Neurology Clinical Trial Database
SELECT
Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes
Year of Publication: 2023
Authors: A. Michael Lincoff, M.D., Kirstine Brown-Frandsen, ..., and Donna H. Ryan
Journal: The New England Journal of Medicine
Citation: N Engl J Med 2023;389:2221-32.
PDF: Clinical Question
In patients with overweight or obesity and preexisting cardiovascular disease but without diabetes, is treatment with semaglutide superior to placebo in reducing the risk of major adverse cardiovascular events?
Study Overview
Objective
Semaglutide - To evaluate whether semaglutide reduces major adverse cardiovascular events (MACE), including stroke, in overweight or obese patients without diabetes but with preexisting cardiovascular disease.
Study Summary
- In patients with obesity but no diabetes, semaglutide significantly reduced major cardiovascular events.
- Stroke risk was modestly reduced, though the difference was not statistically significant.
Intervention
Multicenter, double-blind, placebo-controlled, event-driven superiority trial enrolling 17,604 patients ≥45 years with BMI ≥27 and preexisting cardiovascular disease but no diabetes. Patients were randomized 1:1 to weekly subcutaneous semaglutide 2.4 mg or placebo. Median follow-up was 39.8 months.
Patients per Arm
Semaglutide: 8803; Placebo: 8801
Bottom Line
In patients with preexisting cardiovascular disease and overweight or obesity but without diabetes, once-weekly subcutaneous semaglutide at a dose of 2.4 mg was superior to placebo in reducing the incidence of the primary composite endpoint of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke.
Major Points
- SELECT is the landmark trial demonstrating that GLP-1 receptor agonists reduce cardiovascular events in patients WITHOUT diabetes — the first trial to show CV benefit of an obesity medication in this population. Underlying mechanisms of the benefit remain speculative per the authors.
- 17,604 patients across 804 centers in 41 countries — the largest GLP-1 RA cardiovascular outcome trial. BMI ≥27 with established CVD but NO diabetes (HbA1c <6.5%). Mean follow-up 39.8±9.4 months.
- 20% reduction in 3-point MACE (CV death, nonfatal MI, nonfatal stroke): 6.5% vs 8.0% (HR 0.80, 95% CI 0.72–0.90, p<0.001). NNT = 67 over 3.3 years.
- Nonfatal stroke: 1.7% vs 1.9% (HR 0.93, 95% CI 0.74–1.15) — numerically lower but NOT statistically significant; nonfatal MI drove most of the primary composite reduction (HR 0.72, 0.61–0.85).
- All-cause mortality: HR 0.81 (95% CI 0.71–0.93) and CV death: HR 0.85 (0.71–1.01, p=0.07) narrowly missed the prespecified significance threshold in hierarchical testing.
- Heart failure composite HR 0.82 (95% CI 0.71–0.96) — however, because CV death did not meet the required nominal significance, superiority testing was not performed for the heart failure composite; per the authors, the confidence intervals should not be used to infer definitive treatment effects.
- Semaglutide produced ~9.4% weight loss (vs 0.9% placebo) plus reduced inflammation (hsCRP ↓~38%), improved lipids, and lowered BP. Whether benefit is mediated by weight loss, glycemic improvement, or direct vascular/cardiac GLP-1 receptor effects remains speculative — the authors state that the underlying mechanism is not established.
- GI tolerability was the major limitation: 16.6% permanent discontinuation (vs 8.2%) driven by nausea, diarrhea, and vomiting. ~10% stopped specifically for GI disorders. Dose escalation over 16 weeks mitigated but did not eliminate this.
- 66% of patients had prediabetes (HbA1c ≥5.7%) — SELECT may partly represent diabetes prevention rather than pure obesity treatment. A prespecified analysis showed similar benefit regardless of glycemic status.
- Together with STEP-HFpEF, SURMOUNT, and FLOW, SELECT positioned semaglutide as a transformative cardiometabolic drug — driving the 'GLP-1 revolution' in preventive cardiology and raising questions about population-level obesity pharmacotherapy for CV risk reduction.
Design
Study Type: Multicenter, double-blind, randomized, placebo-controlled, event-driven superiority trial.
Randomization: 1
Blinding: Double-blind.
Enrollment Period: October 2018 through March 2021.
Follow-up Duration: Mean of 39.8±9.4 months.
Centers: 804
Countries: 41 countries
Sample Size: 17604
Analysis: Intention-to-treat.
Inclusion Criteria
- Age 45 years or older.
- Preexisting cardiovascular disease (defined as previous myocardial infarction, previous stroke, or symptomatic peripheral arterial disease).
- Body-mass index (BMI) of 27 or greater.
- No history of diabetes (glycated hemoglobin <6.5% at screening and no treatment with glucose-lowering medication within the previous 90 days).
Exclusion Criteria
- Previous diagnosis of diabetes or glycated hemoglobin ≥6.5%.
- Treatment with any glucose-lowering medication or GLP-1 receptor agonist within the previous 90 days.
- New York Heart Association class IV heart failure.
- End-stage kidney disease or dialysis.
- Cardiovascular or neurologic event within 60 days of enrollment.
- Planned coronary, carotid, or peripheral revascularization.
