<i>Background:</i> Eating disorders are thought to be risk factors for cardiac sudden death secondary to arrhythmia. Results in previous studies on QT interval and QT dispersion, markers of fatal arrhythmia, have been inconsistent. <i>Methods:</i> We prospectively examined 179 female eating disorder patients, being over 18 years old and diagnosed according to the DSM-IV criteria between January 1995 and December 2002, and 52 healthy women. Patients with abnormal plasma electrolytes or taking medications that might influence the electrocardiogram (ECG) were excluded from the study. QT intervals were corrected for heart rate using Bazett’s formula and the nomogram method, which is more reliable at extremely low heart rates than Bazett’s formula. QT dispersion was measured as the difference between the longest and shortest QT intervals. QT intervals and QT dispersion in each patient group were compared with those in the control group. <i>Results:</i> The 164 eligible patients consisted of 43 patients with anorexia nervosa restricting type, 35 with anorexia nervosa binge eating/purging type, 63 with bulimia nervosa purging type, and 23 with bulimia nervosa nonpurging type. There was no significant difference in age between eating disorder patients and controls. QT interval and QT dispersion were significantly longer in all eating disorder subtypes than in the control group. QT interval and QT dispersion were significantly correlated with the rate of body weight loss in bulimia nervosa. <i>Conclusions:</i> QT interval and QT dispersion were prolonged in both anorexia nervosa and bulimia nervosa. Examination of ECG in eating disorder patients without extremely low body weight also appears to be clinically important.