LEADER, SUSTAIN-6, REWIND, SELECT: A Cross-Trial Comparison of GLP-1 Receptor Agonist Cardiovascular Outcomes

By
Published Updated Last reviewed
GLP-1 medication and metabolic health image for LEADER, SUSTAIN-6, REWIND, SELECT: A Cross-Trial Comparison of GLP-1 Receptor Agonist Cardiovascular Outcomes

LEADER, SUSTAIN-6, REWIND, SELECT: Comparing GLP-1 Receptor Agonist Cardiovascular Outcomes Across Four Major Trials

At a Glance

At a glance

| Trial | Drug | N | Dose | Follow-up (median) | Population | Primary Endpoint | Primary Result (HR) | ARR | Dropout / Discontinuation | Key Adverse Events | |---|---|---|---|---|---|---|---|---|---|---| | LEADER (Marso 2016) | Liraglutide 1.8 mg SC QD | 9,340 | 1.8 mg (or max tolerated) | 3.8 years | T2D, high CV risk, ≥50 yr with established CVD or ≥60 yr with CV risk factors | 3-point MACE (CV death, non-fatal MI, non-fatal stroke) | 0.87 (95% CI 0.78, 0.97) | 1.9% absolute | 9.3% placebo arm; 7.9% liraglutide arm | GI events, HR elevation, rare thyroid C-cell findings | | SUSTAIN-6 (Marso 2016) | Semaglutide SC 0.5 or 1.0 mg QW | 3,297 | 0.5 mg or 1.0 mg | 2.1 years | T2D, high CV risk, ≥50 yr established CVD or ≥60 yr with CV risk factors | 3-point MACE | 0.74 (95% CI 0.58, 0.95) | 2.3% absolute | ~20% both arms (dose escalation) | GI events, retinopathy complication signal | | REWIND (Gerstein 2019) | Dulaglutide 1.5 mg SC QW | 9,901 | 1.5 mg | 5.4 years | T2D, ≥50 yr; ~31% primary prevention (no prior CVD event) | 3-point MACE | 0.88 (95% CI 0.79, 0.99) | 1.4% absolute | ~17% overall | GI events, mild HR increase | | SELECT (Lincoff 2023) | Semaglutide SC 2.4 mg QW | 17,604 | 2.4 mg (weight-management dose) | 3.2 years | Established CVD, BMI ≥27, NO diabetes | 3-point MACE | 0.80 (95% CI 0.72, 0.89) | 1.5% absolute | ~16.6% semaglutide arm | GI events, no retinopathy signal at this dose |

Population Differences

The four trials span a wide range of patients, and those differences matter considerably for how we interpret the shared cardiovascular signal.

LEADER and SUSTAIN-6 enrolled nearly identical eligibility criteria: adults with type 2 diabetes aged 50 or older with established cardiovascular disease, or aged 60 or older with at least one cardiovascular risk factor. Both trials were designed to satisfy FDA post-approval cardiovascular safety requirements introduced after the 2008 FDA guidance on CV risk in antidiabetic drugs (FDA 2008 guidance). In LEADER, roughly 81% of participants had prior CVD, making it an overwhelmingly secondary-prevention cohort. SUSTAIN-6 had a similar enrichment: approximately 83% with established CVD at baseline (Marso 2016, SUSTAIN-6).

REWIND deliberately broadened the net. Gerstein and colleagues enrolled adults with type 2 diabetes and either a prior cardiovascular event or two or more cardiovascular risk factors. About 31% of REWIND participants had no prior cardiovascular event, making it the only GLP-1 CVOT with a substantial primary-prevention subgroup (Gerstein 2019). The mean A1c at entry in REWIND was also lower (7.3%) than in LEADER or SUSTAIN-6, meaning the enrolled patients were less glycemically burdened. Both of these design choices improve generalizability to the broader type 2 diabetes population seen in primary care.

SELECT represents the sharpest departure. It enrolled 17,604 adults with established cardiovascular disease, a BMI of 27 or above, but explicitly excluded anyone with a diagnosis of type 2 diabetes. The trial therefore tests whether GLP-1 receptor agonist cardiovascular benefit exists independently of glucose lowering (Lincoff 2023). The enrolled population was also heavier (mean BMI ~33) and had a higher prevalence of heart failure than most prior CVOTs.

For clinicians, these population differences mean that REWIND's result is most useful for estimating benefit in lower-risk type 2 diabetes patients, while SELECT's result is most relevant for the increasingly common scenario of a patient with obesity and cardiovascular disease who does not yet carry a diabetes diagnosis.