Baseline Characteristics
| Characteristic | Control | Active |
|---|---|---|
| Age-yr | 61.6±8.8 | 61.6±8.9 |
| Male sex no. (%) | 6377 (72.5) | 6355 (72.2) |
| White race no. (%) | 7404 (84.1) | 7387 (83.9) |
| BMI | 33.4±5.0 | 33.3±5.0 |
| Waist circumference - cm | 111.4±13.1 | 111.3±13.1 |
| Glycated hemoglobin level-% | 5.78±0.33 | 5.78±0.34 |
| Prediabetes (HbA1c ≥5.7%) - no. (%) | 5819 (66.1) | 5877 (66.8) |
| Previous Myocardial Infarction only - no. (%) | 5944 (67.5) | 5962 (67.7) |
| Previous Stroke only - no. (%) | 1556 (17.7) | 1578 (17.9) |
| eGFR-ml/min/1.73 m² | 82.5±17.3 | 82.4±17.5 |
| Median LDL cholesterol level (IQR)-mg/dl | 78 (61-102) | 78 (61-102) |
| Median Triglyceride level (IQR) - mg/dl | 135 (100-190) | 134 (99-188) |
| Systolic blood pressure - mm Hg | 130.9±15.3 | 131.0±15.6 |
Arms
| Field | Control | Semaglutide 2.4 mg |
|---|---|---|
| Intervention | Matched placebo injection once weekly, subcutaneous, same volume and schedule as semaglutide. Same dose-escalation regimen to maintain blinding. | Subcutaneous semaglutide once weekly with dose escalation over 16 weeks: 0.24 mg (weeks 1–4) → 0.50 mg (weeks 5–8) → 1.0 mg (weeks 9–12) → 1.7 mg (weeks 13–16) → 2.4 mg (week 17+). Semaglutide is a GLP-1 receptor agonist (94% amino acid homology to native GLP-1) with a half-life of ~1 week (albumin binding). Mechanism: reduces appetite via hypothalamic satiety centers, slows gastric emptying, reduces hepatic glucose output, and has direct anti-inflammatory effects on vascular endothelium and cardiomyocytes via GLP-1R. |
| Duration | Mean duration of exposure was 35.1±13.0 months. | Mean duration of exposure was 33.3±14.4 months. |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| A composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke. | Primary | 8.0% (701/8801) | 6.5% (569/8803) | 0.8 | <0.001 |
| Death from cardiovascular causes | Secondary | 3.0% (262/8801) | 2.5% (223/8803) | HR 0.85 (95% CI, 0.71 to 1.01) | 0.07 |
| Heart failure composite end point (CV death, hospitalization, or urgent visit for heart failure) | Secondary | 4.1% (361/8801) | 3.4% (300/8803) | HR 0.82 (95% CI, 0.71 to 0.96) | NA (superiority testing not performed after CV death missed hierarchical significance) |
| Death from any cause | Secondary | 5.2% (458/8801) | 4.3% (375/8803) | HR 0.81 (95% CI, 0.71 to 0.93) | NA (superiority testing not performed after CV death missed hierarchical significance) |
| Serious adverse events | Adverse | 36.4% (3204/8801) | 33.4% (2941/8803) | <0.001 | |
| Adverse events leading to permanent discontinuation | Adverse | 8.2% (718/8801) | 16.6% (1461/8803) | <0.001 | |
| Discontinuation due to gastrointestinal disorders | Adverse | 2.0% (172/8801) | 10.0% (880/8803) | <0.001 | |
| Gallbladder-related disorders | Adverse | 2.3% (203/8801) | 2.8% (246/8803) | 0.04 |
Criticisms
- Secondary prevention only (established CVD required) — cannot be extrapolated to primary prevention. Whether semaglutide benefits obese patients WITHOUT prior CVD remains unknown from SELECT alone.
- Limited diversity: 84% White, 72% male — markedly underrepresents women, Black, and Hispanic populations who have high obesity and CVD burden. Generalizability to these groups is uncertain.
- High discontinuation rate (16.6% vs 8.2%) driven by GI side effects — the as-treated benefit may be larger than ITT, but real-world adherence is likely even worse than in the trial setting.
- Industry-sponsored by Novo Nordisk (sole manufacturer of semaglutide) — raises concerns about study design, conduct, and interpretation bias. Stock price doubled after results, creating substantial commercial incentive.
- 66% prediabetic — SELECT may partly represent diabetes prevention rather than pure obesity treatment. The trial does not cleanly separate the CV benefits of weight loss vs glycemic improvement vs anti-inflammatory effects.
- Hierarchical testing stopped after death from cardiovascular causes did not meet the prespecified nominal significance threshold — as a result, superiority testing was not performed for the heart failure composite or all-cause mortality, and P values for these outcomes are not reported. This limits the strength of any mortality or heart failure claims.
- No comparison to lifestyle intervention, bariatric surgery, or other GLP-1 agonists — the incremental benefit over optimized non-pharmacologic weight management is unknown.
- Cost-effectiveness concerns: semaglutide ~$1,000+/month in many markets. Population-level use for CV prevention would have enormous healthcare cost implications that were not addressed.
- Unknown duration of benefit: does CV protection persist if weight is regained after stopping semaglutide? Post-trial data suggest rapid weight regain, raising questions about indefinite treatment duration.
Funding
Novo Nordisk
Based on: SELECT (The New England Journal of Medicine, 2023)
Authors: A. Michael Lincoff, M.D., Kirstine Brown-Frandsen, ..., and Donna H. Ryan
Citation: N Engl J Med 2023;389:2221-32.
Content summarized and formatted by NeuroTrials.ai.