Event rates in major Phase 3 heart failure trials over the last 20 years

To determine whether control-arm event rates in heart failure (HF) randomized clinical trials (RCTs) published in New England Journal of Medicine from 2004 to 2024 declined over time, despite intensification of background therapy.

We identified Phase 3 HF RCTs published in New England Journal of Medicine (2004-24). Annualized event rates were calculated as events divided by patients multiplied by follow-up to the primary endpoint. Temporal trends were analysed with weighted least squares (weights equal to the number of control-arm patients), with adjustment for follow-up duration.

Thirty-eight trials met criteria; 31 enrolled HF with reduced ejection fraction (HFrEF; 82%). In HFrEF control arms, median age was 67 years [interquartile range (IQR) 64-69], women were 24% (IQR 21-30), left ventricular ejection fraction (LVEF) was 28% (25-31), N-terminal pro-B-type natriuretic peptide (NT-proBNP) was 1700 ng/L (1273-2879), and the New York Heart Association (NYHA) class III-IV was 46% (29-71). Background therapy included β-blockers 91% (82-93), angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (ARBs)/angiotensin receptor-neprilysin inhibitors 93% (90-97), and mineralocorticoid receptor antagonists (MRAs) 52% (42-59). The median control-arm size was 602 patients (308-1533). Across years, therapy use rose (β-blockers +1.04 percentage points per year, P < .001; MRAs +2.05 percentage points per year, P < .001), LVEF increased (+.22% per year, P = .012), NT-proBNP increased (∼+11.4% per year on the log scale, P = .004), and follow-up tended to shorten (P = .052). In control arms, all-cause mortality showed no temporal decline [unadjusted slope +.0019 per year; 95% confidence interval (CI) -.0013 to +.0051; P = .252]; after adjusting for follow-up, the slope was +.0006 per year (P = .711). Longer follow-up was associated with lower annualized mortality (coefficient -.0195 per year; P = .032). Cardiovascular mortality was stable (unadjusted +.0003 per year; 95% CI -.0036 to +.0042; P = .898; follow-up-adjusted -.0012 per year; P = .557). The composite of all-cause death or HF hospitalization increased unadjusted (+.0180 per year; 95% CI +.0063 to +.0298; P = .011) but was not significant after follow-up adjustment (+.0113 per year; P = .111). Enrichment intensity did not rise linearly (coefficient +.032 criteria per year; P = .111), whereas natriuretic-peptide cut-offs were adopted more often [odds ratio (OR) per decade 14.09; 95% CI 1.93-102.75; P = .009). Higher age related to higher mortality (coefficient +.0039 per year; P = .043). An interaction between year and log-transformed NT-proBNP indicated risk-dependent temporal patterns (P = .003).

In major HFrEF trials, control-arm mortality did not decline from 2004 to 2024 despite greater uptake of evidence-based therapy. Risk-enriched enrolment and shorter follow-up likely counterbalanced therapeutic gains.