Nine patients with stable cardiac failure and mean left ventricular ejection fraction of 30% were investigated. All had previously been prescribed a benzodiazepine hypnotic by their home physicians, but the medication had been discontinued for at least 1 month. Subjects were monitored under three conditions: 1) without any sleeping medication, 2) during nasal CPAP administration and 3) at two points during a month-long administration of the benzodiazepine that had initially been prescribed to them. Overall, the benzodiazepine hypnotic improved the sleep fragmentation noted in these patients by decreasing the arousal index from a mean of 18 +/- 6 per hour at baseline to a mean of 9 +/- 6.5 per hour after one month of benzodiazepine therapy. Total nocturnal sleep time was consequently improved [baseline mean nocturnal total sleep time: 313 +/- 27.3 minutes; benzodiazepine mean nocturnal total sleep time: 350 +/- 17.3 minutes (p < 0.0003)], as was sleep efficiency. However, the benzodiazepine hypnotic had no significant effect on central hypopneas or apneas [baseline mean respiratory disturbance index (RDI): 20.5 +/- 5.85 events/hour; mean RDI after 1 month of drug intake: 21.3 +/- 2.5 events/hour]. Nasal CPAP was also ineffective on the disordered breathing. In this group of subjects, respiration was even significantly worsened with nasal CPAP compared to baseline, as indicated by RDI (p < 0.0001), lowest SaO2 (p < 0.0001) and total nocturnal sleep time (p < 0.0001) measurements.