ORIGINAL ARTICLE

Bright Light Treatment in Elderly Patients

With Nonseasonal Major Depressive Disorder
A Randomized Placebo-Controlled Trial
Ritsaert Lieverse, MD; Eus J. W. Van Someren, PhD; Marjan M. A. Nielen, PhD;
Bernard M. J. Uitdehaag, MD, PhD; Jan H. Smit, PhD; Witte J. G. Hoogendijk, MD, PhD

Context: Major depressive disorder (MDD) in elderly Main Outcome Measures: Mean improvement in
individuals is prevalent and debilitating. It is accompa- Hamilton Scale for Depression scores at T1 and T2 using
nied by circadian rhythm disturbances associated with parameters of sleep and cortisol and melatonin levels.
impaired functioning of the suprachiasmatic nucleus, the
biological clock of the brain. Circadian rhythm distur- Results: Intention-to-treat analysis showed Hamilton Scale
bances are common in the elderly. Suprachiasmatic for Depression scores to improve with BLT more than pla-
nucleus stimulation using bright light treatment (BLT) cebo from T0 to T1 (7%; 95% confidence interval, 4%-
may, therefore, improve mood, sleep, and hormonal 23%; P=.03) and from T0 to T2 (21%; 7%-31%; P=.001).
rhythms in elderly patients with MDD. At T1 relative to T0, get-up time after final awakening in
the BLT group advanced by 7% (P ⬍ .001), sleep effi-
Objective: To determine the efficacy of BLT in elderly
ciency increased by 2% (P=.01), and the steepness of the
patients with MDD. rise in evening melatonin levels increased by 81% (P=.03)
compared with the placebo group. At T2 relative to T0,
get-up time was still advanced by 3% (P=.001) and the 24-
Design: Double-blind, placebo-controlled randomized
hour urinary free cortisol level was 37% lower (P=.003)
clinical trial.
compared with the placebo group. The evening salivary cor-
tisol level had decreased by 34% in the BLT group com-
Setting: Home-based treatment in patients recruited from
pared with an increase of 7% in the placebo group (P=.02).
outpatient clinics and from case-finding using general
practitioners’ offices in the Amsterdam region. Conclusions: In elderly patients with MDD, BLT improved
mood, enhanced sleep efficiency, and increased the upslope
Participants: Eighty-nine outpatients 60 years or older melatonin level gradient. In addition, BLT produced con-
who had MDD underwent assessment at baseline (T0), tinuing improvement in mood and an attenuation of cor-
after 3 weeks of treatment (T1), and 3 weeks after the tisol hyperexcretion after discontinuation of treatment.
end of treatment (T2).
Trial Registration: clinicaltrials.gov Identifier
Intervention: Three weeks of 1-hour early-morning BLT NCT00332670
(pale blue, approximately 7500 lux) vs placebo (dim red
light, approximately 50 lux). Arch Gen Psychiatry. 2011;68(1):61-70

M
AJOR DEPRESSIVE DIS - tivity.13,14 Light induces specialized light-
order (MDD) is fre- sensitive retinal ganglion cells to release glu-
quently accompanied tamate in the SCN through a monosynap-
by symptoms sugges- tic pathway called the retinohypothalamic
tive of circadian dys- tract.15-17 Bright light treatment also targets
function,1 such as abnormal sleep-wake pat- depression-associated neurotransmitter sys-
terns,2 altered social rhythms,3 and diurnal tems (serotonin, noradrenalin, and dopa-
mood swings.4 These symptoms have, there- mine) and targets the same brain structures
fore, been related to impaired functioning as antidepressant drug treatments.18,19 In pri-
of the suprachiasmatic nucleus (SCN), the mates, subcortical projections of retinal neu-
circadian pacemaker of the brain.5-8 Activa- rons not only involve the SCN but also the
tion of the SCN has been hypothesized as serotonergic raphe nucleus.20,21 Elderly
one of the mechanisms of bright environ- people expose themselves less frequently to
mental light (bright light treatment [BLT]) bright environmental light.22,23 Moreover,
Author Affiliations are listed at on mood,9,10 sleep,11,12 circadian rhythms,11 with aging, photoreception declines.24 Con-
the end of this article. and hypothalamic-pituitary axis (HPA) ac- certedly, these age-related changes may re-

