Résumé : BACKGROUND: The present study further examined relationships between postdexamethasone cortisol plasma values and sleep electroencephalogram (EEG) parameters. METHODS: The dexamethasone suppression test (DST) and polysomnographic recordings were performed in a sample of 300 inpatients with primary major depressive disorder (MDD) (102 men and 198 women, mean age 44 +/- 12 years, range 20-74 years) consecutively admitted to Erasme Hospital (Brussels, Belgium) between 1981 and 1992. RESULTS: The DST was abnormal in 40% of the sample. Postdexamethasone cortisol plasma values at 4:00 PM were significantly influenced by age, but not by gender. They were also significantly and positively correlated with weight loss, total scores on the Hamilton Depression Rating Scale, total scores on the Newcastle Scale, percentage of awakenings during sleep, and percent of stage 1. They were significantly and negatively correlated with percent of stage 2, slow-wave sleep, and REM sleep. Multiple regression analyses were conducted in two successive steps. First among clinical variables, only age and depressive symptom severity remained correlated with postdexamethasone plasma cortisol values. In the second step, with age and severity held constant, postdexamethasone plasma cortisol values were positively associated with amount of wake time and stage 1, and negatively with amount of slow-wave sleep. CONCLUSIONS: These findings provide further indirect support for an overarousal state in MDD with sympathoadrenal system hyperactivity and impaired sleep continuity. They also underline the importance of taking into account various clinical confounding factors in the interpretation of both DST and sleep EEG results.