Did the REWIND trial have a formal extension study?

No. As of mid-2026, no formal post-trial extension with continued participant follow-up has been published. Several prespecified substudies (cognitive, renal, stroke) analyzed outcomes within the original trial duration, but none followed participants after drug discontinuation.

Was the cardiovascular benefit driven entirely by stroke?

Stroke was the only individual MACE component to reach statistical significance (HR 0.76). Both non-fatal MI and cardiovascular death trended toward benefit but did not reach significance individually. The composite endpoint was significant (HR 0.88, p = 0.026), indicating that the benefit was not attributable to stroke alone.

How does REWIND's follow-up duration compare to other GLP-1 trials?

REWIND had the longest median follow-up of any GLP-1 RA CVOT at 5.4 years. LEADER followed for 3.8 years, SUSTAIN-6 for 2.1 years, and SELECT for 3.3 years. This longer observation period allowed better assessment of late-emerging safety signals.

Did dulaglutide show kidney benefits in REWIND?

Yes. Dulaglutide reduced new-onset macroalbuminuria by 23% (HR 0.77). The composite renal endpoint also favored dulaglutide. Sustained eGFR decline showed a favorable trend but was not independently significant.

Is there evidence that the MACE benefit persists after stopping dulaglutide?

No direct evidence exists because no post-cessation follow-up study has been published. The Kaplan-Meier curves did not converge during the on-treatment period, but what happens after discontinuation remains unknown.

What was the rate of pancreatitis in REWIND?

Acute pancreatitis occurred in 0.3% of dulaglutide patients vs. 0.2% on placebo over 5.4 years. This difference was not statistically significant and represents one of the most reassuring long-term datasets for pancreatic safety in the GLP-1 RA class.

Does REWIND apply to patients without diabetes?

No. REWIND enrolled only patients with type 2 diabetes. The SELECT trial later demonstrated cardiovascular benefit for semaglutide in patients without diabetes but with obesity and established CVD.

Why was dulaglutide's MACE reduction smaller than semaglutide's?

Multiple factors likely contribute: REWIND's lower-risk population (69% primary prevention), longer follow-up allowing background therapy convergence, and possible differences in drug potency. No head-to-head cardiovascular comparison between GLP-1 RAs exists, so direct potency conclusions cannot be drawn from cross-trial comparisons.

Did REWIND influence clinical guidelines?

Yes. REWIND contributed to the ADA and ESC recommendations favoring GLP-1 RAs for patients with T2D and cardiovascular risk, not just those with established cardiovascular disease. Its inclusion of primary prevention patients was central to broadening guideline recommendations.

What dose of dulaglutide was used in REWIND?

Only the 1.5 mg weekly dose was studied. Current prescribing ranges from 0.75 mg to 4.5 mg. Whether higher or lower doses produce different cardiovascular effects is unknown.

REWIND Extension Data and What Happened After the Trial Ended

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Parameter | Detail | |---|---| | N | 9,901 | | Intervention | Dulaglutide 1.5 mg subcutaneous weekly | | Comparator | Matched placebo | | Median follow-up | 5.4 years | | Primary endpoint | First occurrence of MACE (CV death, non-fatal MI, non-fatal stroke) | | Key result | HR 0.88 (95% CI 0.79, 0.99; p = 0.026) | | Population | T2D with established CVD or CV risk factors (age ≥50 with CVD, age ≥55 with risk factors, age ≥60 with ≥2 risk factors) |

Why the Original REWIND Trial Still Generates Questions

Most GLP-1 receptor agonist cardiovascular outcomes trials (CVOTs) enrolled patients with established atherosclerotic disease. REWIND was different. Roughly 69% of its participants had cardiovascular risk factors but no prior event, making it the first large GLP-1 RA trial to test a primary prevention hypothesis alongside secondary prevention. The 12% MACE reduction was statistically significant, but the trial's design raised a specific set of follow-up questions that the original publication could not fully resolve.

Did the benefit persist after the drug was stopped? Was the stroke signal real or a statistical artifact amplified by modest overall event rates? And did long-term exposure surface safety concerns that a 5.4-year median could miss?

What the Original Publication Left Unresolved

Three issues stood out when the primary results appeared in The Lancet in 2019.

Primary vs. secondary prevention separation. The prespecified subgroup analysis showed no significant interaction between patients with and without prior CVD (interaction p = 0.97). Both subgroups trended toward benefit. But critics noted the primary prevention subgroup's confidence interval was wide, and the event rate was lower than projected, limiting statistical power for a definitive claim in that population alone.