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
61

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 Therefore, we conducted a double-blind, placebo-
Table 1. Demographic Characteristics of the 89 Randomized controlled, randomized clinical trial that included assess-
Study Patients by Treatment Assignment ment of SCN function from cortisol profiles, rise in evening
melatonin levels, and actigraphic sleep estimates.
Placebo BLT
Group Group P
Characteristic (n=47) (n=42) Value a METHODS
Age, mean (SD), y 69.00 (6.6) 69.67 (8.5) .69
Sex, No. F/M (% female) 30/17 (64) 28/14 (67) .52 The present study was executed in accordance with the Hel-
Lives with partner, No. (%) 18 (38) 16 (38) .52 sinki Declaration.39 Approvals were obtained from the Dutch
Body height, mean (SD), m 1.69 (0.08) 1.68 (0.07) .55 authorities and the medical ethical committee (METIGG
Body weight, mean (SD), kg 76.37 (14.2) 73.88 (15.1) .45 [Medisch-ethische Toetsingscommissie Instellingen Geesteli-
BMI, mean (SD) 26.79 (4.6) 26.15 (4.6) .54 jke Gezondheidszorg], Utrecht). In particular, the medical ethi-
MMSE score, mean (SD) 28.47 (1.8) 27.60 (2.0) .04 cal committee consented to the blinding procedure and the way
CIRS comorbidity scores, information was provided to the patients.
mean (SD)
Total pathology 5.33 (2.8) 4.82 (2.9) .55
Illness severity composite 0.30 (0.2) 0.24 (0.2) .33 PARTICIPANTS
Comorbidity composite 0.96 (1.1) 0.83 (1.0) .59
SCID-IV/DSM-IV comorbid Based on the literature, a moderate response was expected.40-42 With
diagnoses (DSM-IV code) the use of conventional values for ␣ (.05) and ␤ (.80) for 2-tailed
Panic disorder with tests with equal groups, the sample size was determined to be 63
agoraphobia (300.21) patients per arm,43 resulting in a total number of 126 patients. In-
Severe 1 0 ⬎.99
clusion started on January 22, 2003, and lasted until August 22,
Moderate 0 1 ⬎.99
2007 (4.5 years). By then, 89 patients were included (for depres-
Light 2 1 ⬎.99
sion characteristics, see eTable 1 http://www.genpsychiatry
Hypochondria (300.7) 0 1 ⬎.99
Social phobia (300.23) 0 1 ⬎.99
.com). Taking into account the conservative power analysis and
Alcohol abuse (305.00) the limiting resources and perspectives for subsequent inclu-
Sustained full remission 2 3 ⬎.99 sion rates, it was decided not to include more patients.
Early partial remission 0 1 ⬎.99 We recruited study participants from outpatient clinics, ad-
Cannabis abuse (305.20) 0 1 ⬎.99 vertisements, and referrals by general practitioners. Candi-
dates were 60 years or older and first selected using the 15-
Abbreviations: BLT, bright light treatment; BMI, body mass index item version of the Geriatric Depression Scale. Individuals with
(calculated as weight in kilograms divided by height in meters squared); Geriatric Depression Scale scores of 5 or more were screened
CIRS, Cumulative Illness Rating Scale55; MMSE, Mini-Mental State by interview (n=444) to establish whether they met the eligi-
Examination; SCID-IV, Structured Clinical Interview for DSM-IV Axis I bility criteria. Exclusions were categorized as psychiatric
Disorders.44,56 (n=154), neurological (n=22), ophthalmological (n=17), re-
a Calculated as comparisons of BLT and placebo, using 2-tailed t tests
(continuous variables) or ␹2 tests (discrete variables). Statistically significant search incompatibility (n=101), and miscellaneous (n=9)44,54,55
test values are in bold type. (Table 1). In addition, 52 individuals refused to participate.

DIAGNOSIS AND QUANTIFICATION

sult in insufficient stimulation of the SCN,22,23 thought to OF SEVERITY
be involved in the attenuated neuronal activity in the SCN
at advanced age.25 Bright light treatment could therefore be Depression was diagnosed using the Structured Clinical Inter-
hypothesized to be particularly suitable in the management view for DSM-IV Axis I Disorders.56 Severity was rated with the
of MDD in elderly patients, which is important because of Structured Interview Guide for the Hamilton Scale for Depres-
the less favorable adverse-effect profile of antidepressants sion (HAM-D)–Seasonal Affective Disorder Version,48,49 a struc-
in this population. tured interview yielding total score, the original HAM-D score,47
the 8-item Atypical Symptom Scale score, and the HAM-D6,
The beneficial effect of BLT in seasonal affective disor- the 6-item core version score.50,57 Furthermore, the Montgomery-
der is well accepted,26 with early onset of action27 and mild Åsberg Depression Rating Scale46 was used to allow compari-
adverse-effect profiles.28 Results of controlled BLT trials in son of our results with other randomized controlled trials in
nonseasonalMDDarepromisingbutinconclusive,especially depression in elderly patients. Interviews were performed by a
with respect to efficacy in elderly patients with MDD.26,29-35 trained physician (R.L.) and qualified research psychologists
Reviews emphasize the need for further study because of the (including M.M.A.N.), all blind to assignments (Table 2).
greatdiversityofstudydesignsandtherelativelysmallsample
sizes.30,36 We showed that bright light attenuated the devel- STUDY DESIGN
opment of depressive symptoms in elderly residents of group
care facilities.11 To our knowledge, double-blind, placebo- We used a randomized, double-blind, placebo-controlled de-
controlled,randomizedclinicaltrialsofsufficientsamplesize sign to compare the antidepressive effects of BLT and placebo.
to evaluate the efficacy of BLT in elderly patients diagnosed Permuted block randomization in subsets of 10 was per-
formed, with separate randomizations for the strata of patients
as having MDD have not been performed, although some
who used and did not use antidepressants. The 2 randomization
studies suggested BLT might have favorable effects.34,35,37,38 lists, prepared by an independent researcher (B.M.J.U.) not in-
Our hypotheses were 2-fold. First, we expected BLT to volved in the recruitment and using a computer-generated table,
lower depressive symptoms. Second, we expected this to were transferred to a sequence of sealed opaque envelopes.
be mediated by improved circadian functioning, as indi- Study patients were informed that the primary goal of the
rectly indicated by enhanced sleep and hormone rhythms. study was to investigate spectrum-dependent efficacy differ-

