Fifteen patients with left ventricular heart failure (LVF) without known breathing disorders during sleep had full-night recordings of sleep and breathing to study the incidence and impact of the apnea of Cheyne-Stokes breathing. This group showed a marked degree of sleep-related breathing abnormalities, 40% demonstrating Cheyne-Stokes breathing with five or more central apneas per hour of sleep. Cheyne-Stokes breathing during sleep in patients with LVF predicted an increased short-term mortality rate. All six patients with LVF and Cheyne-Stokes breathing with more than five apneas per hour of sleep were dead within six months, while only three of nine patients without recurrent apnea died within six months, a significant difference (P less than .05) even in this small group. Among seven patients with LVF studied with the polysomnogram, there were statistically significant differences between the Cheyne-Stokes and non-apnea groups in total sleep time, awakening per hour, and the number of arterial desaturations. Although sleep disturbances have been anecdotally described in patients with LVF, no previous investigation has determined the incidence and impact of Cheyne-Stokes breathing during sleep in LVF. Our findings that Cheyne-Stokes breathing predicts an adverse short-term mortality rate confirm the clinical impression that Cheyne-Stokes breathing is a poor prognostic sign in LVF.