Abstract Body: The cardioprotective effects of GLP1-Receptor Agonists (GLP1-RA) and SGLT2 Inhibitors (SGLT2i) are incompletely understood. We compared these two drug classes and their effect on HFpEF hospitalizations and examined the hypothesis that weight loss is a significant mediator of these agents’ HFpEF protection.

Methods: Using the VA Informatics and Computing Infrastructure, we generated a national cohort of veterans with type 2 diabetes on metformin who initiated SGLT2i, GLP1-RA or Insulin Glargine for the first time from 1/1/18 to 12/31/21. The index date was the date of first outpatient prescription fill. We defined 6-month weight change as change from baseline weight (closest outpatient value prior to index date within 1 year) to 6 months (all averaged outpatient weights from index date to 6 months). Those with prior prescriptions for the study drugs from 1/1/08 and missing baseline or 6-month weights were excluded. HFpEF hospitalizations were defined by a previously-validated natural language processing algorithm using inpatient HF diagnostic codes and cardiac imaging dated closest to and within 180 days of the new code with ejection fraction ≥50%. Follow up was through 3/31/23. We performed mediation analyses to separate the direct effects (independent of 6-month weight change) and indirect effects (mediated through 6-month weight change) of the drug classes on time to HFpEF hospitalizations, adjusted for demographics, comorbidities, baseline labs and other medications, and diabetes severity.

Results: Overall, 161,405 veterans initiated one of the study drugs and 105,439 had weight data. After excluding 760 veterans with HFpEF hospitalizations in the first 6 months, 104,679 were included in the analysis. Baseline demographics included mean age 64±11 years, 94% male and 75% white. Baseline BMI was 33±6 kg/m². There were 1494 HFpEF hospitalizations/ 256,101 years of follow-up. Figure 1 shows weight over time by study drug class. Table 1 shows mediation analyses results.

Conclusion: GLP1-RA lowered the risk of HFpEF hospitalizations versus the active comparator insulin glargine. This cardioprotection was not mediated by weight loss. GLP1-RA and SGLT2i show equivalent risk reduction. These agents' direct cardiac and vascular effects deserve further exploration.