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
62

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 ences between blue and red. Investigators were blinded to the ance (ANOVA) was applied to completers (20 patients in the
condition because the lamps were delivered at the patients’ BLT group and 20 in the placebo group) with the T0 cortisol
homes by protocol-blinded instructors, who were also in- level as the covariate. To evaluate whether MDD was associ-
formed that the study aimed for spectrum-dependent efficacy ated with HPA alterations, age- and sex-matched nonde-
differences. Patients were asked not to discuss any details of pressed control patients were recruited from general practi-
their condition with the interviewers. In 2 cases, patients did tioners’ offices. We excluded controls with Geriatric Depression
reveal their assignment, after which the interviewer was re- Scale45 scores larger than 0, a lifetime history of psychiatric dis-
placed. Before the light box was installed, patients completed orders, any somatic condition that could interfere with HPA
a 4-item expectations questionnaire (eTable 2 ). functioning, or any required medications other than sporadic
use of aspirin. Valid urine samples were obtained from 8 men
STUDY INTERVENTION and 14 women with a mean (SD) age of 68.9 (6.4) years.

Patients were randomly assigned to receive bright pale blue or Saliva Cortisol Levels. At T0, T1, and T2, we collected saliva
dim red light treatment therapy at home using 2 light boxes (Phil- samples using cotton dental rolls (Salivette; Sartstedt Ltd, Num-
ips Bright Light Energy HF 3304; Koninklijke Philips Electron- brecht, Germany), including 4 sequential single samples at 30-
ics NV, Eindhoven, the Netherlands). Concealed within the light minute intervals starting 30 minutes after final awakening and
boxes, a single-layer filter was wrapped around the fluorescent 4 sequential samples at hourly intervals starting 4 hours be-
tubes: a mist-blue filter (Model 061; Lee Filters, Andover, En- fore the predicted bedtime (supplementary text; available at http:
gland) with high-throughput pale blue (7500 lux) for the active //www.ggzingeest.nl/saliva-sampling). The samples were col-
condition and a blood-red filter (Model 789; Lee Filters) with low- lected the following day to be delivered to the laboratory, where
throughput red (50 lux) for the placebo condition (eFigure 1). they were centrifuged and stored at −85°C. All samples were
Dim red light can be considered to be biologically inactive58 analyzed in a single batch using a cortisol assay on an immu-
(supplementary Appendix A). noanalyzer system (Roche Cobas assay on an Elecsys system;
Given the proposed interaction between exposure inten- Roche Diagnostics, Mannheim, Germany). The detection limit
sity and duration for the efficacy of BLT,9 we chose an expo- was 0.07 µg/dL (to convert to nanomoles per liter, multiply by
sure of 60 minutes in the early morning at about 7500 lux. For 27.588), and the intra-assay and interassay variability coeffi-
BLT of nonseasonal depression in elderly patients, there is no cients were less than 10%. For determination of the diurnal time
consensus with respect to optimal timing, dosage, and treat- course of saliva cortisol levels, only days with at least 7 of 8
ment duration. We chose 3 weeks of daily light exposure valid samples were included in analyses. A skewed cosine func-
(Figure 1) because most studies thus far used short-term treat- tion72 was fitted to each day using SPSS statistical software, ver-
ment of up to 1 week37,41,42,59-69 and because the Cochrane re- sion 16.0.2 (SPSS, Inc, Chicago, Illinois), providing the most
view of studies of BLT in nonseasonal affective disorder con- parsimonious rhythmic diurnal curve description that allows
cluded that BLT may be effective in as little as 1 week.31 for skewness, an undisputed property of the cortisol curve. Areas
under the curves for the morning and evening (ie, 9 AM to
OUTCOME MEASURES 1 PM and 5 to 9 PM) were calculated for subsequent analyses.
Assessments were performed at the following 3 time points Saliva Melatonin Levels. At T0, T1, and T2, 4 sequential saliva
(Figure 1): just before the start of light treatment (baseline [T0]), samples were collected using the cotton dental rolls (Salivette)
immediately on completion of the 3-week treatment interval at hourly intervals starting from 4 hours before predicted bed-
(T1), and 3 weeks after the end of the treatment (T2). time under dim light conditions (supplementary text and Ap-
The primary outcome was determined to be the change in pendix B). The samples were collected the following day to be
