← All trial deep-dives

EXSCEL trial deep-dive: exenatide once-weekly CV outcomes in T2D (Holman 2017)

Last verified 2026-05-28 · Phase 3/4 · Completed (results published Sep 2017) · NCT01144338 ↗

By Eli Marsden · Founding Editor
Editorially reviewed (not clinically reviewed) · How we verify contentLast reviewed

EXSCEL (Exenatide Study of Cardiovascular Event Lowering) was a pragmatic, multicenter, double-blind, placebo-controlled, event-driven phase 3/4 trial sponsored by Amylin Pharmaceuticals and subsequently AstraZeneca, co-conducted by the Duke Clinical Research Institute and the University of Oxford Diabetes Trials Unit. Investigators randomized 14,752 adults with type 2 diabetes — with or without prior cardiovascular disease — 1:1 to once-weekly subcutaneous extended-release exenatide 2 mg or matching placebo, layered on usual care. The primary endpoint was a three-point major adverse cardiovascular event composite of cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke. Median follow-up was 3.2 years. Results published by Holman and colleagues in the New England Journal of Medicine on September 28, 2017 reported a hazard ratio of 0.91 for the primary outcome, which met the prespecified non-inferiority margin (P<0.001) but missed superiority (P=0.06). EXSCEL is the trial that complicated the early GLP-1 cardioprotection narrative: with LEADER and SUSTAIN-6 already positive and ELIXA neutral, exenatide's miss reframed the discussion from a class effect toward molecule-specific cardiovascular pharmacology.

Enrollment
14,752
Duration
Median follow-up 3.2 years (interquartile range 2.2 to 4.4 years); event-driven design
Drug
Exenatide once-weekly (Bydureon)
Population
Adults aged 18 or older with type 2 diabetes and an HbA1c between 6.5% and 10.0%, with or without prior cardiovascular disease. Approximately 73.1% had a history of prior cardiovascular events (secondary prevention) and 26.9% did not (primary prevention). Mean age 62.0 years, 38.0% female, mean HbA1c 8.0%, mean diabetes duration 12 years, mean BMI 31.8 kg/m², mean eGFR 76.4 mL/min/1.73 m². Background therapy reflected real-world type 2 diabetes care: 77% on metformin, 47% on sulfonylureas, 46% on insulin at baseline.

Primary endpoint

Three-point major adverse cardiovascular events (MACE): cardiovascular death, nonfatal myocardial infarction, or nonfatal stroke (time-to-first-event, intention-to-treat)

Treatment arm

839 of 7,356 (11.4%); 3.7 events per 100 person-years

Comparator

905 of 7,396 (12.2%); 4.0 events per 100 person-years

Treatment difference: Hazard ratio 0.91 (95% CI 0.83 to 1.00); P<0.001 for non-inferiority, P=0.06 for superiority

EXSCEL met the prespecified non-inferiority margin required for the FDA's post-2008 cardiovascular safety mandate but narrowly missed the conventional threshold for superiority. The 9% point estimate for relative reduction sat between the neutral ELIXA result and the clearly positive LEADER (HR 0.87) and SUSTAIN-6 (HR 0.74) results, complicating any clean class-effect narrative.

Secondary endpoints

EndpointTreatmentComparatorDifference
Cardiovascular death

Directionally favorable but did not reach statistical significance and was not protected from multiplicity given the missed primary superiority test.

340 of 7,356 (4.6%)383 of 7,396 (5.2%)Hazard ratio 0.88 (95% CI 0.76 to 1.02)
Nonfatal myocardial infarction354 of 7,356 (4.8%)383 of 7,396 (5.2%)Hazard ratio 0.95 (95% CI 0.82 to 1.10)
Nonfatal stroke187 of 7,356 (2.5%)218 of 7,396 (2.9%)Hazard ratio 0.86 (95% CI 0.70 to 1.05)
Death from any cause (all-cause mortality)

All-cause mortality crossed the nominal significance threshold, but per the prespecified hierarchical testing plan it was not formally significant because the primary superiority test had failed. Hypothesis-generating only.

507 of 7,356 (6.9%)584 of 7,396 (7.9%)Hazard ratio 0.86 (95% CI 0.77 to 0.97)
Hospitalization for heart failure

No effect on heart-failure hospitalization, mirroring the broader GLP-1 class pattern and contrasting with the clear heart-failure benefit later seen with SGLT2 inhibitors.

219 of 7,356 (3.0%)231 of 7,396 (3.1%)Hazard ratio 0.94 (95% CI 0.78 to 1.13)
Hospitalization for acute coronary syndromeReported as supportive endpointReported as supportive endpointHazard ratio 0.97 (95% CI 0.87 to 1.07)
HbA1c change from baseline (least-squares mean difference vs placebo across the trial)

Modest glycemic benefit in a real-world background-therapy population; smaller than in head-to-head exenatide vs comparator trials because of more permissive rescue-therapy use in EXSCEL.

