Relationships between fatty liver and coronary heart disease (CHD) and stroke risk remain ill defined. We investigated whether fatty liver is a predictor of CHD and stroke risk. Until December 2000 we followed 2,024 atomic bomb survivors (775 men: 62.0 +/- 9.9 years old; 1,249 women: 63.2 +/- 8.4 years old) who had basic examinations between November 1990 and October 1992 for clinical and laboratory CHD risk factors and fatty liver and who were initially free of CHD and stroke. Forty-nine cases of CHD and 84 cases of stroke were observed. At the time of the baseline examinations, significant clinical associations were found between fatty liver and obesity (p<0.001), hypertension (p<0.001), dyslipidemia (p<0.001), and glucose intolerance (p<0.001). A slight but nonsignificant association was found between fatty liver and hyperuricemia (p=0.07) as well. By using multiple Cox regression analyses, age (relative risk [RR] 1.05, 95% confidence interval [CI] 1.01-1.08), smoking (RR 2.20, 95% CI 1.02-4.74), hyperuricemia (RR 2.30, 95% CI 1.08-4.89), and fatty liver (RR 2.53, 95% CI 1.06-6.06) were shown to be significant predictors of CHD, whereas age (RR 1.08, 95% CI 1.06-1.10), smoking (RR 2.06, 95% CI 1.14-3.72), and hypertension (RR 2.14, 95% CI 1.38-3.30) predicted stroke risk. Fatty liver, which clusters clinical and laboratory CHD risk factors, is an independent predictor of CHD, but not of stroke. Fatty liver should be followed as a feature of metabolic syndrome, with the aim of preventing CHD.