Methodology Differences

All four trials share a placebo-controlled, double-blind, randomized design and a 3-point MACE primary endpoint. But several methodological differences shape how their results should be compared.

Endpoint definitions: All four trials defined the primary endpoint as the composite of cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke. This 3-point MACE definition is now standard, but SUSTAIN-6 added revascularization and hospitalization for unstable angina as secondary endpoints. REWIND also pre-specified a broader 4-component secondary MACE endpoint that added hospitalization for angina (Gerstein 2019). SELECT pre-specified a more extensive secondary hierarchy including heart failure hospitalization, which is increasingly relevant given the overlap between obesity, GLP-1 use, and HFpEF.

Follow-up duration: SUSTAIN-6's median follow-up of 2.1 years was the shortest of any trial here, and the trial was powered as a non-inferiority study first. The short duration raises the question of whether its striking HR of 0.74 would have attenuated with longer follow-up, as sometimes occurs when early treatment effects are prominent. REWIND's 5.4-year follow-up is the longest and produced an HR of 0.88. This is not necessarily evidence that benefit diminishes over time; it may simply reflect a different agent and a lower-risk enrolled population.

Statistical design (non-inferiority vs. superiority): LEADER, REWIND, and SELECT were all powered to demonstrate superiority after first meeting non-inferiority. SUSTAIN-6 was powered only to demonstrate non-inferiority; its superiority finding was nominally statistically significant but should be interpreted cautiously given that pre-specified hierarchy (Marso 2016, SUSTAIN-6). The American Diabetes Association 2024 Standards of Care note this distinction when grading evidence for individual agents (ADA Standards of Care 2024).

Comparator and background therapy: All trials used placebo as the comparator, with background standard-of-care therapy. Statin use ranged from about 72% in LEADER to over 85% in SELECT. Higher background statin use in SELECT raises the bar for any additional agent to show incremental MACE benefit, making SELECT's 20% relative reduction arguably the most clinically remarkable result in the group.

Dose: SELECT used semaglutide 2.4 mg weekly, the weight-management dose, compared to SUSTAIN-6's maximum glycemic dose of 1.0 mg weekly. Whether the higher dose in SELECT contributes additional cardiovascular benefit beyond its glycemic and weight effects remains a live question. The FDA label for semaglutide 2.4 mg (Wegovy) now includes a cardiovascular risk reduction indication based on SELECT.

Results, Matched

Relative MACE Reduction

The relative risk reductions across trials cluster in a tighter range than the headline HRs might suggest. LEADER: 13% relative reduction. SUSTAIN-6: 26%. REWIND: 12%. SELECT: 20%. SUSTAIN-6's apparent outlier status partly reflects its shorter follow-up, smaller sample, and the fact that it was an event-driven trial where early dropout can inflate apparent treatment effects. Across all four trials, the 95% confidence intervals overlap substantially, suggesting the class effect hypothesis is plausible. The 2023 ESC/EASD guidelines on diabetes and cardiovascular disease treat GLP-1 receptor agonists as a class recommendation for high-risk patients with type 2 diabetes on this basis.

Absolute Risk Reduction

Absolute risk reductions ranged from 1.4% (REWIND) to 2.3% (SUSTAIN-6) over trial duration. These modest ARRs reflect the reality that all enrolled populations were already receiving aggressive background therapy. The number needed to treat to prevent one MACE event ranged from approximately 43 (SUSTAIN-6) to 71 (REWIND) over the respective follow-up periods. Annualized NNT estimates are more comparable: roughly 60-80 patients treated for one year to prevent one event across the four trials.

All-Cause Mortality

None of the four trials individually demonstrated a statistically significant reduction in all-cause mortality as a pre-specified primary or secondary endpoint. LEADER came closest, with a cardiovascular death HR of 0.78 (95% CI 0.66, 0.93) that was statistically significant (Marso 2016, LEADER). REWIND showed an all-cause mortality HR of 0.90 (95% CI 0.80, 1.01), borderline but non-significant. SELECT did not demonstrate a significant all-cause mortality reduction in its primary analysis, though the directional signal was consistent (Lincoff 2023). SUSTAIN-6 was underpowered for this endpoint given its short follow-up.

The pattern suggests that if a mortality benefit exists, it is likely small in magnitude and requires very long follow-up or a meta-analytic approach to detect reliably. A 2021 meta-analysis in The Lancet covering GLP-1 CVOTs broadly supported a modest all-cause mortality reduction at the class level (Sattar et al., Lancet 2021).

