This study was designed to determine the impact of central sleep apnea with or without Cheyne-Stokes respiration (CSR) on morbidity and mortality. Central sleep apnea was found in 77 male general medical ward in-patients. Cheyne-Stokes respiration was found in 49 of the 77 men; in 15 men, CSR was severe, ie, > or = 25 percent of the night spent in CSR, in 34 men CSR was mild (1 to 25 percent CSR). Twenty-eight men had central sleep apnea but no CSR. An additional 31 patients had no sleep apnea and no CSR. The patients with severe CSR had more central apneas, more, but shorter desaturations, more awakenings and more wake time during the night, but spent more time in bed than those with no CSR or no apnea. Radiographic evidence was consistent with an association of CSR and heart failure. In addition, patients with severe CSR were at almost twice the risk of dying compared with those with no apnea and had a shorter survival time. Nevertheless, we could not confirm that CSR was an independent predictor of elevated mortality risk, implying that some other factors specific to severe CSR predispose these patients to shorter survival time.