Methods
Under a prospective, case-control design, 90 outpatients with probable or definite endogenous major depression (Schedule for Affective Disorders and Schizophrenia and Research Diagnostic Criteria) were stratified into abnormal and normal sleep subgroups (on the basis of an empirically validated electroencephalographic sleep profile) and more severe and less severe depression subgroups (on the basis of pretreatment Hamilton scores). Response to 16 weeks of treatment was analyzed for both intention-to-treat and completers (n=82) samples. Outcomes during a 36-month prospective follow-up were assessed with survival analyses.

Results
Abnormal sleep profiles and higher pretreatment depression severity were independently associated with poorer outcomes on several analyses. The association between sleep abnormality and cognitive behavior therapy response was not significant in the completers analyses, however, largely because of differential attrition. During follow-up, pretreatment depression severity was predictive of relapse and a lower recovery rate, whereas sleep abnormality was predictive of a lower recovery rate and a higher risk of recurrence.

Conclusions
Depressed patients characterized by higher severity and/or an abnormal electroencephalographic sleep profile were relatively less responsive to cognitive behavior therapy. These associations are hypothesized to result from a constellation of neurophysiologic disturbances that interfere with the acquisition, application, and implementation of the skills emphasized in cognitive behavior therapy.