The standard methods of scoring sleep patterns do not ensure an accurate clinical impression of sleep quality. This is important especially in depressed insomniacs because persistent poor sleep increases the likelihood of recurrent depressive episodes. Changes in cyclic alternating patterns (CAP) in sleep have been shown to reflect corresponding changes in sleep quality. We evaluated the effects of nefazodone on CAP and standard sleep architecture in depressed insomniacs. The study was a single-center, single-blind, 6-week treatment of nefazodone hydrochloride followed by placebo withdrawal in 16 subjects meeting the DSM-IV criteria for depression who had a score of at least 18 on the 17-item Hamilton Depression Rating Scale, with insomnia-related items 4, 5, and 6 having a total score of 3 or greater. A mean daily dose of 339.1 ± 141.7 mg at endpoint of nefazodone significantly reduced Hamilton Depression Scores from 21.7 ± 3.0 on baseline to 5.8 ± 5.3 (P< .05) by the end of the study. Polysomnography showed an improvement in sleep latency and sleep efficiency (P< .05), but no alterations in rapid-eye-movement or slow-wave sleep. Subjective estimates of sleep quality improved throughout the study, but CAP rates did not show a significant improvement. The disparity between CAP rates and sleep quality in depressed insomniacs is discussed.