A population-based cohort study was conducted to assess the relationship between total mortality and self-reported sleep patterns as regards not only to sleep duration but also subjective sleep quality. A total of 5,322 inhabitants in Gifu Prefecture, Japan, completed a self-administered questionnaire on health status and lifestyles including habitual sleep patterns, and were followed-up for an average of 11.9 years. Relative risks were computed by using Cox proportional hazards models. Both longer and shorter sleep, compared to 7-8 hour-sleep, was related to significantly increased risk of total mortality in males (relative risk [RR] for > or = 10 hours = 1.94, and RR for < 7 hour = 1.90), but not in females. Females complaining of poor awakening state experienced a higher mortality risk compared to those who woke up normally (RR: 1.97). Males who usually fell asleep easily showed a marginally lower mortality risk compared to those who fell asleep normally (RR: 0.70). Female users of sleeping pills were at an elevated risk (RR: 1.89). These findings were almost unchanged after adjustment for sleep duration and other confounders. Poor self-reported quality of sleep seemed to be associated with an increased risk of mortality independently of sleep duration.