[Polysomnographic aspects of antidepressive therapy]

Sleep disorders are very common in depression. They have been quantitatively and qualitatively described by polysomnographical recordings. It is of interest to know how treatments act on polysomnographical data. In this article, we propose to evaluate five treatment approaches proposed for depressive illness. Pharmacological treatment induces marked changes in sleep continuity, sleep architecture and REM sleep. However, no specific sleep profiles emerge neither for each treatment class nor for molecules within the same pharmacological class. But actually, sleep data cannot be viewed as markers of treatment response. Psychotherapeutic interventions have only few effects on sleep of depressed patients. Treatment efficiency does not seem to be correlated to abnormal sleep parameters. Sleep deprivation induces marked changes in nearly all sleep parameters and in temporal distribution of sleep stages during the night. Actually, the efficiency of sleep deprivation cannot longer be explained by suppression of REM sleep. Sleep deprivation has only a transient effect and treatment indications are therefore secondary. Sleep parameters do not distinguish responders from non-responders. Sleep deprivation shows that there is a depressogenic effect of sleep in the end of the night. Bright light therapy shows marked changes in sleep continuity parameters. Among all studies that examine the impact of treatment on sleep EEG, ECT has received little attention. The few studies available are either case studies or with poor effectifes++. For this reason and because of methodological bias, results are heterogeneous and no definite conclusions can be drawn. But all of them agree that ECT modifies sleep EEG. So, changes in polysomnographical data cannot predict response to any treatment. Prospective sleep studies are difficult to realise on a great number of patients explaining absence of treatment predictors.