Semaglutide Weight Loss Evidence: A Detailed Review

The evidence supporting semaglutide for weight loss derives from multiple sources: (1) rigorous Phase 3 randomised controlled trials (STEP 1–4 program and SELECT); (2) comparative observational data from health systems and registries; (3) mechanistic studies in animal models and human neuroimaging; and (4) post-hoc analyses of trial datasets evaluating subgroups, body composition, and durability. This hierarchy of evidence provides robust support for semaglutide's weight loss efficacy, with randomised trial data representing the gold standard for efficacy claims.

The volume of evidence is substantial: the STEP program alone enrolled >4,500 participants, while SELECT enrolled >17,600. Meta-analytical synthesis of these trials provides high-confidence estimates of average weight loss and response rate distributions. Furthermore, observational data from real-world populations (non-trial settings with diverse demographics, comorbidities, and treatment adherence patterns) helps contextualize trial findings and identify potential predictors of response.

The STEP 1 trial (primary weight management indication) enrolled >1,900 participants and demonstrated mean weight loss of 15.3% with semaglutide 2.4 mg versus 2.6% with placebo—a 12.7 percentage point difference. Response rates were robust: 86% achieved ≥5% weight loss, 69% achieved ≥10%, and 35% achieved ≥15%. STEP 2 (type 2 diabetes) showed slightly lower average weight loss (10.5% vs. 3.2% placebo) due to baseline metabolic differences. STEP 3 and 4 showed consistent efficacy. SELECT, powered for cardiovascular outcomes rather than maximal weight loss, demonstrated 10.4% weight loss with semaglutide versus 1.5% placebo.

Meta-analyses of all STEP trials show mean weight loss with semaglutide 2.4 mg of approximately 12–15% across diverse populations. By comparison, prior weight loss medications (phentermine, orlistat, phendimetrazine) achieved mean weight losses of approximately 3–5%. Bariatric surgery (gastric bypass, sleeve gastrectomy) achieves 20–35% weight loss but involves surgical risk. Semaglutide thus occupies an intermediate position—substantially more effective than prior medical management but less dramatic than surgery, with a favourable safety profile.

Body composition analyses from STEP trials using dual-energy X-ray absorptiometry demonstrated that semaglutide-induced weight loss is predominantly from fat mass (approximately 80% of total weight loss), with relative preservation of lean mass (approximately 20% of weight loss). This is favourable compared to caloric restriction alone, which typically produces 60–70% fat loss and 30–40% lean loss. The preservation of lean mass during semaglutide treatment supports sustained resting energy expenditure and may reduce the metabolic adaptation that complicates long-term weight maintenance.

Secondary metabolic endpoints improved with weight loss: fasting glucose, insulin levels, HOMA-IR (an insulin resistance marker), and inflammatory markers (C-reactive protein) all decreased. Cardiovascular risk factors improved: systolic and diastolic blood pressure decreased by 2–4 mmHg, LDL cholesterol by 5–10%, and triglycerides by 10–15%. These cardiometabolic improvements likely exceed those expected from weight loss alone, suggesting direct metabolic benefits of GLP-1 agonism. Physical function assessments showed improvements with weight loss, including increased 6-minute walk test distance and reduced dyspnoea.

Post-hoc subgroup analyses of STEP and SELECT trials evaluated whether efficacy varied by baseline characteristics. Weight loss was generally consistent across age groups (though slightly lower in older adults), gender (no substantial differences), ethnicity (though trial populations were predominantly white), and baseline BMI (efficacy was preserved in very obese populations with BMI >40 kg/m²). HbA1c at baseline did not substantially predict weight loss response. However, some analyses suggested that baseline appetite dysregulation and certain genetic polymorphisms in GLP-1 receptor expression might predict superior response, though these findings require confirmation.

Response heterogeneity is notable: approximately 20–30% of trial participants achieved minimal weight loss (<5%), while approximately 10–15% achieved exceptional weight loss (>20%). The basis for this heterogeneity—whether genetic, microbiotal, psychological, or metabolic—remains incompletely understood. Identifying predictive biomarkers of response is an active research area with important implications for precision medicine approaches to obesity treatment.

While the STEP and SELECT trials demonstrated sustained weight loss during active treatment (68–104 weeks), evidence on long-term durability beyond the trial period is more limited. Observational data from clinical practice show that weight regain occurs after semaglutide discontinuation, with most patients regaining 50–70% of lost weight within 1–2 years. This pattern suggests that semaglutide is not a 'cure' for obesity but rather a treatment requiring ongoing use to maintain weight loss. Some evidence suggests that early, intensive lifestyle intervention coupled with semaglutide may enable better weight maintenance after medication discontinuation, but rigorous data are lacking.

Discontinuation rates due to adverse effects, particularly gastrointestinal intolerance, range from 4–8% in trials but may be higher in real-world settings with less intensive support. Additionally, cost barriers and access limitations in many healthcare systems limit real-world adoption. These practical considerations are important for evaluating the population-level impact of semaglutide as a weight loss intervention.