GLP-1 Receptor Agonists and Cardiorenal Outcomes in Type 1 Diabetes

Key Takeaways

• GLP-1 receptor agonist initiation was associated with lower risks of major adverse cardiovascular events and end-stage kidney disease than noninitiation.
• Initiation was also associated with lower hospitalization for heart failure and major adverse liver events, alongside greater likelihood of reaching 5%, 10%, and 15% weight loss thresholds.
• Hospitalization risks for diabetic ketoacidosis and severe hypoglycemia were not increased, gastrointestinal events were more frequent, and age and HbA1c subgroup findings were generally consistent.

In original research in Nature Medicine, GLP-1 receptor agonist initiation in type 1 diabetes was associated with lower 5-year major adverse cardiovascular event risk than noninitiation, with an overall HR of 0.85. End-stage kidney disease risk was also lower over the same period. The analysis drew on a large national electronic health record cohort and evaluated associations rather than treatment effects. Major cardiorenal endpoints moved in the same direction overall.

Investigators used a sequential target-trial emulation with de-identified Optum Labs electronic health record data from more than 60 health systems. The cohort included 174,678 people with type 1 diabetes who contributed 6,092,537 person-trials across 135 monthly target trials from January 2013 through March 2024. Eligible participants used insulin, had baseline BMI and HbA1c values, and had at least 12 months of prior engagement. The comparison was GLP-1 receptor agonist initiation versus noninitiation, and propensity score overlap weighting left weighted standardized mean differences below 10%. Mean age was 43 years, 47% were female, and the prespecified primary outcomes were major adverse cardiovascular events and end-stage kidney disease.

At 5 years, major adverse cardiovascular event risk was 4.3% with initiation and 5.0% with noninitiation, for a risk difference of −0.7% and a 95% CI from −1.2% to −0.2%. The corresponding hazard ratio for MACE was 0.85 with a 95% CI of 0.77 to 0.95. End-stage kidney disease risk at 5 years was 1.6% versus 1.9%, with a risk difference of −0.3%, a 95% CI from −0.6% to 0%, and an HR of 0.81. Its 95% CI for the hazard ratio was 0.69 to 0.95, and median follow-up in the main intention-to-treat analysis was 38 months. Both prespecified primary endpoints moved in the same direction.

Secondary outcomes also favored initiation, with lower heart failure hospitalization, HR 0.82, 95% CI 0.71–0.94, and lower major adverse liver events, HR 0.72, 95% CI 0.60–0.85. Initiators were also more likely to reach 5%, 10%, and 15% weight loss thresholds during follow-up. For hospitalization-based safety endpoints, no increased risk was seen for diabetic ketoacidosis, HR 0.83, 95% CI 0.76–0.90, or severe hypoglycemia, HR 0.82, 95% CI 0.72–0.94. Gastrointestinal adverse events were more frequent among initiators, although the risk difference was not statistically significant. The secondary results also included a more frequent, but not statistically significant, gastrointestinal signal.

Exploratory analyses by age and baseline HbA1c were generally consistent, and per-protocol results with 7 months of median follow-up were less precise but similar. Additional sensitivity analyses supported the main direction of results. Traffic accidents showed no significant between-group difference, HR 0.96 with 95% CI 0.77–1.20, and E-values were 1.63 for MACE and 1.77 for ESKD. The authors noted possible residual and unmeasured confounding, incomplete exclusion of type 1 diabetes misclassification despite a validated algorithm, and hospitalization-based capture of severe events. Insulin-dose data were unavailable, individual agents were not evaluated separately, and weight-loss findings could include unintentional loss without a sustainability assessment, and the authors said large randomized trials are warranted.