ORIGINAL ARTICLE Morning vs Evening Light Treatment of Patients With Winter Depression Alfred J. Lewy, MD, PhD; Vance K. Bauer, MA; Neil L. Cutler, BA; Robert L. Sack, MD; Saeeduddin Ahmed, MD; Katherine H. Thomas, MD; Mary L. Blood, MS; Jeanne M. Latham Jackson, MD Background: According to the phase-shift hypoth- were obtained 7 times during the study to assess circa- esis for winter depression, morning light (which dian phase position. causes a circadian phase advance) should be more antidepressant than evening light (which causes a Results: Morning light phase-advanced the dim-light mela- delay). Although no studies have shown evening light tonin onset and was more antidepressant than evening light, to be more antidepressant than morning light, investi- which phase-delayed it. These findings were statistically gations have shown either no difference or morning significant for both crossover and parallel-group compari- light to be superior. The present study assesses these sons. Dim-light melatonin onsets were generally delayed light-exposure schedules in both crossover and in the patients compared with the controls. parallel-group comparisons. Conclusions: These results should help establish the im- Methods: Fifty-one patients and 49 matched controls portance of circadian (morning or evening) time of light were studied for 6 weeks. After a prebaseline assess- exposure in the treatment of winter depression. We rec- ment and a light/dark and sleep/wake adaptation base- ommend that bright-light exposure be scheduled imme- line week, subjects were exposed to bright light at ei- diately on awakening in the treatment of most patients ther6to8AM or7to9PM for 2 weeks. After a week of with seasonal affective disorder. withdrawal from light treatment, they were crossed over to the other light schedule. Dim-light melatonin onsets Arch Gen Psychiatry. 1998;55:890-896 HE DISCOVERY that human phase advance and in the evening causes melatonin production can a phase delay.8,9,11-18 Critical to the PSH is be suppressed by bright that bright-light exposure should be more light1 led to the initial test antidepressant when it is scheduled in the of bright light in treating morning than when it is scheduled in the winter depression.2,3 Subsequently, Rosen- evening, because most patients with SAD T 4 thal and coworkers coined the term sea- are expected to have phase-delayed circa- sonal affective disorder (SAD) for this dis- dian rhythms when depressed in the order, described its clinical features, and winter. However, some clinicians recom- conducted the first controlled investiga- mend exposure at other times of the day, tion in which comparatively dim light was particularly if morning is not conve- used as a placebo control. A form of re- nient.19 Therefore, for both practical and current depression marked by an annual theoretical reasons, it is important to know onset from midautumn to early winter, if circadian time of exposure affects light’s SAD may affect more than 10 million antidepressant efficacy. Americans.5 Its prevalence is greatest in fe- males, and it is often accompanied by See also pages 861, atypical symptoms of fatigue, hypersom- nia, carbohydrate craving, and weight gain. 863, 875, and 883 From the Sleep and Mood Among the first hypotheses for SAD6,7 Disorders Laboratory, is the phase-shift hypothesis (PSH),8-10 There are no investigations that dem- Department of Psychiatry, Oregon Health Sciences which postulates that bright light is anti- onstrate evening light (EL) to be more an- University, Portland. Dr Ahmed depressant in patients with SAD at least tidepressant than morning light (ML). is now with the Palo Alto in part by causing a circadian phase ad- Studies are divided between those that Health Care System, Palo Alto, vance. It is well established that bright- show ML to be superior and those that Calif. light exposure in the morning causes a show no difference (see the “Comment” ARCH GEN PSYCHIATRY/ VOL 55, OCT 1998 890 ©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/30/2021 SUBJECTS AND METHODS instructed to sit 45° to the light with their eyes at midfix- ture level and to gaze across it once or twice a minute, rather than stare directly at it. SUBJECTS A prebaseline assessment was followed by a baseline week, 2 weeks of a first in-home light treatment period (I-1 After approval by the institutional review board at Oregon and I-2), a withdrawal week, and 2 weeks of a second light Health Sciences University, Portland, the study was treatment period (II-1 and II-2). Subjects were assigned to conducted each January and February during 4 winters receive either ML first (6 to 8 AM; MLF group) or EL first (1989-1992) in Portland, (45°N latitude). Subjects were (7 to 9 PM; ELF group), balanced for age, sex, and pre- recruited through advertisements in area newspapers, tele- baseline SIGH-SAD rating. Before the study, subjects