Abstract:
Background:
Heart failure (HF) following an acute myocardial infarction (post-MI HF) has been studied as an additional sub-type of HF to broaden the indications for HF drugs. Post-MI HF and HFrEF are pathophysiologically similar and share pharmacotherapies. In this meta-analysis, we examined the concordance between all-cause mortality data for drugs indicated for HFrEF and post-MI HF. We used our analysis to calculate the projected all-cause mortality hazard ratios (HRs) for the pending dapagliflozin (DAPA-MI) and empagliflozin (EMPACT-MI) post-MI HF trials.
Methods:
Using CenterWatch and UpToDate, we identified all FDA-approved drugs for NYHA Class II to IV HFrEF. We searched each of these drugs on FDALabel and ClinicalTrials.gov to identify their registration trials measuring all-cause mortality for HFrEF and, if available, in the post-MI setting—including trials where participants displayed a left ventricular ejection fraction of <40% (“post-MI HF”). For each of the included studies, we extracted the all-cause mortality HRs, their 95% confidence intervals, and the control-group used. For all drugs studied in both indications, we plotted the all-cause mortality HRs for HFrEF against those for post-MI (HF) and calculated the linear regressions.
Results:
This meta-regression pooled data from 29 completed trials underlying 20 drugs. Two pending trials were also analyzed. Nine drugs (metoprolol, carvedilol, spironolactone, eplerenone, sacubitril-valsartan, lisinopril, enalapril, valsartan, losartan) had all-cause mortality data in both HFrEF and post-MI generally, with a linear coefficient of determination of 0.93. Five of these drugs (carvedilol, eplerenone, sacubitril-valsartan, valsartan, losartan) were studied in both HFrEF and non-acute post-MI HF, displaying a linear coefficient of determination of 0.99. Using our model, we predict the all-cause mortality HRs that will be observed in the EMPACT-MI and DAPA-MI trials will be 0.85 and 0.89, respectively.
Conclusions:
In this meta-regression of registration trials for drugs studied in both HFrEF and post-MI (HF), all-cause mortality effects were highly concordant. We also find asymmetries in the assessment of HF drug indications, whereby drugs are seldom assessed for an all-cause mortality benefit in both HFrEF and in post-MI HF. Future studies may use these results to guide future HF RCT development.