Glucagon-like peptide-1 receptor agonists (GLP1-RA) have established cardiovascular benefits in people with type 2 diabetes (T2D) and obesity, and recent evidence supports benefit in obesity with heart failure with preserved ejection fraction. Their effects in patients who have both T2D and heart failure with reduced ejection fraction (HFrEF) are less well defined. The authors set out to evaluate associations between GLP1-RA initiation and cardiovascular outcomes in a real-world cohort of patients with T2D and HFrEF.
The study is a retrospective cohort analysis using pooled electronic health record data from the TriNetX Research Network. The authors extracted records of patients with T2D and HFrEF who were newly initiated on GLP1-RA therapy during the period May 13, 2026 to June 6, 2026. One-to-one propensity-score matching was used to assemble comparable treatment and control cohorts for time-to-event analyses over a one-year follow-up period.
Patients included had diagnostic data indicating T2D and HFrEF and documentation of initiation of a GLP1-RA within the specified study window. The comparator group comprised matched individuals with similar baseline characteristics who were not initiated on GLP1-RA. The authors report that, after matching, two well-balanced cohorts of 38,879 patients each were formed.
One-to-one propensity-score matching was performed to reduce confounding between GLP1-RA users and controls. The primary analytic approach was time-to-event analysis comparing one-year incidences between matched cohorts. The predefined primary endpoints were one-year all-cause mortality and a composite cardiovascular outcome. Secondary outcomes included the individual components of the composite (acute myocardial infarction, ischemic stroke, acute heart failure) and any in-hospital encounter. Absolute risk differences (ARD) with 95% confidence intervals and hazard ratios (HR) with 95% confidence intervals were reported.
The primary composite outcome comprised acute myocardial infarction, ischemic stroke, and acute heart failure. Secondary outcomes reported were the individual components of that composite and the occurrence of any in-hospital encounter during follow-up. All outcomes were assessed over a one-year period after cohort entry.
After propensity-score matching, each cohort included 38,879 patients. Key comparative results at one year were:
All-cause mortality: 4.2% in the GLP1-RA group versus 12% in controls; absolute risk difference (ARD) –7.8 (95% CI –8.17 to –7.41); HR 0.4 (95% CI 0.34–0.38).
Primary composite outcome (acute MI, ischemic stroke, acute HF): 32.5% in the GLP1-RA group versus 48.3% in controls; ARD –15.8 (95% CI –16.47 to –15.11); HR 0.6 (95% CI 0.57–0.60).
The GLP1-RA cohort was at lower risk for each reported secondary outcome compared with controls, per the authors’ results.
The manuscript presents these differences as associations observed in this large real-world dataset.
In this retrospective, propensity-matched analysis of TriNetX data, initiation of GLP1-RA therapy in patients with T2D and HFrEF was associated with substantially lower one-year all-cause mortality and lower incidence of a composite cardiovascular outcome compared with matched controls. The authors conclude that GLP1-RA therapy was associated with reduced risk of mortality and cardiovascular events in this population in routine clinical data.
It is important to note that these findings derive from observational, secondary data and the report is a medRxiv preprint that has not undergone peer review. The reported associations do not establish causation.
The authors report that the study was exempt from Institutional Review Board approval because it used de-identified secondary data from the TriNetX Research Network. They state that all analyses used de-identified data consistent with HIPAA de-identification standards. The authors declare no competing interests. The dataset and analyses are reported as available via the TriNetX platform; the manuscript provides a link to live.trinetx.com for data access.