Stroke drove the composite. Non-fatal stroke was reduced by 24% (HR 0.76 to 95% CI 0.61, 0.95). Non-fatal MI showed a non-significant 4% reduction. Cardiovascular death showed a non-significant 9% reduction. This raised the question of whether dulaglutide's benefit was genuinely concentrated on cerebrovascular outcomes or whether the MI and CV death components simply needed more events to reach significance.

HbA1c convergence over time. The glycemic separation between arms narrowed as the trial progressed and investigators were permitted to intensify background therapy. By trial end, the HbA1c difference was roughly 0.3%, down from an early peak of about 0.6%. This made it harder to attribute the MACE reduction purely to glucose lowering and pointed toward direct vascular or anti-inflammatory mechanisms.

Post-Trial Substudies and Extended Analyses

The HealthRX Durability Assessment Framework for CVOTs

To evaluate whether a cardiovascular benefit from a trial like REWIND is durable, we apply four criteria. Each criterion gets a rating of confirmed, probable, or unresolved based on available post-trial data.

| Criterion | Question | REWIND Status | |---|---|---| | On-treatment persistence | Did the hazard ratio remain stable across time-on-drug quartiles? | Confirmed. Kaplan-Meier curves separated early and did not converge through 5.4 years. | | Component consistency | Did ≥2 of 3 MACE components trend in the same direction? | Probable. All three trended toward benefit; stroke was significant, MI and CV death were not. | | Subgroup robustness | Did major prespecified subgroups (age, sex, baseline CVD) show consistent direction? | Confirmed. No significant interaction across 14 prespecified subgroups. | | Post-cessation signal | Is there evidence the benefit persists or reverses after drug discontinuation? | Unresolved. No formal post-trial washout follow-up has been published as of mid-2026. |

This framework highlights REWIND's central limitation: we lack a formal extension study with post-drug-cessation follow-up comparable to what LEADER-EXTEND or SUSTAIN-FORTE provided for other GLP-1 RAs.

Stroke-Specific Follow-Up

The most clinically significant post-hoc analysis examined stroke outcomes in detail. Published in 2020, this analysis confirmed that the 24% reduction in non-fatal stroke was driven primarily by ischemic events. Hemorrhagic stroke events were too few to analyze separately. The stroke benefit appeared consistent regardless of baseline statin use, blood pressure, or atrial fibrillation history.

This finding matters because no other GLP-1 RA trial at the time had demonstrated a statistically significant stroke reduction as a standalone component. SUSTAIN-6 with semaglutide showed a significant stroke reduction (HR 0.61), but in a smaller, higher-risk population. REWIND's result in a lower-risk cohort strengthened the biological plausibility of GLP-1-mediated cerebrovascular protection.

Cognitive Outcomes (REWIND-Cog)

A prespecified substudy assessed cognitive function using the Montreal Cognitive Assessment (MoCA) and the Digit Symbol Substitution Test (DSST). Published results showed a modest attenuation of cognitive decline in the dulaglutide arm over approximately 5 years, though the clinical significance of the difference (roughly 0.5 points on MoCA) remained debatable. This substudy could not distinguish whether the cognitive signal was secondary to stroke prevention or reflected a direct neuroprotective effect.

Kidney Outcomes

A renal composite outcome analysis demonstrated that dulaglutide reduced new-onset macroalbuminuria (HR 0.77 to 95% CI 0.68, 0.87) and the composite renal endpoint. The sustained eGFR decline component showed a favorable trend but did not reach significance independently. These kidney findings have since influenced the ADA Standards of Care recommendations for GLP-1 RA use in patients with T2D and albuminuria.

Long-Term Safety: What Emerged Over 5+ Years

REWIND's extended duration relative to earlier CVOTs (LEADER ran 3.8 years median, SUSTAIN-6 ran 2.1 years) provided a longer window for safety signal detection.

Gastrointestinal Events

Nausea occurred in 19.5% of dulaglutide-treated patients vs. 6.4% on placebo. Diarrhea occurred in 14.8% vs. 8.4%. These rates are consistent with the dulaglutide FDA prescribing information. Discontinuation due to GI adverse events was 5.7% in the dulaglutide group. The long trial duration showed that GI events clustered in the first 3 to 6 months and diminished over time.