HAM-D score at T1 relative to T0. Secondary efficacy out- delivered to the laboratory to be centrifuged and stored at −85°C.
come measures were (1) change in HAM-D score at T2 rela- Concentrations were determined using an assay with a limit
tive to T0 to investigate whether immediate responses would of sensitivity of 0.2 ng/L (to convert to picomoles per liter, mul-
last after treatment discontinuation and (2) the dichotomized tiply by 4.305) (Bühlmann Laboratories AG, Schönenbuch, Swit-
treatment response for T1 relative to T0 and T2 relative to T0 zerland) and intra-assay and interassay coefficients of 2.6% and
(with responders vs nonresponders defined according to whether 20.1%. For determination of a rise in melatonin levels, only days
the HAM-D score decreased by at least 50%). with at least 3 of 4 valid samples were included in the analyses.
Because melatonin levels were so low that commonly ap-
Endocrine Outcome Measures plied methods were not applicable, we could obtain a measure
of the steepness of the evening rise only, which may have bio-
Urinary Cortisol Levels. Urinary free cortisol (UFC) levels dur- logical relevance and which has been proposed before as a pa-
ing a 24-hour period provide a noninvasive valid estimation of rameter of use.72 Therefore, we used a mixed-effect linear regres-
overall daily cortisol production.70 Collections were per- sion model to estimate treatment effects on the slope (steepness
formed at home at T0, T1, and T2. Urine was collected in 3-L of melatonin level rise) and intercept (timing of the melatonin
polyethylene bottles starting after the first voided urine after level rise) of the evening rise (supplementary Appendix B).
awakening and included the first voided urine on the follow-
ing day. The UFC level was determined by radioimmunoassay
using a commercially available antibody kit (Coat-A-Count; Di- Actigraphic Estimates of Sleep and Light Exposure
agnostic Product Corporation, Siemens, Los Angeles, Califor-
nia). Analysis procedures and limits of detection reported for Actigraphy, the continuous assessment of activity with a watch-
assays performed at the VU University Medical Center Labo- sized nondominant wrist-worn recorder (Actiwatch-L; Cam-
ratory are published by the manufacturer and available on re- bridge Neurotechnology, Cambridge, England), is a validated
quest. Completeness of collection was ascertained by inter- technique to obtain estimates of sleep.11,73,74 Patients wore ac-
views documenting urine losses. Only complete collections, with tigraphs throughout their participation and were instructed not
creatinine within the normal range of 0.06 to 1.20 mg/dL per to remove them when taking a bath or shower. Patients kept a
24 hours (to convert to micromoles per liter, multiply by 88.4) diary of bedtimes and get-up times after final awakening. The
were included in analysis.71 Repeated-measures analysis of vari- sleep analysis software (Sleepwatch; Cambridge Neurotech-

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
63

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 nology) was used to obtain estimates of sleep parameters, in- An adverse event was recorded only if it increased relative to
cluding total sleep time, sleep efficiency (ie, the percentage of baseline and the previous rating. Group differences in frequen-
actual sleep between sleep onset and final awakening), and sleep cies were compared using ␹2 statistics.
onset latency (ie, the time between lights out and sleep onset).
A light sensor integrated in the actigraphs was used to evalu- STATISTICAL ANALYSES
ate whether treatment adherence was supported by increased
intensity recording during the time intervals of BLT and to evalu- Baseline characteristics were compared using 2-sided t tests for
ate compliance with dim-light requirements during saliva sam- continuous data and ␹2 statistics and 2-tailed Fisher exact tests
pling for characterization of the evening rise in melatonin lev- for categorical data with the use of SPSS 16.0.2 software (Table 1
els (supplementary Appendix A). and eTable 1).

Adverse Events
444 Assessed for eligibility
At baseline and at the end of every week during treatment, pa-
tients were systematically interviewed about 28 possible ad-
355 Excluded
verse effects by blinded raters. Each item was rated on a 4-point 52 Refused to participate
scale (0 indicates absent; 1, mild; 2, moderate; and 3, severe). 303 Not meeting inclusion criteria
154 Psychiatric reasons
22 Neurological reasons
17 Ophthalmological reasons
101 Research incompatibility
1-3 wk 3 wk 3 wk reasons
9 Miscellaneous
Actigraphy