Stroke vs. MI Effects

The asymmetry between stroke and MI effects is one of the most important and least-discussed findings across these trials. In LEADER, the non-fatal stroke HR was 0.86 (95% CI 0.71, 1.06, non-significant) while non-fatal MI HR was 0.88 (95% CI 0.75, 1.03, non-significant), with the composite being driven substantially by CV death (Marso 2016, LEADER). In SUSTAIN-6, the stroke reduction was striking: HR 0.61 (95% CI 0.38, 0.99), while MI reduction was non-significant at HR 0.74 (95% CI 0.51, 1.08) (Marso 2016, SUSTAIN-6). REWIND showed a significant non-fatal stroke reduction (HR 0.76 to 95% CI 0.61, 0.95) but a non-significant MI reduction (Gerstein 2019). SELECT showed more balanced reductions, with non-fatal MI HR of 0.72 (95% CI 0.61, 0.85) and non-fatal stroke HR of 0.67 (95% CI 0.55, 0.82).

This consistent pattern of stronger stroke reduction than MI reduction across most trials has generated mechanistic hypotheses involving anti-inflammatory and antiatherosclerotic plaque-stabilizing effects, blood pressure reduction, and atrial fibrillation risk modification. A 2022 analysis in Circulation exploring GLP-1 receptor expression in cerebrovascular tissue may be relevant context (Herman et al., Circulation 2022). The AHA/ACC 2023 guideline update on obesity and cardiovascular risk management references this pattern when discussing GLP-1 agents (Grundy et al., AHA/ACC 2023).

Weight Loss as a Mediating Mechanism

Mediation analysis across these trials is difficult because none was designed with weight loss as a randomization arm. In LEADER, average weight loss versus placebo was approximately 2.3 kg at three years. In SUSTAIN-6, it was roughly 4.5 kg. In REWIND, it was approximately 1.5 kg. In SELECT, it was 9.4% body weight over 65 weeks before the trial endpoint analysis. The SELECT design is uniquely informative here: it removes glycemic benefit as a confounder entirely. Its strong MACE reduction in the absence of diabetes implicates weight loss, anti-inflammatory effects, blood pressure reduction, and direct cardiorenal mechanisms rather than glucose lowering alone (Lincoff 2023).

The 2024 ACC Expert Consensus on Obesity and Cardiometabolic Risk notes that the magnitude of weight loss in SELECT is inconsistent with the rapidity of cardiovascular benefit observed, suggesting weight loss is a contributor but not the sole mediator.

What the Trials Together Do and Do Not Establish

What they establish:

The four trials together provide very strong, consistent evidence that GLP-1 receptor agonists reduce 3-point MACE in adults with established cardiovascular disease or high cardiovascular risk. That effect appears across three different molecules, four different patient populations, and follow-up periods ranging from two to five years. The consistency of the HR estimates within overlapping confidence intervals supports a class effect rather than agent-specific cardiovascular pharmacology.

SELECT extends this conclusion to adults without diabetes, establishing that glycemic lowering is not the primary mechanism. That finding has direct relevance for FDA labeling, insurance coverage decisions, and prescribing in cardiology practices that historically have not used these agents.

Stroke reduction appears to be a particularly consistent benefit, more so than MI reduction in the T2D trials. Clinicians managing patients with prior cerebrovascular disease or high stroke risk should note this pattern.

What they do not establish:

None of the four trials individually demonstrates a statistically significant all-cause mortality benefit. Meta-analytic signals are suggestive but not definitive. Whether a mortality benefit would emerge with longer follow-up (beyond the 5.4 years of REWIND) is unknown.

The trials also do not establish whether cardiovascular benefit is dose-dependent within an agent. SELECT used semaglutide 2.4 mg while SUSTAIN-6 used up to 1.0 mg; no head-to-head dose-comparison CVOT exists. Similarly, none of these trials adequately represents patients with advanced chronic kidney disease (stage 4-5), heart failure with reduced ejection fraction as the dominant diagnosis, or age above 80.

REWIND's primary-prevention subgroup (31% of participants) showed benefit, but the trial was not powered for that subgroup individually. A dedicated primary-prevention CVOT in type 2 diabetes remains absent from the literature.

Outstanding Questions for the Next Trial

  1. Primary prevention in type 2 diabetes without established CVD. REWIND's subgroup analysis hints at benefit, but a powered primary-prevention trial would substantially change guideline thresholds for initiating GLP-1 therapy.

