Summary: In 4017 patients with acute myocardial infarction and impaired left ventricular systolic function, but no diabetes or chronic heart failure, dapagliflozin 10 mg won more often than placebo on a hierarchical cardiometabolic composite (win ratio 1.34; 95% CI 1.20 to 1.50), driven mainly by the cardiometabolic components. Cardiovascular death or heart failure hospitalisation did not differ (2.5% vs 2.6%; hazard ratio 0.95; 95% CI 0.64 to 1.40).
PICO Summary
| Element | Detail |
|---|---|
| Population | 4017 patients with acute myocardial infarction and impaired left ventricular systolic function, without prior diabetes or chronic heart failure; international, registry-based, randomised, double-blind, placebo-controlled trial; approximately 1 year of treatment. |
| Intervention | Dapagliflozin 10 mg once daily (n=2019). |
| Comparison | Placebo once daily (n=1998). |
| Outcome | Hierarchical composite of death, heart failure hospitalisation, non-fatal myocardial infarction, atrial fibrillation or flutter, type 2 diabetes, NYHA class at the last visit, and body weight decrease of 5% or more at the last visit: win ratio 1.34 (95% CI 1.20 to 1.50); p<0.001, mainly driven by the added cardiometabolic outcomes. Cardiovascular death or heart failure hospitalisation: 50 of 2019 (2.5%) vs 52 of 1998 (2.6%); HR 0.95 (95% CI 0.64 to 1.40). Rates of other cardiovascular events were low, without nominally significant differences. No safety concerns were identified. The win ratio analysis was adopted during the trial because of low event accrual. |
Expert Commentary
My verdict is neutral. The headline win ratio of 1.34 (95% CI 1.20 to 1.50) is statistically convincing but clinically thin. The analysis was changed during the trial to a hierarchical composite because too few events were accruing, and the abstract states that the result was driven mainly by the added cardiometabolic components, which include new type 2 diabetes and a weight decrease of 5% or more. On the conventional outcome of cardiovascular death or heart failure hospitalisation, events occurred in 2.5% on dapagliflozin and 2.6% on placebo (hazard ratio 0.95; 95% CI 0.64 to 1.40). The limitation is therefore low power for hard events: with 50 and 52 events over approximately 1 year, the trial can neither show nor exclude a worthwhile effect. No safety concerns were identified. The trial was funded by the manufacturer. Can I use this with my patients? Not yet as a routine treatment after myocardial infarction. In a patient without diabetes or heart failure, I would not start dapagliflozin after myocardial infarction on this evidence alone, although it appears safe where there is another indication. Longer follow-up of this cohort for heart failure and death would be the most useful next step.
References
James S, Erlinge D, Storey RF, et al. Dapagliflozin in myocardial infarction without diabetes or heart failure. NEJM Evid. 2024;3(2):EVIDoa2300286. doi:10.1056/EVIDoa2300286. PMID: 38320489.