Dim red light treatment

89 Randomized

Bright blue light treatment

42 Allocated to bright blue light 47 Allocated to dim red light

GDS-15 GDS-15 ≥ 5 T0 T1 T2
• Actigraph • Inclusion or exclusion • Psychometry • Psychometry
• Instructions • MDD on SCID-IV • 24-h Urinary • 24-h Urinary 2 Withdrawn during intervention (T1) 3 Withdrawn during intervention (T1)
• General practitioner cortisol level cortisol level
• Saliva samples • Saliva samples 1 Medication switch 1 Medication switch
Randomization 8 times 8 times 1 Subjective worsening depression 1 Subjective worsening depression
4 Withdrawn during follow-up (T2) 1 Not willing to continue study
• Psychometry
• Expectancy 1 Not willing to continue study 6 Withdrawn during follow-up (T2)
• 24-h Urinary cortisol level after intervention 3 Not willing to continue study
• Saliva samples 8 times 1 Hospital admission because of after intervention
(cortisol and melatonin levels)
somatic illness 3 Went on vacation after
Instruction of 2 Went on vacation after intervention
• Light box intervention

Figure 1. Diagram of the study protocol. Assessments of psychometry and 40 Underwent analysis 44 Underwent analysis
hypothalamic-pituitary axis function and parameters of sleep and circadian
rhythmicity were conducted before the start of 3 weeks of light treatment
(T0), after 3 weeks of treatment (T1), and 3 weeks after discontinuation of Figure 2. Flow of study patients. Every randomized patient started treatment.
treatment (T2). Wrist actigraphy was continuously measured. GDS-15 Five patients discontinued the intervention and refused follow-up. Analyses
indicates 15-item version of the Geriatric Depression Scale; MDD, major fulfill intention-to-treat characteristics because none of the patients assigned
depressive disorder; and SCID-IV, Structured Clinical Interview for DSM-IV to a condition switched to another condition and because analyses involved
Axis I Disorders. all available observations of all patients.

Table 2. Outcomes in Depression Ratings a

Placebo Group, Mean (SD) BLT Group, Mean (SD) Change From T0 to T1 b Change From T0 to T2 b

Mean Test Statistic Cohen Mean Test Statistic Cohen

Outcome T0 T1 T2 T0 T1 T2 (95% CI) b (P Value) c dd (95% CI) b (P Value) c dd
HAM-D e 16.2 (4.6) 10.4 (6.3) 10.8 (6.5) 18.6 (5.7) 10.1 (6.1) 8.6 (6.5) 2.6 (0.3-4.9) F1,81 =3.94 (.03) f 0.50 4.5 (2.4-6.6) F1,81 =11.39 (.001) f 0.93
(n=84)
BCF g 16.0 (4.7) 10.6 (6.3) 10.9 (6.4) 18.4 (5.6) 10.4 (6.1) 8.9 (6.5) 2.6 (0.3-4.8) F1,86 =3.18 (.04) f 0.48 4.4 (2.3-6.5) F1,86 =8.846 (.004) f 0.80
(n=89)
CA g 15.7 (4.3) 9.9 (6.1) 10.3 (6.3) 18.5 (5.6) 9.9 (6.1) 8.2 (6.5) 2.7 (0.2-5.2) F1,71 =3.14 (.04) f 0.50 4.9 (2.6-7.1) F1,71 =11.39 (⬍.001) f 1.01
(n=74)

Abbreviations: BCF, baseline carried forward analysis; BLT, bright light treatment; CA, completers analysis; CI, confidence interval; HAM-D, Hamilton Scale for
Depression47; T0, baseline; T1, after 3 weeks of treatment; T2, 3 weeks after discontinuation of treatment.
a Outcome descriptions are given in the “Outcome Measures” subsection of the “Methods” section.
b Indicates differences between BLT and placebo in the change from T0 to each patient’s own end point for the change in depression rating.
c Calculated as part of the repeated-measures analysis of covariance (ANCOVA), using T0 depression rating and Mini-Mental State Examination scores as
covariates. Statistically significant test values are depicted in bold type.
d Computed as the difference between the means, M − M , divided by the pooled standard deviation, sigma (␴
1 2 pooled) of both groups.
e The intention-to-treat analysis used the last observation carried forward.
f With repeated-measures ANCOVA, using the T0 rating as a covariate was significant.
g Performed as a sensitivity analysis.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
64

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 Treatment effect analyses fulfilled intention-to-treat criteria
because none of the patients assigned to one condition switched †
12 BLT
to another, and analyses involved all observations of all patients Placebo
until study end or withdrawal. The primary efficacy outcome analy- ∗

Mean Change From T0 in HAM-D Score

10
sis consisted of repeated-measures ANOVA with baseline HAM-D
scores as covariates. Ancillary analyses consisted of analysis of
covariance (ANCOVA) on HAM-D scores from T0 to T2 scores. 8

