Treatment of SAD
Treatment studies of light therapy have shown increasingly rigorous methodology with larger sample sizes, less diagnostic heterogeneity, longer treatment periods, and parallel instead of crossover designs. Wavelength of light used in light therapy was examined in two studies. In one study, the ultraviolet (UV) spectrum did not add to the therapeutic efficacy of light therapy [14*]. Because of the potential harmful effects of long-term W exposure, light therapy devices should have W filters that block wavelengths below 400 nm. In a comparison light box study, cool-white fluorescent lights were as effective as full-spectrum fluorescent lights [15], adding evidence to other studies showing that various light sources (including incandescent lights) are effective for treating SAD.

Devices other than light boxes were also studied for light therapy. Two recent studies, with the largest sample sizes in light therapy studies to date, used a light visor [16*,17*]. In both studies, there was no relationship between the intensity of light and various measures of response to treatment, despite the fact that very low intensity light (60 lux) was used. This contrasts to most light box studies where a dear intensity-response relationship is found. Several explanations may explain this discrepancy. The proximity of the visor light source to the eye may increase the amount of light that reaches the retina, as compared to a light box. Lux, a unit of illumination, may also not be the best measure of the biologic or therapeutic effect of light. There is increasing evidence that even low illumination can affect biologic parameters [18], so that for some patients, light as low as 100 lux may be therapeutically effective. Finally, although the response rate was high in both studies (over 60% by strictly defined criteria), a non-specific (placebo) effect of light therapy must also be considered. In this regard, a light box study by Eastman and associates [19**] using a non-light control condition (a negative ion generator that, unknown to subjects, was turned off), found no differences between the control condition and bright light treatment (7000 lux for 1 hour in the morning). However, the response rate for the bright light condition (29%) was unusually low compared to other treatment studies. The selection criteria and unusually sunny weather during
the course of their study may have excluded more light-responsive patients. Thus, the issue of placebo effects in light therapy remains unresolved.

The Seattle group conducted a series of studies investigating dawn simulation in SAD [20,21,22*]. Dawn simulation uses a device that gradually increases illumination exposure, while the patient is sleeping, to simulate a summer dawn during the winter. Significant improvement occurred using dawn simulation compared to various control conditions, despite a final illumination as low as 250 lux.

Two groups studying predictors for light therapy found that hypersomnia and hyperphagia predicted clinical response [23*,24,25*]. Another study, however, reported that only high consumption of sweets in the latter half of the day predicted response to treatment [26*]. Of interest is that prospective measures of sleep and eating were used in the latter study, whereas the other studies used global patient self-report.

Light therapy has been considered a rather benign treatment with few side effects. A systematic report of side effects to light therapy using a light visor showed that approximately 20% of patients reported mild side effects, including headache, eyestrain, and "feeling wired" [27]. A more controversial topic is the potential for prolonged bright light exposure to produce harmful effects on the retina. The intensities of light used in light therapy regimens are not considered harmful to the human retina based on short term studies, but the retinal effects of long term bright light exposure are not known. Some investigators have called for routine ophthalmologic evaluation prior to starting light therapy because of the small, potential risk of aggravating previously unrecognized retinal conditions (e.g. macular degeneration) [28*]. Others suggest ophthalmologic screening only in patients with a history of pre-existing retinal disease, patients taking highly photosensitizing drugs, and the elderly [29*]. Empiric data are still sparse, but a recently reported five-year prospective study of patients on chronic light therapy has not shown any significant clinical or electrophysiologic changes in the eyes [30].

Finally, antidepressant drugs are also being studied in SAD. An open study showed efficacy of bupropion in treating SAD [31*]. One case-study suggested that citaloprim, a selective serotonin reuptake inhibitor, was as effective as light therapy [321. Fluoxetine was reported to be as effective as light therapy for SAD [33*], and results from at least two double-blind studies of serotonin reuptake inhibitors in SAD will soon be available. What remains a question is whether a combination of medications and light therapy is more effective than either alone.