Half of patients with chronic heart failure (CHF) have preserved left-ventricular ejection fraction (LVEF), but few treatments have specifically been assessed in such patients. In previous studies of patients with CHF and low LVEF or vascular disease and preserved LVEF, inhibition of the renin-angiotensin system is beneficial. We investigated the effect of addition of an angiotensin-receptor blocker to current treatments. Between March, 1999, and July, 2000, we randomly assigned 3023 patients candesartan (n=1514, target dose 32 mg once daily) or matching placebo (n=1509). Patients had New York Heart Association functional class II–IV CHF and LVEF higher than 40%. The primary outcome was cardiovascular death or admission to hospital for CHF. Anaysis was done by intention to treat. Median follow-up was 36·6 months. 333 (22%) patients in the candesartan and 366 (24%) in the placebo group experienced the primary outcome (unadjusted hazard ratio 0·89 95% Cl 0·77–1·03, p=0·118; covariate adjusted 0·86 0·74–1·0, p=0·051). Cardiovascular death did not differ between groups (170 vs 170), but fewer patients in the candesartan group than in the placebo group were admitted to hospital for CHF once (230 vs 279, p=0·017) or multiple times. Composite outcomes that included non-fatal myocardial infarction and non-fatal stroke showed similar results to the primary composite (388 vs 429; unadjusted 0·88 0·77–1·01, p=0·078; covariate adjusted 0·86 0·75–0·99, p=0·037). Candesartan has a moderate impact in preventing admissions for CHF among patients who have heart failure and LVEF higher than 40%. Published online Sept 1, 2003 http://image.thelancet.com/extras/03art7419web.pdf