ReducedReferenceBetween-group difference −0.53 percentage points (95% CI −0.57 to −0.50)
Body weight change from baseline (least-squares mean difference vs placebo)

Weight reduction was clinically modest compared with semaglutide and tirzepatide trials, reflecting both exenatide's lower receptor-activation magnitude and the pragmatic trial design.

ReducedReferenceBetween-group difference −1.27 kg (95% CI −1.40 to −1.13)

Adverse events

EventTreatment rateComparator rate
Injection-site reactions

Once-weekly extended-release exenatide uses a microsphere depot formulation that produces visible subcutaneous nodules in a meaningful minority of patients; this was a known class-distinct tolerability issue versus liquid GLP-1 formulations.

Higher with exenatideLower with placebo
Confirmed acute pancreatitis

No increased pancreatitis signal — consistent with other large GLP-1 CVOTs.

31 of 7,344 (0.4%)39 of 7,389 (0.5%)
Pancreatic cancer

No imbalance; addresses a regulatory concern that had been raised for the GLP-1 class.

20 of 7,344 (0.3%)24 of 7,389 (0.3%)
Medullary thyroid carcinoma

No cases observed in either arm during the trial.

0 of 7,3440 of 7,389
Severe hypoglycemia

Slightly higher with exenatide in absolute terms but not driven by the GLP-1 mechanism itself — concomitant insulin and sulfonylurea use in this pragmatic population was the dominant driver.

247 of 7,344 (3.4%)219 of 7,389 (3.0%)
Permanent discontinuation of trial drug

High discontinuation rates in both arms reflect the pragmatic design and the long follow-up; the gap between arms was small.

43.0%45.2%

Subgroup analyses

  • Prior cardiovascular disease (secondary prevention, 73.1% of cohort): Hazard ratio 0.90 (95% CI 0.82 to 0.99)

    Point estimate slightly more favorable than the overall trial; in the secondary-prevention subgroup the upper 95% bound did not cross 1.0, but this was a prespecified subgroup, not a primary test.

  • No prior cardiovascular disease (primary prevention, 26.9% of cohort): Hazard ratio 0.99 (95% CI 0.79 to 1.23)

    No signal in the primary-prevention subgroup, consistent with the broader pattern that GLP-1 cardioprotection is most evident in established cardiovascular disease.

  • Baseline eGFR <60 mL/min/1.73 m² (chronic kidney disease): Hazard ratio 0.88 (95% CI 0.76 to 1.01)

    Directional benefit in the renal-impairment subgroup, but the prespecified test for treatment-by-subgroup interaction across eGFR strata was not significant.

Clinical significance

EXSCEL is the trial that broke the early GLP-1 class-effect narrative. By 2017, LEADER (liraglutide, 2016, HR 0.87) and SUSTAIN-6 (semaglutide, 2016, HR 0.74) had both shown clear cardiovascular benefit, while ELIXA (lixisenatide, 2015) had been neutral. EXSCEL was the largest and longest of the GLP-1 cardiovascular outcomes trials at that point, and its near-miss superiority result — HR 0.91 with a 95% confidence interval whose upper bound touched but did not cross 1.0 — forced clinicians and regulators to abandon a simple class-effect view. The dominant interpretation since 2017 has been that GLP-1 cardiovascular benefit depends on the specific molecule, its half-life, its tissue distribution, and possibly its degree of receptor agonism. Once-weekly exenatide, with a shorter functional half-life and a microsphere depot formulation, did not deliver the same cardioprotection as longer-acting liraglutide or semaglutide. REWIND (dulaglutide, 2019, HR 0.88) and AMPLITUDE-O (efpeglenatide, 2021, HR 0.73) later restored the class story for the longer-acting agents, while HARMONY OUTCOMES (albiglutide, 2018, HR 0.78) added a positive result for another weekly molecule that was nonetheless withdrawn from the market for commercial reasons. EXSCEL itself stopped being clinically actionable in the United States in 2024, when Bydureon was discontinued.

Frequently asked questions

Why did EXSCEL miss superiority?

The primary three-point MACE hazard ratio in EXSCEL was 0.91 with a 95% confidence interval of 0.83 to 1.00 and a P-value of 0.06 for superiority — a near miss rather than a flat result. Several design and population factors contributed. EXSCEL was a pragmatic trial that enrolled both primary- and secondary-prevention patients, with about a quarter of participants having no prior cardiovascular disease at baseline, diluting the absolute event rate. The trial allowed open-label glucose-lowering therapy in both arms, including other GLP-1 receptor agonists and SGLT2 inhibitors, which narrowed the between-arm separation over time. Discontinuation of the assigned trial drug was high in both arms (about 43% on exenatide and 45% on placebo by end of follow-up), reducing the effective on-treatment exposure. Once-weekly extended-release exenatide also has a different pharmacologic profile than liraglutide or semaglutide — a shorter functional half-life, a microsphere depot that produces variable plasma concentrations, and weaker brain penetrance — which most reviewers now believe explains the bulk of the gap rather than trial design alone.