were vision interviews, and referrals from health professionals. asked whether they thought ML or EL would be more an- About 1900 patients were sent a Seasonal Pattern Assess- tidepressant, equally antidepressant, or make no differ- ment Questionnaire,20 a 21-item Beck Depression Inven- ence; they could also answer “don’t know.” During the study, tory,21 and a health and sleep screening questionnaire that subjects were asked to avoid unscheduled bright-light ex- inquired about medical and psychiatric history. More than posure (such as in the supermarket) from 6 PM to 8 AM and 80 potential subjects were interviewed. not to travel outside the time zone. Subjects were also asked To be admitted into the study, all patients had to (1) to maintain a 10 PM to 6 AM sleep schedule and to com- meet DSM-III-R criteria22 for moderate to severe major de- plete a written daily sleep and mood diary. For most sub- pressive disorder (without psychotic episodes) or bipolar jects, sleep diary data were also obtained for 2 weeks be- disorder (depressed or not otherwise specified) with a fore entry into the study. Morning wake-up times were winter-type seasonal pattern; (2) score 20 or higher on the verified each day by having subjects call the laboratory and Structured Interview Guide for the Hamilton Depression leave a voice-mail message. Rating Scale–Seasonal Affective Disorder Version (SIGH- SAD)23 with a Hamilton Depression scale24 score of 10 or MELATONIN more and an atypical score of 5 or more25; (3) report that a depression developed during the fall or winter and remit- Subjects were admitted to the Clinical Research Center at ted the following spring for at least the 2 preceding years; Oregon Health Sciences University for a prebaseline visit (4) be in good physical health; (5) not be suicidal; (6) not followed by 6 weekly visits for assessment of the DLMO (a have used psychotropic medications for the previous 4 weeks very reliable marker of circadian phase position26) and be- or other medications that interfere with endogenous mela- havioral (29-item SIGH-SAD) ratings by 2 independent blind tonin production; and (7) not have other psychiatric or raters (including, at times, N.L.C., M.L.B., and J.M.L.J.). medical illnesses. Control subjects did not have notable Subjects also completed a Beck Depression Inventory each medical or psychiatric problems (including seasonal fluc- week. Under dim light (,50 lux), blood samples (3-5 mL) tuations in mood, sleep, appetite, and energy) and were were drawn every 30 minutes from 6 to 11 PM (sometimes taking no medications that could interfere with mood or later). (Dim light is necessary in sighted people to avoid endogenous melatonin production. the acute suppressant effect of light on melatonin1; we cur- A total of 56 patients and 52 controls participated. rently recommend light intensities of ,30 lux for the Two patients did not complete the study, 3 were found to DLMO.) Samples were immediately centrifuged, and plasma have other psychiatric problems after beginning the study, was separated and frozen for later analysis. Melatonin was and 2 had immeasurable melatonin levels (which are assayed by means of a modification of the highly sensitive excluded from any dim-light melatonin onset [DLMO] and specific gas chromatographic–negative chemical ion- analyses). Three controls did not complete the study. ization mass spectrometric method.27 The DLMO was de- Controls and patients were matched by age and sex as fined as the first interpolated point above 10 pg/mL that closely as possible. Informed consent was obtained from continued to rise. all participants. STATISTICS LIGHT TREATMENT For continuous data, repeated-measures analyses of vari- The light fixtures contained two 40-W cool-white fluores- ance (ANOVAs) were performed, followed by group or cent tubes (Hughes Lighting Technologies, Lake Hopat- paired t tests, where appropriate. For categorical data, x2 cong, NJ). To determine the distance to obtain 2500 and Fisher exact tests were used. Unless otherwise stated, lux, intensities were measured for each light fixture using results described as statistically significant refer to P#.05 a light meter with a photometric sensor. Subjects were on a 2-tailed test. section). This may be similar to the pharmacological lit- treatment period) and crossover comparisons. We hy- erature, in which antidepressant medication has been shown pothesized that ML would be more antidepressant to be superior or equal, but rarely inferior, to placebo. than EL. Morning light has been found to be more antide- pressant primarily in crossover studies, whereas most RESULTS parallel-group studies have failed to demonstrate a sta- tistically significant benefit of ML. The following inves- Fifty-one patients with SAD (45 women) and 49 con- tigation studied a large number of patients, so as to have trols (40 women) completed the study.