The causes and risks of death, and role of severity of obstructive sleep apnea (OSA), obesity, and pulmonary function (PF) in OSA patients treated with or without continuous positive airway pressure (CPAP) have been questioned. Using the Okinawa Nakamura Sleep (ONSLEEP) registry, we studied 4000 patients with an obstructive apnea-hypopnea index (AHI) of >5 events/h. Kaplan–Meier analysis determined survival rates based on use of CPAP therapy and OSA severity. Multivariate Cox proportional hazard analysis determined effects of AHI, body mass index (BMI), PF, and use of CPAP. A total of 135 deaths (3.4%) were registered at the end of follow-up period (62.0 ± 43.4 months, mortality rate 8 per 1000 patient-years). Main causes of death were cardiovascular diseases. Multivariate predictors of mortality were male sex, age, BMI, and PF. Although both AHI and use of CPAP tended to affect prognosis, both effects were insignificant. Mortality rate was ∼9-fold higher in non-CPAP users with pulmonary impairment (PI) than non-CPAP users with normal lung function. The mortality rate of non-obese (BMI < 25 kg/m2) OSA patients with PI was ∼10-fold higher than that of non-obese patients with normal PF. In patients with OSA, body weight and PI, but not AHI, independently predict mortality. CPAP therapy reduced the risk of death in OSA. The risk of mortality was lowest, as 3 per 1000 patient-year, with normal PF and non-CPAP treatment.