The SELECT trial demonstrated that semaglutide 2.4 mg weekly reduced the three-point MACE composite by 20% versus placebo in adults with overweight or obesity and established CVD but without diabetes. The hazard ratio was statistically significant at HR 0·80 and the absolute risk reduction was approximately one and a half percentage points over a mean follow-up of approximately 40 months.
What Was the SELECT Trial Design and Who Was Enrolled?
SELECT was a randomised double-blind placebo-controlled event-driven cardiovascular outcomes trial enrolling 17,604 adults aged 45 years or older with a BMI of at least 27 kg/m² and established atherosclerotic CVD without diabetes. Participants were randomised 1:1 to once-weekly subcutaneous semaglutide 2.4 mg or matched placebo titrated over 16 weeks.
The trial was conducted across 804 sites in 41 countries and was powered at 90% to detect a relative risk reduction of 17% in the primary endpoint. Established CVD was defined as prior myocardial infarction or prior stroke or symptomatic peripheral arterial disease.
Mean baseline BMI was approximately 33 kg/m² and mean age was approximately 62 years and approximately 72% of participants were male. The primary composite endpoint comprised cardiovascular death and nonfatal MI and nonfatal stroke. Secondary endpoints included each MACE component individually and a heart failure composite and all-cause mortality.
The trial was published in the New England Journal of Medicine in November 2023 by Lincoff and colleagues. It required 1225 primary endpoint events to trigger the final analysis and provide the pre-specified statistical power for the superiority test.
What Was the Exact Magnitude of the Primary MACE Reduction?
A primary endpoint event occurred in 6·5% of patients on semaglutide versus 8·0% of patients on placebo across the 17,604-person trial population. The hazard ratio of HR 0·80 was statistically significant and the absolute risk reduction of approximately one and a half percentage points yields a number needed to treat of roughly 67 over the trial period.
The Kaplan–Meier curves began to diverge at approximately 6 months and continued to separate throughout the observation period. This pattern suggests a sustained rather than early-only benefit and the confidence interval excludes unity.
The relative risk reduction of 20% was consistent across the prespecified primary analysis. It was not materially altered by sensitivity analyses. The trial required 1225 primary endpoint events to provide the pre-specified statistical power.
How Did Secondary Cardiovascular Endpoints Perform?
Beyond the primary MACE composite semaglutide reduced the heart failure composite endpoint and all-cause mortality and nonfatal MI and nonfatal stroke with hazard ratios of HR 0·82 and HR 0·81 and HR 0·72 and HR 0·67 respectively. Cardiovascular death showed a directionally consistent but nominally non-significant reduction.
The hierarchical testing procedure preserved the family-wise error rate across secondary endpoints. All-cause mortality reaching statistical significance is a notably high evidentiary bar for a cardiovascular outcomes trial in this population. The heart failure composite included urgent heart failure visits and hospitalisations.
A subsequent JAMA Cardiology analysis confirmed that semaglutide significantly reduced total cardiovascular hospitalisations and not merely first events. This is relevant to the overall burden of disease in a population with established CVD. These findings indicate broad cardiovascular benefit rather than an effect confined to a single event type.
Is the Cardiovascular Benefit Independent of Weight Loss?
Pre-specified mediation analyses indicate that the MACE reduction was only partially explained by weight loss and that weight reduction accounted for less than half of the observed cardiovascular benefit. Residual benefit is attributed to direct GLP-1 receptor-mediated effects on inflammation and endothelial function and atherosclerotic plaque biology independent of adiposity reduction.
Semaglutide reduced high-sensitivity C-reactive protein independently of the degree of weight loss achieved as demonstrated in substudy analyses published in JACC in 2024. GLP-1 receptors are expressed on macrophages and endothelial cells and cardiomyocytes providing mechanistic plausibility for direct vascular effects. Macrophage polarisation from pro-inflammatory M1 to anti-inflammatory M2 phenotype has been documented in preclinical models of GLP-1 receptor agonism.
