Worsening heart failure is not a single endpoint: a severity-gradient perspective from digoxin in rheumatic heart disease.

The composite of cardiovascular or all-cause death and worsening heart failure (WHF) has become the principal efficacy endpoint of heart failure trials. Yet WHF is not one event: it spans outpatient oral-diuretic escalation, urgent intravenous-diuretic visits, hospitalisation, and HF-related death-tiers that differ markedly in frequency, ascertainment, and prognostic weight. A composite WHF benefit therefore cannot be properly interpreted without knowing which tier carries the signal, and we argue that this tier is both phenotype- and setting-dependent. Two trials published on the same day in 2026 make the point with unusual clarity. In the Dig-RHD trial of digoxin in rheumatic heart disease, the benefit concentrated in outpatient diuretic events, with neutral results at hospitalisation and death; in a contemporaneous meta-analysis of digitalis glycosides in reduced or mildly reduced ejection fraction, the same drug class reduced worsening heart failure mainly through hospitalisation, again without a mortality benefit. We use this contrast to propose a severity-gradient framework for reporting and interpreting worsening heart failure, and discuss its implications for trial design, for the translation of guideline-based evidence across health systems, and for the use of inexpensive therapies such as digoxin in low- and middle-income settings.