Direct comparison of total sleep deprivation and late partial sleep deprivation in the treatment of major depression.

Background: There are clinical as well as experimental indications that--contrary to what is generally assumed--late partial sleep deprivation (LPSD) is not as effective as total sleep deprivation (TSD) in the treatment of depression.

Method: We conducted a randomised balanced crossover study with 39 in-patients with major depression (mainly unipolar) in which both procedures LPSD and TSD were compared within a 1-week interval. Response was defined as a reduction of > or =30% in the 6-item Hamilton Depression Rating Scale and/or one of two self-rating scales (Adjective Mood Scale, Visual Analogue Scale).

Results: Overall response rate on the day after was low (0-53%, depending on the rating used). TSD proved slightly and in about half of the comparisons also significantly more effective than LPSD. In general, first treatments were more effective than second treatments; there were 10-20% second day responses; in up to 10% of the treatments patients worsened after sleep deprivation (using the same absolute criteria as for therapeutic response).

Limitations: Non-blind rating, intentional and unintentional napping (microsleep) was not recorded, mainly unipolar depressives.

Conclusions: Total sleep deprivation seems to be more effective than late partial sleep deprivation. We believe that there might be a dose-response relationship between hours of lost sleep and therapeutic effect within the range of 1 night.