To analyze the interaction effect of antidepressants, it was added

to the repeated-measures model. Subgroup analyses of the pos- 6
sible effects of antidepressants, age, sex, melancholy, atypical fea-
tures, seasonality, recurrent course, treatment resistance, late on- 4
set, and duration of depression were preplanned.
Numbers needed for treatment were computed according 2
to the methods of Sacket et al,75 with 95% confidence intervals
(CIs) computed using the method of Altman.76 0
For dropouts after the T1 assessment, the principle of last ob- T1 T2
servation carried forward was used for depression scales. As sec-
ondary sensitivity analyses, we performed a baseline (T0) car- Figure 3. Changes in the Hamilton Scale for Depression (HAM-D) from
ried forward analysis and a T2-completers analysis (Figure 2). baseline (T0) in groups receiving bright light treatment (BLT) and placebo for
We used mixed-effect regression analysis (MLwiN soft- nonseasonal major depressive disorder in elderly individuals. Bars indicate
ware; Institute of Education, London, England) to evaluate treat- standard deviations. T1 indicates after 3 weeks of treatment; T2, 3 weeks
ment effects on saliva cortisol and melatonin levels and diary after discontinuation of treatment. *P ⬍ .05. †P= .001.
and actigraphic sleep estimates to account for the variable num-
ber of valid days within patients, without having to discard pa- cebo groups (all P⬎.05, ANOVA). Responders in the BLT
tients because of partially missing data. group had more pessimistic expectations concerning im-
Based on the literature finding that dim red light treatment provement without treatment than did placebo respond-
never had a more favorable outcome than BLT on depression
ers (BLT group, 4.55 [1.14]; placebo group, 3.56 [1.01];
ratings,31 we justified 1-sided testing on the primary outcome
of depression ratings at T1. All other significance levels for ef- F1,29 =5.08; P=.03). Mean expectations in nonresponders
fects (ie, at T2) were set at P⬍.05 with 2-sided testing. Means in the BLT and placebo groups did not differ (all P⬎.05,
and 95% CIs are provided. Secondary analyses were not ad- ANOVA). Expectations did not predict treatment re-
justed for multiple comparisons and should therefore be re- sponse (r=0.03; P=.81) (eTable 2).
garded as descriptive and exploratory. Where not otherwise in-
dicated, data are expressed as mean (SD). TREATMENT ADHERENCE

RESULTS Adherence to treatment was supported by the fact that only

BLT-assigned patients showed elevated light exposure ex-
clusively during the treatment intervals (supplementary Ap-
DEMOGRAPHIC AND pendix A).
CLINICAL CHARACTERISTICS
TREATMENT EFFECT ON
We included and randomized 89 patients, with 42 allo- DEPRESSION RATINGS
cated to the BLT condition and 47 to the placebo con-
dition (Figure 2). There were no hospitalizations or sui- The intention-to-treat analysis showed significantly more
cides or other deaths. T0 to T1 improvement in HAM-D scores in patients in
Randomization was balanced with respect to demo- the BLT group (43%; 8.5 [95% CI, 6.8-10.3] points) than
graphic and comorbidity characteristics and psychiatric co- in the placebo group (36%; 5.8 [4.0-7.6] points), the dif-
morbid diagnoses (Table 1). Groups were not balanced with ference being 7% (4%-23%; F1,81 =3.94; 1-sided P=.03, with
regard to Mini-Mental State Examination score (mean pla- HAM-D and Mini-Mental State Examination scores at T0
cebo group score, 28.5 [1.8]; mean BLT group score, 27.6 as covariates). Ancillary analyses of treatment effects af-
[2.0]; P=.04) or the pretreatment HAM-D score (mean pla- ter discontinuation at T2 likewise showed significantly
cebo group score, 16.0 [4.7]; mean BLT group score, 18.4 more T0 to T2 improvement in HAM-D scores in the BLT
[5.6]; P=.03). Baseline values were therefore used as co- group (54%; 10.0 [95% CI, 8.6-12.0] points) than in the
variates in all effect analyses. The number of patients who placebo group (33%; 5.4 [3.9-6.9] points), the differ-
received psychotherapy in the past was smaller in the pla- ence being 21% (7%-31%; repeated-measures ANCOVA,
cebo group than in the BLT group (21 [45%] vs 31 [74%]; F1,81 =11.39; P=.001, with HAM-D and Mini-Mental State
␹2 =7.0; P=.007). Three patients in the placebo group dis- Examination scores at T0 as covariates) (Table 2 and
continued before T1 and 6 after T1. Two patients in the Figure 3).
BLT group discontinued before T1 and 4 after T1 (Figure 2). At T1, 20 patients in the BLT group (50%) were re-
sponders vs 18 (41%) in the placebo group (␹12 = 0.70;
EXPECTANCY P = .20) (eTable 3 and Figure 4). The difference be-
came significant at T2, with 23 responders in the BLT
None of the 4 expectancy scores differed significantly be- group (58%) vs 15 in the placebo group (34%) (␹12 =3.76;
tween the treatment groups (all P⬎.05, ANOVA) or be- P= .05). The number needed to treat for HAM-D score
tween responders and nonresponders in the BLT or pla- improvement at T2 was 5 (95% CI, 1-151) (eTable 3).