Is exenatide still available?

Once-weekly extended-release exenatide (Bydureon and Bydureon BCise) was discontinued in the U.S. market by AstraZeneca in 2024. Twice-daily exenatide (Byetta) was discontinued in the U.S. years earlier. Outside the United States, exenatide-containing products remain available in some markets, but the molecule's commercial footprint has been replaced almost entirely by semaglutide, dulaglutide, liraglutide, and tirzepatide. For prescribers in the United States, there is no longer a branded exenatide product on the market, and there is no FDA-approved generic version of either the once-weekly or twice-daily formulation as of 2026.

Why was Bydureon discontinued in the U.S.?

AstraZeneca cited commercial rather than safety reasons. Bydureon required a complex weekly mixing step with a microsphere suspension that patients found cumbersome compared with the prefilled liquid pens used for semaglutide and dulaglutide. Injection-site nodules — a known consequence of the depot formulation — were a persistent tolerability complaint. Weight loss with Bydureon (approximately 1 to 2 kg in EXSCEL) was modest compared with the 5-10% body-weight reductions clinicians and patients were by then seeing with semaglutide. The EXSCEL near-miss on cardiovascular superiority meant that exenatide could not match the cardioprotective labeling that competing GLP-1 agents had earned. Combined with patent expiry and an unfavorable head-to-head competitive position, the product became uneconomic to keep in the U.S. portfolio, and AstraZeneca announced withdrawal in 2024.

How does EXSCEL inform GLP-1 class cardiovascular effects?

EXSCEL was the trial that ended the simple class-effect hypothesis for GLP-1 cardiovascular protection. Before EXSCEL reported in September 2017, the two completed superiority trials for the class — LEADER for liraglutide and SUSTAIN-6 for semaglutide — had both been positive, and many commentators expected exenatide to extend the pattern. Its near-miss result, in the largest GLP-1 cardiovascular trial completed at that time, established that cardioprotection is molecule-dependent. The current synthesis, supported by REWIND (dulaglutide, positive), AMPLITUDE-O (efpeglenatide, positive), and HARMONY OUTCOMES (albiglutide, positive), is that the longer-acting human-GLP-1-backbone agents tend to show cardiovascular benefit, while exendin-4-derived agents with shorter functional half-lives (exenatide, lixisenatide) tend not to. This distinction is now built into how the FDA and EMA evaluate cardiovascular labeling claims — each agent must be judged on its own outcomes trial rather than inherited from the class.

Should patients on exenatide switch to a different GLP-1 for cardiovascular protection?

In the United States this is essentially a historical question because Bydureon and Byetta were both discontinued — patients who were on exenatide have already been transitioned, typically to semaglutide or dulaglutide. From a pure outcomes-evidence standpoint, a patient with type 2 diabetes and established cardiovascular disease is better served by an agent with a positive cardiovascular trial: liraglutide (LEADER), semaglutide (SUSTAIN-6), dulaglutide (REWIND), or efpeglenatide (AMPLITUDE-O), all of which now carry or supported labeled cardiovascular-risk-reduction claims in some form. Tirzepatide (the dual GIP/GLP-1 agonist used in Mounjaro and Zepbound) does not yet have a completed cardiovascular outcomes trial in type 2 diabetes — SURPASS-CVOT is ongoing — but its glycemic and weight-loss effects exceed those seen with any GLP-1 monotherapy. Any switch decision should be made with the prescribing clinician, weighing renal function, cost, coverage, injection burden, and the patient's full cardiovascular risk profile.

References

  1. 1.Holman RR, Bethel MA, Mentz RJ, Thompson VP, Lokhnygina Y, Buse JB, Chan JC, Choi J, Gustavson SM, Iqbal N, Maggioni AP, Marso SP, Öhman P, Pagidipati NJ, Poulter N, Ramachandran A, Zinman B, Hernandez AF; EXSCEL Study Group. Effects of Once-Weekly Exenatide on Cardiovascular Outcomes in Type 2 Diabetes. N Engl J Med. 2017. PMID: 28910237.
  2. 2.Amylin Pharmaceuticals, LLC; AstraZeneca. Exenatide Study of Cardiovascular Event Lowering Trial (EXSCEL): A Trial to Evaluate Cardiovascular Outcomes After Treatment With Exenatide Once Weekly in Patients With Type 2 Diabetes Mellitus. ClinicalTrials.gov, NCT01144338. 2017. https://clinicaltrials.gov/study/NCT01144338
  3. 3.AstraZeneca. Bydureon (exenatide extended-release) product portfolio communication — U.S. market discontinuation, 2024. AstraZeneca corporate disclosure. 2024. https://www.astrazeneca.com/our-therapy-areas/cardiovascular-renal-and-metabolism.html

Related trial deep-dives