This study was aimed at investigating the relationships between sleep EEG abnormalities and hypothalamo pituitary adrenal (HPA) and hypothalamo pituitary thyroid (HPT) disturbances in major depressive disorder. Post dexamethasone (DXM) cortisol levels and the dual TSH response to 08:00 h and 23:00 h TRH administration were determined after a 2 weeks wash-out period in a group of 113 DSM-IV major depressed patients (72 females aged 44.3+/-13.0 and 41 males aged 45.7+/-11) who were consecutively admitted to undergo sleep EEG recordings. Post-DXM cortisolemia, 08:00 and 23:00 post-TRH TSH values, time spent in rapid eye movement sleep (REMS), in slow wave sleep (SWS), and in stage 2 as well as time awake after sleep onset were introduced in a principal component (PC) analysis. The four 3 PC scores explaining up to 74% of the data set were further calculated for each patients and used in a cluster analysis. A three-cluster solution was retained. Controlling for the effects of age and gender, patients belonging to these three clusters could clearly be differentiated on the basis of their neuroendocrine responses and on their sleep EEG profiles. Compared to the two other clusters, cluster I (n=26) patients showed the most severe sleep continuity disturbances. Post-DXM cortisol escape and sleep architecture disturbances (consisting of a shortening of REMS latency and a decreased SWS) identified patients belonging to cluster II (n=39). Patients in cluster III (n=48) had the lowest TSH response to TRH and the less marked sleep EEG alteration. Clinical or demographic variables were unable to differentiate the three clusters. Our results suggest that different biological dysfunctions could each underlie particular neuroendocrine and sleep EEG disturbances in major depression.