Subgroup analyses stratified by weight-loss quartile showed consistent MACE reduction across all four quartiles including patients who lost minimal weight. This weight-loss-independent signal distinguishes semaglutide's cardiovascular profile from that of purely weight-loss-mediated interventions. It has direct implications for how the mechanism is interpreted in clinical and research contexts.
Was the Benefit Consistent Across Patient Subgroups?
The MACE reduction in SELECT was broadly consistent across prespecified subgroups including age and sex and geographic region and BMI category and baseline HbA1c and type of qualifying CVD event and background statin use. No subgroup demonstrated a statistically significant interaction with treatment effect supporting generalisability of the primary result across the enrolled population.
Patients with prior MI and prior stroke and symptomatic peripheral arterial disease all showed directionally consistent hazard ratios below unity. The benefit was observed in both sexes though the trial enrolled approximately 72% male participants limiting precision of the female-specific estimate.
Age subgroups below and above 65 years showed similar relative risk reductions. This is clinically relevant because older patients with established CVD carry the highest absolute event rates and therefore derive the greatest absolute benefit per treated patient. These data support the robustness of the trial's primary finding.
What Does the SELECT Safety Data Show for This Population?
Overall serious adverse events were significantly less frequent in the semaglutide group than in the placebo group at approximately 33% versus 36%. However adverse events leading to permanent discontinuation occurred more than twice as often with semaglutide at approximately 17% versus 8% driven predominantly by gastrointestinal events during the dose-escalation phase.
Rates of acute pancreatitis and gallbladder disease and diabetic retinopathy were numerically higher in the semaglutide group consistent with the known pharmacological profile of GLP-1 receptor agonists. The SELECT protocol used a 16-week titration schedule to mitigate gastrointestinal burden. No new cardiac safety signals emerged from the trial data.
Heart rate increased modestly with semaglutide at approximately 1 to 2 beats per minute which is a known class effect of GLP-1 receptor agonists. Hypoglycaemia rates were low in both arms given the non-diabetic population. The approximately 17% discontinuation rate for adverse events means that real-world persistence may substantially influence population-level cardiovascular benefit.
Clinicians should evaluate individual gastrointestinal risk and gallbladder history and personal or family history of medullary thyroid carcinoma or multiple endocrine neoplasia type 2 which remain formal contraindications. Renal function monitoring is advisable in patients with pre-existing chronic kidney disease given dehydration risk from gastrointestinal side effects.
How Does This Evidence Compare With Tirzepatide's Cardiovascular Outcomes Data?
As of 2026 semaglutide 2.4 mg holds the only completed randomised cardiovascular outcomes trial in non-diabetic patients with established CVD and obesity with a confirmed hazard ratio of HR 0·80 for MACE. Tirzepatide's dedicated cardiovascular outcomes trial SURMOUNT-MMO is ongoing and has not yet reported primary MACE endpoint data making a direct comparison impossible.
SURMOUNT-MMO is a randomised double-blind event-driven trial designed to assess tirzepatide's effect on cardiovascular morbidity and mortality in adults with obesity but without diabetes mirroring the SELECT eligibility framework. Estimated completion dates project into 2027. SELECT therefore remains the sole RCT-level evidence base for MACE reduction with a GLP-1 receptor agonist in this specific population.
Post-hoc modelling using SURMOUNT-5 data suggested tirzepatide may produce greater predicted 10-year CVD risk reduction than semaglutide based on surrogate endpoints. However surrogate-based projections do not substitute for event-driven outcomes data. Until SURMOUNT-MMO reports SELECT's hazard ratio of HR 0·80 stands as the reference-standard cardiovascular outcome for this drug class in non-diabetic obesity with established CVD. How Do You Cycle GH Peptides Without Crashing Endogenous Production in 2026? What Does 2026 Research Reveal About BPC-157 for Musculoskeletal Healing — Regeneration or Risk? What Does the 2026 Clinical Evidence Actually Show for BPC-157 in Shoulder Rotator Cuff Tears?