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
65

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 A B
30 30
Placebo
BLT

20 20
T1 HAM-D Score

T2 HAM-D Score
10 10

0 0

0 10 20 30 0 10 20 30
T0 HAM-D Score T0 HAM-D Score

Figure 4. Scatterplots of individual patients’ Hamilton Scale for Depression (HAM-D) scores at baseline (T0), after 3 weeks of treatment (T1) (A), and 3 weeks
after discontinuation of treatment (T2) (B). Treatment consisted of bright light treatment (BLT) or placebo. Points that fall below the solid diagonal represent
patients who improved. Points that fall below the dashed diagonal in the gray shaded area represent patients whose scores were reduced by 50% or more relative
to baseline.

As sensitivity analyses, the baseline carried forward and cance (ANCOVA, F1,38 =3.663; P=.06). Significance was
completers analyses showed results comparable to those reached by T2 when the 24-hour UFC level had de-
of the intention-to-treat analysis (Table 2 and eTable 3). creased by 17% (−0.98 [95% CI, −1.73 to 0.24] µg) in
Analyses on other depression ratings produced similar re- the BLT group and had increased by 20% (0.98 [0.16-
sults, with some significant and others as trends only 1.81] µg) in the placebo group, resulting in a difference
(Table 3, eTable 4, and Figure 5). of 37% (ANCOVA, F2,37 =6.78; P =.003) (Figure 6). At
T2, 24-hour UFC excretion of patients undergoing BLT
EFFECT MODIFICATION BY ANTIDEPRESSANT no longer differed from that of the healthy controls
USE AND DEPRESSION SUBTYPE (P=.47). Thus, in contrast to the placebo group, the mean
24-hour UFC excretion in the BLT group was signifi-
Fourteen patients in the BLT group (33%) and 18 in the cantly lowered (supplementary Appendix C and eFig-
placebo group (38%) used antidepressants. Analyses re- ure 3). To investigate whether the increased 24-hour UFC
vealed no effect of antidepressants on the HAM-D scores excretion in placebo-treated patients could be ex-
(F1,71 =1.46; P=.24) or interaction of antidepressants with plained by nonresponse, placebo nonresponders were
treatment effect at T2 (F1,71 =0.001; P=.98). Likewise, there compared with placebo responders, which showed that
was no significant effect on HAM-D score or the inter- nonresponders had higher 24-hour UFC excretion than
action of treatment by patient characteristics, including responders (5.53 [3.34] vs 3.94 [1.29] µg/dL) but with-
age (F1,71 =0.41; P = .67), sex (F1,71 = 0.50; P = .61), melan- out reaching statistical significance (P=.07).
choly (F2,138 =0.23; P=.79), atypical features (F2,138 =0.59;
P=.55), seasonality (Global Seasonality Score; F1,71 =0.85; SALIVA CORTISOL LEVELS
P=.43), recurrent course (F1,71 = 1.13; P = .33), treatment
resistance (F2,138 = 1.68; P = .18), late onset (F2,138 =1.25; Seven patients (8%) refused saliva sampling. In sum, 1537
P=.29), or short duration (F2,138 = 0.02; P = .10) at T2. samples from 177 series were used from 5:30 AM until
3:15 AM. The skewed cosine model showed a goodness
24-HOUR URINARY CORTISOL EXCRETION of fit of R2 =0.79 (SD, 0.10). During the course from T0
to T2, the area under the curve during the evening (5-9
Nine patients (10%) refused to collect urine, and 3 oth- PM) showed a stronger decrease with BLT than with pla-
ers had incontinence. Seventy-two urine collections at cebo, reaching significance for the contrast between T2
T0 and 40 at both T1 and T2 (20 in each group) were and T0 (BLT, 34% decrease from T0 at 0.10 [95% CI, 0.07-
considered valid. Mean T0 24-hour UFC excretion was 0.12] µg/dL per minute to T2 at 0.05 [0.04-0.09] µg/dL
5.65 (3.73) µg, which was significantly higher than the per minute; placebo, 7% increase from T0 at 0.08 [0.05-
mean 24-hour UFC excretion of controls (4.31 [2.07] µg; 0.11] µg/dL per minute to T2 at 0.10 [0.04-0.15] µg/dL
P=.01) (supplementary Appendix C). per minute; P =.02). The morning area under the curve
From T0 to T1, 24-hour UFC excretion decreased by showed a similar decrease that was stronger during and
7.3% (−0.36 [95% CI, −1.76 to 1.04] µg) in the BLT group after BLT than placebo, although the difference did not
and increased by 32.3% (1.49 [0.36-2.61] µg) in the pla- reach significance (eFigure 4). The findings indicate that
cebo group, a difference that did not yet reach signifi- BLT accelerated the diurnal decline in saliva cortisol level.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
66

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 Table 3. Outcomes in Supplementary Depression Ratings a

Placebo Group, Mean (SD) BLT Group, Mean (SD) Change From T0 to T1 b Change From T0 to T2 b

