Covariate adjusted reanalysis of the I-Preserve trial

BackgroundThe CHARM-Preserved trial suggested that the renin-angiotensin system (RAS) inhibitor candesartan might have been beneficial in heart failure with preserved ejection fraction (HFpEF); however, this hypothesis was not supported by the findings of I-Preserve with irbesartan.AimsTo re-analyse the results of I-Preserve, adjusting for imbalances in baseline variables that may have influenced the trial outcomes.MethodsCox proportional hazards models with covariate adjustment for baseline variables, including age, sex, medical history, physiological and laboratory variables.ResultsIn I-Preserve, 763 (37.0%) participants in the placebo group and 742 (35.9%) in the irbesartan group experienced the primary composite outcome (death from any cause or hospitalization for heart failure, myocardial infarction, unstable angina, arrhythmia, or stroke). The prespecified analysis of this outcome, stratifying for the use of ACEi at baseline, gave a hazard ratio (HR) of 0.95 (95% confidence interval, 0.86–1.05); p = 0.35. Adjusting the effect of treatment for key prognostic baseline variables, gave a HR of 0.89 (0.80–0.99); p = 0.033. Similar findings were observed for the composite of cardiovascular death or HF hospitalization.ConclusionAdjusting for imbalances in baseline variables that influence outcomes (or the response to therapy or both) can improve the power around the estimate of the effect of treatment and may alter its statistical significance. Along with the CHARM-Preserved results, these findings suggest that angiotensin-receptor blockers may have a modest effect in HFpEF.