm); percentage change in WHtR. Improvement in BMI category (healthy <25, overweight 25–<30, class I 30–<35, class II 35–<40, class III ≥40 kg/m2) was assessed at week 104. Achievement of sex- and race-specific WC risk cutoffs (Asian women <80 cm; non-Asian women <88 cm; Asian men <88 cm; non-Asian men <102 cm) was evaluated at week 104, stratified by baseline BMI <35 or ≥35 kg/m2. Follow-up and adherence: Overall, 97.1% (semaglutide) and 96.8% (placebo) completed the trial; 30.6% and 27.0%, respectively, did not complete study drug. A first on-treatment analysis censored follow-up at >35 days off treatment to assess weight change among those adhering. Statistical analysis: In-trial analyses followed intention-to-treat, including all randomized patients irrespective of adherence. Continuous endpoints were analyzed by ANCOVA with treatment as a fixed factor and baseline value as covariate. Missing data at landmark visits (e.g., week 104) were imputed via multiple imputation (separately by arm), generating 500 datasets; Rubin’s rules were used. Binary endpoints used logistic regression with treatment and baseline value as covariates, after multiple imputation of missing outcomes. Subgroup analyses included subgroup and treatment-by-subgroup interaction terms. Confidence intervals were not adjusted for multiplicity. Analyses used SAS 9.4 TS1M5.

- Population: 17,604 patients (72.3% male), mean age 61.6 years, mean BMI 33.3 kg/m2; enrolled Oct 2018–Mar 2021 across 41 countries. - Overall weight trajectory: With semaglutide, mean weight loss continued through week 65 and was sustained to week 208. - Mean change at 208 weeks (in-trial): body weight −10.2% with semaglutide vs −1.5% with placebo; estimated treatment difference (ETD) −8.7 percentage points (95% CI −9.42 to −7.88; P<0.0001). First on-treatment analysis: −11.7% vs −1.5%; ETD −10.2 points (95% CI −11.0 to −9.42; P<0.0001). - Waist circumference at 208 weeks: −7.7 cm with semaglutide vs −1.3 cm placebo; ETD −6.4 cm (95% CI −7.18 to −5.61; P<0.0001). - Waist-to-height ratio at 208 weeks: relative reduction −6.9% with semaglutide vs −1.0% with placebo; ETD −5.87 percentage points (95% CI −6.56 to −5.17; P<0.0001). - Categorical weight loss at week 104 (in-trial): With semaglutide, 67.8% achieved ≥5%, 44.2% ≥10%, 22.9% ≥15%, 11.0% ≥20%, and 4.9% ≥25% loss vs placebo 21.3%, 6.9%, 1.7%, 0.6%, and 0.1%, respectively (P<0.001). - BMI category improvement at week 104: 52.4% with semaglutide vs 15.7% with placebo improved category; 12.0% vs 1.2% achieved healthy BMI (<25 kg/m2). Proportion with BMI ≥30 kg/m2 fell from 71.0% to 43.3% with semaglutide vs 71.9% to 67.9% with placebo. - Sex- and race-specific WC cutoff achievement (baseline BMI <35 kg/m2): Below cutoff at baseline 15.0% (semaglutide) and 14.3% (placebo); at week 104, 41.2% (semaglutide) vs 18.0% (placebo) below cutoff. - Subgroups (week 104): Weight and WC reductions favored semaglutide across prespecified subgroups. Women had greater mean weight loss difference than men (−11.1% vs −7.5% vs placebo; P<0.0001 for interaction). Younger age associated with slightly greater loss; higher baseline BMI associated with slightly greater percentage loss. Asian race/region showed slightly smaller ETDs (race −7.27%; region Asia −7.30%). No significant differences by ethnicity, glycemic status, or renal function. - Safety: Serious adverse events (SAEs) occurred in 33.4% (semaglutide) vs 36.4% (placebo) overall (P<0.001). SAE rates (events per 100 patient-years) by BMI class for semaglutide vs placebo: <30: 43.23 vs 50.48; 30–<35: 43.54 vs 49.66; 35–<40: 51.07 vs 52.73; ≥40: 47.06 vs 60.85; no heterogeneity detected. Hepatobiliary and gastrointestinal SAEs did not differ between arms within BMI classes. - Tolerability: Adverse events leading to permanent discontinuation: 16.6% (semaglutide) vs 8.2% (placebo) (P<0.001). Discontinuations with semaglutide increased as baseline BMI category decreased. Cumulative incidence plots showed graded increase in discontinuation for semaglutide across time and lower BMI classes. - Cardiovascular context: As previously reported in SELECT, semaglutide reduced major adverse cardiovascular events by 20% (HR 0.80; 95% CI 0.72–0.90; P<0.001).

This prespecified SELECT analysis demonstrates that once-weekly subcutaneous semaglutide 2.4 mg produces substantial, durable weight loss, with continued reduction to week 65 and maintenance through 4 years, alongside favorable changes in waist circumference and waist-to-height ratio. These anthropometric improvements translated into shifts to lower BMI categories and a higher proportion of patients falling below sex- and race-specific waist circumference risk thresholds. Clinically meaningful weight loss was observed consistently across sexes, races, body sizes, geographic regions, glycemic states, and renal function levels, with somewhat greater relative benefits in women and in those with higher baseline BMI or WHtR. Safety was favorable: semaglutide was associated with fewer SAEs across BMI classes, primarily driven by fewer cardiovascular events and infections, and no increase in hepatobiliary or gastrointestinal SAEs by BMI class. Tolerability varied by BMI, with higher discontinuation rates in lower BMI categories, possibly reflecting higher effective drug exposure. Together with the established 20% MACE reduction in SELECT, these data support semaglutide 2.4 mg as a safe, durable weight-management therapy that could materially reduce the public health burden associated with obesity and cardiometabolic disease.

In adults with established CVD and overweight or obesity but without diabetes, semaglutide 2.4 mg once weekly produced clinically significant, durable weight loss and improvements in waist circumference and waist-to-height ratio over up to 4 years, with consistent benefits across diverse subgroups and a favorable serious adverse event profile. These findings, the longest trial data for semaglutide’s weight effects to date, support broad use of semaglutide for weight management and cardiovascular risk reduction in this population. Future research should assess cardiovascular risk reduction and weight-management efficacy in populations not included in SELECT (e.g., Asian individuals with BMI <27 kg/m2) and incorporate detailed body composition measures (fat mass, lean mass, muscle mass) to better understand physiological changes underlying weight loss.

- SELECT was a secondary prevention cardiovascular outcomes trial; findings should not be extrapolated to primary prevention or to all individuals with overweight/obesity. - Limited sample sizes within some racial/ethnic subgroups may have reduced power to detect differential effects. - Individuals with BMI <27 kg/m2 (including some Asian populations qualifying for treatment at lower BMI/WC thresholds) were not included; results may not generalize to these groups. - Lack of direct body composition assessments (fat and lean mass) limits insight into compartmental changes. - Greater discontinuation rates in lower BMI categories may reflect higher drug exposure or other unmeasured factors; mechanisms require further study. - Percentage weight loss was smaller in BMI <30 kg/m2 than in higher BMI classes, potentially reflecting proximity to a physiologic weight ‘settling point’; implications warrant investigation.