The Candesartan Antihypertensive Survival Evaluation in Japan (CASE-J) trial assessed cardiovascular outcomes in high-risk hypertensive patients receiving either candesartan or amlodipine. The aim of this study was to examine the role of pre-existing diabetes or obesity on these outcomes as a sub-analysis of the trial. We examined the influence of pre-existing diabetes on cardiovascular morbidity and mortality using a multivariate Cox regression model. The cardiovascular morbidity and mortality of candesartan and amlodipine were compared between subgroups with or without pre-existing diabetes or by body mass index (BMI) category, and new-onset diabetes was compared by BMI category. Pre-existing diabetes greatly increased the cardiovascular mortality and morbidity, regardless of the allocated drugs. Furthermore, all-cause mortality was significantly higher with amlodipine than with candesartan among patients with BMI >or=27.5 kg m(-2) (adjusted hazard ratio (HR)=0.32; range=0.13-0.75; P=0.009). New-onset diabetes occurred significantly less frequently with candesartan than with amlodipine, with an adjusted HR of 0.66 (P=0.043). Furthermore, the increase in new-onset diabetes was dependent on BMI among patients receiving amlodipine, whereas no such dependency was observed for candesartan (interaction P=0.016). In conclusion, preexisting diabetes increased the risk of experiencing a cardiovascular event among high-risk Japanese hypertensive patients. Candesartan treatment may suppress all-cause death and reduce the incidence of new-onset diabetes in patients with obesity.