Although several studies have shown that sleep deprivation is associated with increased slow wave sleep during the recovery night, the effects of sleep deprivation on cortisol and growth hormone (GH) secretion the next day and recovery night have not been assessed systematically. We hypothesized that increased slow wave sleep postsleep deprivation is associated with decreased cortisol levels and that the enhanced GH secretion is driven by the decreased activity of the HPA axis. After four consecutive nights in the Sleep Laboratory, 10 healthy young men were totally deprived of sleep during the fifth night, and then allowed to sleep again on nights six and seven. Twenty-four hour blood sampling was performed serially every 30 minutes on the fourth day, immediately following the previous night of sleep and on the sixth day, immediately after sleep deprivation. Eight-hour sleep laboratory recording, including electroencephologram, electro-oculogram and electromyogram. Plasma cortisol and GH levels using specific immunoassay techniques. Mean plasma and time-integrated (AUC) cortisol levels were lower during the postdeprivation nighttime period than on the fourth night (P < 0.05). Pulsatile analysis showed significant reduction of both the 24 h and daytime peak area (P < 0.05) and of the pulse amplitude (P < 0.01), but not of the pulse frequency. Also, the amount of time-integrated GH was significantly higher for the first 4 h of the postdeprivation night compared to the predeprivation night (P < 0.05). Cross-correlation analyses between the absolute values of the time-series of each hormone value and percentage of each sleep stage per half hour revealed that slow wave sleep was negatively correlated with cortisol and positively correlated with GH with slow wave sleep preceding the secretion of these hormones. In contrast, indices of sleep disturbance, i.e. wake and stage 1 sleep, were positively correlated with cortisol and negatively correlated with GH. We conclude that sleep deprivation results in a significant reduction of cortisol secretion the next day and this reduction appears to be, to a large extent, driven by the increase of slow wave sleep during the recovery night. We propose that reduction of CRH and cortisol secretion may be the mechanism through which sleep deprivation relieves depression temporarily. Furthermore, deep sleep has an inhibitory effect on the HPA axis while it enhances the activity of the GH axis. In contrast, sleep disturbance has a stimulatory effect on the HPA axis and a suppressive effect on the GH axis. These results are consistent with the observed hypocortisolism in idiopathic hypersomnia and HPA axis relative activation in chronic insomnia. Finally, our findings support previous hypotheses about the restitution and immunoenhancement role of slow wave (deep) sleep.