We studied seven male patients with moderate to severe congestive heart failure (CHF) [left ventricular ejection fraction (LVEF) = 22.4 +/- 6.7; mean +/- SD] in a double-blind crossover trial to determine the effects of temazepam 15 mg on arousability, sleep architecture, Cheyne-Stokes respiration (CSR) and nighttime oxygen saturation. Sleep architecture was not markedly improved with temazepam. There was no significant change in total sleep time (TST) (383.1 +/- 14.1 minutes to 396.6 +/- 15.4 minutes, p = ns) (mean +/- SE, placebo vs. temazepam) or total wake time (TWT) (96.9 +/- 14.0 vs. 81.4 +/- 14.0 minutes, p = ns). Sleep stage proportions did not change appreciably except for a reduction in stage 1 sleep (6.7 +/- 1.2% vs. 4.0 +/- 1.0%, p < 0.05). Microarousals per hour of sleep decreased with temazepam (21.1 +/- 2.7/hour vs. 13.9 +/- 2.1/hour placebo, p < 0.05), with the largest change occurring in stage 2 (24.9 +/- 5.4/hour vs. 15.0 +/- 3.1/hour, p < 0.05). Wake time during sleep (WDS) was reduced from 82.5 +/- 11.7 minutes to 54.5 +/- 9.4 minutes, p < 0.03. Daytime alertness was improved with temazepam as was indicated by an increase in mean latency to sleep [multiple sleep latency test (MSLT) = 7.1 +/- 2.4 vs. 5.7 +/- 2.0 minutes, p < 0.04) on days following treatment with temazepam. There was no significant change in CSR as a percentage of TST (38.7 +/- 13.6% vs. 32.5 +/- 11.8%, p = ns). However, the apnea/hypopnea index (AHI) (10% filter) was decreased in stage 1 (28.1 +/- 9.7/hour vs. 15.6 +/- 8.2/hour). Overnight oxygen saturation did not change with temazepam (95.1 +/- 0.6% both nights) and the percentage of TST spent below 90% oxygen saturation was minimal for both conditions (1.5 +/- 1.1% vs. 2.2 +/- 1.7%, p = ns). We conclude that CHF patients with CSR experience frequent arousals and that these arousals can be reduced with temazepam. There was an improvement in daytime somnolence. There was no worsening of nighttime oxygen saturation.