Pancreatic Safety

Acute pancreatitis occurred in 0.3% (dulaglutide) vs. 0.2% (placebo), a difference that was not statistically significant. Pancreatic cancer occurred in 0.1% in both arms. These low rates over 5.4 years of median exposure represent the most reassuring long-term pancreatic safety dataset available for a GLP-1 RA at the time of publication.

| Safety Signal | Dulaglutide (%) | Placebo (%) | Notable? | |---|---|---|---| | Nausea | 19.5 | 6.4 | Expected, early-onset | | Diarrhea | 14.8 | 8.4 | Expected | | Acute pancreatitis | 0.3 | 0.2 | Not significant | | Pancreatic cancer | 0.1 | 0.1 | No signal | | Medullary thyroid cancer | 0 | 0 | No events | | Injection site reactions | 2.5 | 0.7 | Mild |

Thyroid Safety

Zero cases of medullary thyroid carcinoma occurred in either arm over the trial's duration. Calcitonin levels were not systematically collected, which limits the utility of this datapoint for detecting subclinical C-cell changes. The FDA label retains the boxed warning for thyroid C-cell tumors based on rodent data, not clinical events.

The Missing Extension: What We Do Not Have

Unlike liraglutide (LEADER), semaglutide (SELECT), and tirzepatide (SURPASS-CVOT, ongoing), dulaglutide does not have a published post-trial extension with continued follow-up after drug discontinuation. This gap means we cannot answer three specific questions.

First, does the MACE benefit persist, attenuate, or reverse once dulaglutide is stopped? For liraglutide, post-trial observational data suggested some persistence of benefit, but selection bias in follow-up cohorts makes interpretation difficult.

Second, does body weight regain after cessation eliminate part of the cardiovascular benefit? REWIND participants lost a modest 1.5 kg on average, far less than the 4 to 6 kg seen in newer GLP-1 RA trials. If the cardiovascular benefit were weight-mediated, the modest weight loss in REWIND would argue against that mechanism.

Third, do very long-term exposures (beyond 7 to 8 years) surface delayed oncologic or immune signals? The 5.4-year median provides reasonable reassurance, but GLP-1 RA use is increasingly lifelong, and the question of decade-scale safety remains open across the entire drug class.

How REWIND Fits the Broader GLP-1 RA Cardiovascular Evidence

| Trial | Drug | N | Median FU | MACE HR | Primary prevention included? | |---|---|---|---|---|---| | LEADER | Liraglutide | 9,340 | 3.8 yr | 0.87 | No (established CVD only) | | SUSTAIN-6 | Semaglutide SC | 3,297 | 2.1 yr | 0.74 | No | | REWIND | Dulaglutide | 9,901 | 5.4 yr | 0.88 | Yes (69% primary prevention) | | SELECT | Semaglutide SC | 17,604 | 3.3 yr | 0.80 | No (established CVD, no T2D) | | HARMONY | Albiglutide | 9,463 | 1.6 yr | 0.78 | No |

REWIND's unique contribution is the inclusion of primary prevention patients and the longest median follow-up in this class. Its effect size (12% relative risk reduction) is the most conservative, which may reflect the lower-risk population or may reflect a true difference in potency between dulaglutide and other GLP-1 RAs. Head-to-head cardiovascular comparisons within the class do not exist.

Limitations Worth Repeating

The investigators acknowledged several constraints. The open-label glycemic management after randomization introduced background therapy imbalance. By year 5, dulaglutide-treated patients were less likely to be on insulin, sulfonylureas, and SGLT2 inhibitors than placebo patients, because their glycemic control was better. This means the "placebo" arm was actually receiving more aggressive glucose-lowering co-therapy, which would bias the trial toward the null, potentially underestimating dulaglutide's true effect.

The 1.5 mg dose was the only dose tested. Current prescribing uses doses from 0.75 mg to 4.5 mg. Whether the cardiovascular benefit scales with dose is unknown.

Geographic heterogeneity was substantial (24 countries), which strengthens generalizability but introduces variation in background care standards that could dilute effect estimates.

Frequently asked questions

›

References

Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and cardiovascular outcomes in type 2 diabetes (REWIND): a double-blind, randomised placebo-controlled trial. Lancet. 2019;394(10193):121-130. PubMed
Gerstein HC, Hart R, Colhoun HM, et al. The effect of dulaglutide on stroke: an exploratory analysis of the REWIND trial. Lancet Diabetes Endocrinol. 2020;8(9):755-766. PubMed
Cukierman-Yaffe T, Gerstein HC, Colhoun HM, et al. Effect of dulaglutide on cognitive impairment in type 2 diabetes: an exploratory analysis of the REWIND trial. Lancet Neurol. 2020;19(7):582-590. PubMed
Gerstein HC, Colhoun HM, Dagenais GR, et al. Dulaglutide and renal outcomes in type 2 diabetes: an exploratory analysis of the REWIND randomised, placebo-controlled trial. Lancet. 2019;394(10193):131-138. PubMed
Dulaglutide (Trulicity) prescribing information. U.S. Food and Drug Administration. FDA Label
American Diabetes Association Professional Practice Committee. Standards of Care in Diabetes, 2023. Diabetes Care. 2023;46(Suppl 1). PubMed