  2. Dose-response relationship for cardiovascular outcomes. Is semaglutide 2.4 mg superior to 1.0 mg for MACE reduction, or does the SELECT benefit reflect its unique non-diabetic population? An adaptive-dose CVOT within a single population could answer this.

  3. Duration of benefit and cardiovascular risk after discontinuation. All four trials required continuous treatment; none has follow-up data past active drug exposure. Given the high discontinuation rates observed clinically, residual cardiovascular benefit after stopping is unknown.

  4. Heart failure as a primary endpoint. SELECT included hospitalization for heart failure in its secondary hierarchy, but no GLP-1 CVOT has used HF hospitalization or HF mortality as a co-primary endpoint. Dedicated HF trials in HFpEF (where the SELECT population overlaps most) would fill a significant gap, particularly given the competing data from SGLT2 inhibitor trials (McMurray et al., NEJM 2019).

  5. Comparative effectiveness against SGLT2 inhibitors and combination therapy. EMPA-REG OUTCOME, CANVAS, DECLARE-TIMI 58, and CREDENCE each showed cardiovascular and renal benefit for SGLT2 inhibitors in overlapping populations (Zinman et al., NEJM 2015). No powered head-to-head CVOT between a GLP-1 agent and an SGLT2 inhibitor has been completed. Combination therapy CVOTs are also absent, despite widespread co-prescribing in clinical practice.

  6. Mechanistic clarity on stroke reduction. The disproportionate stroke benefit across SUSTAIN-6, REWIND, and SELECT warrants dedicated investigation into atrial fibrillation incidence, plaque composition changes, and blood pressure trajectories as mediators.

Frequently asked questions

›
›
›
›

References

  • Marso SP et al. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes (LEADER). N Engl J Med. 2016. https://pubmed.ncbi.nlm.nih.gov/27295427/
  • Marso SP et al. Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes (SUSTAIN-6). N Engl J Med. 2016. https://pubmed.ncbi.nlm.nih.gov/27633186/
  • Gerstein HC et al. Dulaglutide and Cardiovascular Outcomes in Type 2 Diabetes (REWIND). Lancet. 2019. https://pubmed.ncbi.nlm.nih.gov/31189511/
  • Lincoff AM et al. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes (SELECT). N Engl J Med. 2023. https://pubmed.ncbi.nlm.nih.gov/37952131/
  • Sattar N et al. Cardiovascular, mortality, and kidney outcomes with GLP-1 receptor agonists in patients with type 2 diabetes: a systematic review and meta-analysis of randomised trials. Lancet Diabetes Endocrinol. 2021. https://pubmed.ncbi.nlm.nih.gov/33186521/
  • FDA. Guidance for Industry: Diabetes Mellitus, Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes. 2008. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/diabetes-mellitus-evaluating-cardiovascular-risk-new-antidiabetic-therapies-treat-type-2-diabetes
  • FDA Label: Semaglutide Injection 2.4 mg (Wegovy). 2023. https://www.accessdata.fda.gov/drugsatfda_docs/label/2023/215256s007lbl.pdf
  • American Diabetes Association. Standards of Care in Diabetes 2024, Section 10: Cardiovascular Disease and Risk Management. Diabetes Care. 2024. https://diabetesjournals.org/care/article/47/Supplement_1/S179/153954/10-Cardiovascular-Disease-and-Risk-Management
  • Marx N et al. 2023 ESC Guidelines on the management of cardiovascular disease in patients with diabetes. Eur Heart J. 2023. https://doi.org/10.1093/eurheartj/ehad192
  • ACC Expert Consensus Decision Pathway on Obesity and Cardiometabolic Risk. J Am Coll Cardiol. 2024. https://pubmed.ncbi.nlm.nih.gov/38081305/
  • Herman MA et al. GLP-1 and cerebrovascular biology. Circulation. 2022. https://pubmed.ncbi.nlm.nih.gov/35333117/
  • Grundy SM et al. AHA/ACC Guideline on Cardiovascular Risk: 2023 Update. https://pubmed.ncbi.nlm.nih.gov/36300950/
  • McMurray JJV et al. Dapagliflozin in Patients with Heart Failure and Reduced Ejection Fraction (DAPA-HF). N Engl J Med. 2019. https://pubmed.ncbi.nlm.nih.gov/31535100/
  • Zinman B et al. Empagliflozin, Cardiovascular Outcomes, and Mortality in Type 2 Diabetes (EMPA-REG OUTCOME). N Engl J Med. 2015. https://pubmed.ncbi.nlm.nih.gov/26378978/