Bright-Light Treatment of Winter Depression


Synopsis: This is one of the few well-designed placebo-controlled studies of bright-light therapy for seasonal affective disorder. The study demonstrated that bright-light therapy is an effective treatment for seasonal depression and that morning treatment may be superior to evening treatment.

Source: Eastman CI, et al. Bright-light treatment of winter depression. Arch Gen Psychiatry 1998;58:883-889.

Bright-light therapy (also known as photo-therapy) has become an increasingly popular treatment modality for patients with seasonal shifts in mood states, such as seasonal affective disorder (SAD). As Eastman and associates acknowledge, the efficacy of bright-light therapy has been difficult to definitively prove because of the difficulty in designing a suitable placebo control. A placebo control is necessary in studies of depression, due to the high response rates (20-40%) of depressed patients to placebo treatment. Eastman et al used a novel device, a negative ion generator, which, when deactivated, served as the placebo control to the bright-light therapy. The deactivated ion generators had flashing lights and made a hissing sound, similar to active ion generators. The subjects were told that they might receive treatment with an active or deactivated ion generator. The bright-light therapy provided 6000 lux (at a distance of 15 inches from light box to subject). Morning light was also compared to evening light. All subjects met generally accepted criteria for SAD. The active study period was four weeks, and subjects were told that they would be assigned to one of four treatment groups (morning or evening ion generator or morning or evening bright-light therapy). No subjects actually received evening ion generator treatment, and few subjects received active ion generator therapy. Thus, three treatment groups were created (morning light [n = 33], evening light [n = 31], and placebo [deactivated ion generator n = 32]). All treatments lasted for 90 minutes per day, six days per week. Eastman et al monitored the expectations of the subjects as a measure of the effectiveness of the placebo control. Sleep schedules were tightly controlled, with many subjects waking up earlier than usual to leave enough time for the 90-minute treatment sessions. Adherence to the sleep schedule was closely monitored. The statistical methods were appropriate for the study design. The results showed that all groups had equal expectations that their treatment would work, suggesting that the method of placebo control was appropriate and effective. An analysis of the sleep schedule data revealed that all treatment groups received approximately one hour less sleep than before the study. In terms of the efficacy of bright-light therapy, it required several separate and more stringent definitions of response to demonstrate statistically significant results. For example, all three treatments produced substantial reductions in depression ratings over the four-week study. Furthermore, a repeated-measures analysis of variance (ANOVA) of change scores in depression ratings compared to baseline failed to show any difference between the treatment groups. Using a standard criterion of response as a 50% reduction in depression ratings at weeks 3 and 4 also failed to reveal any differences among the groups. However, statistically and clinically relevant differences among the groups emerged when subjects were categorized by a strict definition of "complete remission" for both weeks 3 and 4. The morning bright-light therapy produced many more complete responders (55%) than either the evening bright-light therapy (28%) or the placebo treatment (16%). Eastman et al justified the multiple serial data analysis methods on the basis of the unusual differences in the distribution of the data points between the morning light therapy and the placebo treatment.


This is one of the few well-designed placebo-controlled studies of bright-light therapy for SAD. Although some post-hoc data analysis was required to demonstrate differences among the groups, Eastman et al have convincingly demonstrated that bright-light therapy works and that morning treatment may be superior to evening treatment. Replication of this study with larger numbers of subjects is warranted. Future studies should use the criterion of complete remission from depressive symptoms as a measure of response, rather than simply reductions in rating scale scores. Clinically, bright-light therapy is relatively noninvasive, simple to use, and affordable for many patients. Some insurance companies are now reimbursing patients for light boxes, which generally cost between $200 and $500. There are few side effects, other than activation, headache, and eye strain, and many patients prefer phototherapy to antidepressant medication. Recognizing SAD in the office setting remains a challenge to many primary care physicians. Along with routine questions regarding sleep, appetite, energy, and mood, one needs to ask about seasonal shifts in these mood and neurovegetative symptoms. It is also important to remember that many SAD patients also have bipolar disorder and that bright-light therapy can occasionally trigger hypomania, mania, or mood cycling. Therefore, inquiring about a history of abnormal mood elevation is important.