Mean Test Statistic Cohen Mean Test Statistic Cohen

Outcome T0 T1 T2 T0 T1 T2 (95% CI) b (P Value c) dd (95% CI) b (P Value c) dd
HAM-D6 8.9 (2.4) 5.6 (3.2) 5.4 (3.8) 9.3 (3.0) 4.6 (3.3) 4.0 (3.0) 1.3 (−0.1 to 2.7) F1,81 =5.14 (.01) e 0.41 1.8 (0.4 to 3.1) F1,81 =8.78 (.004) e 0.58
(n=84)
ATYP-8 6.4 (4.3) 4.2 (3.8) 4.4 (3.2) 7.5 (4.7) 3.8 (2.9) 3.2 (3.9) 1.45 (−0.4 to 3.4) F1,81 =2.25 (.07) 0.33 2.1 (0.2 to 4.1) F1,81 =7.02 (.01) e 0.52
(n=84)
SIGH-SAD 24.4 (8.4) 16.1 (9.2) 16.3 (10.2) 28.1 (9.0) 15.3 (8.9) 12.7 (9.3) 4.5 (0.9 to 8.1) F1,81 =4.84 (.02) e 0.55 7.7 (4.1 to 11.2) F1,81 =13.13 (⬍.001) e 0.95
(n=84)
MADRS 25.2 (6.8) 16.9 (9.8) 16.1 (9.2) 24.7 (6.5) 14.4 (0.6) 12.7 (10.6) 2.1 (−1.4 to 5.5) F1,81 =3.04 (.04) 0.26 3.3 (−0.5 to 6.9) F1,81 =4.32 (.04) 0.41
(n=84)

Abbreviations: ATYP-8, Atypical Symptom Scale; BLT, bright light treatment; CI, confidence interval; HAM-D6, Hamilton Scale for Depression 6-item core version
(consisting of depressed mood, self-depreciation and guilt feelings, work and interests, psychomotor retardation, psychic anxiety, and general somatic)50; MADRS,
Montgomery-Åsberg Depression Rating Scale46; SIGH-SAD, Structured Interview Guide for the Hamilton Scale for Depression–Seasonal Affective Disorder Version48,49;
T0, baseline; T1, after 3 weeks of treatment; T2, 3 weeks after discontinuation of treatment.
a Outcome descriptions are given in the “Outcome Measures” subsection of the “Methods” section.
b Indicates differences between BLT and placebo in the change from T0 to each patient’s own end point for the change in depression rating.
c Calculated as part of the repeated-measures analysis of covariance (ANCOVA), using T0 depression rating and Mini-Mental State Examination scores as covariates.
Statistically significant test values are depicted in bold type.
d Computed as the difference between the means, M − M , divided by the pooled standard deviation, sigma (␴
1 2 pooled) of both groups.
e With repeated-measures ANCOVA, using the T0 rating as a covariate was significant.

SALIVA MELATONIN LEVEL T1, total sleep time decreased in the BLT group from 6
hours 52 minutes (95% CI, 6 hours 31 minutes to 7 hours
Seven hundred fifty-six samples were considered valid. 14 minutes) to 6 hours 37 minutes (6 hours 17 minutes
At T1 relative to T0, the steepness of the melatonin rise to 6 hours 57 minutes), which was a significantly stron-
increased by 109% in the bright blue light condition (from ger decrease (P=.03) than occurred in the placebo group
0.48 [95% CI, 0.27-0.69] to 1.00 [0.50-1.49] ng/L/h), (from 6 hours 42 minutes [6 hours 23 minutes to 7 hours
whereas it decreased by 11% in the dim red light condi- 1 minute] to 6 hours 22 minutes [6 hours to 6 hours 45
tion (from 0.32 [CI, 0.17-0.47] to 0.28 [0.09-0.47] ng/ minutes]). No significant differences remained at T2
L/h). This differential change, being 81%, was signifi- (P=.47). From T0 to T1, sleep efficiency increased in the
cant (P=.03). A similar differential change between T0 BLT group from 76.8% (95% CI, 74.1%-79.5%) to 77.9%
and T2 did not reach significance. No significant changes (75.5%-80.4%), which was a significantly stronger in-
in regression intercept (ie, onset phase) were found crease (2%, P =.01) than the change that occurred in the
(supplementary Appendix B). The findings indicate that placebo group (from 75.9% [73.5%-78.4%] to 75.6%
BLT enhanced the evening rise in saliva melatonin level. [73.2%-78.0%]). No significant differences remained at
T2 (P=.61). No differential changes occurred in sleep on-
SLEEP set latency from T0 to T1 (P = .53) or from T0 to T2
(P =.70). The findings indicate that BLT decreases total
At baseline, there were no statistically significant group sleep duration by advancing get-up time after final awak-
differences with respect to self-reported habitual bed- ening and increases sleep efficiency.
time (mean, 11:21 PM [1 hour 12 minutes]) or get-up time
after final awakening (mean, 8:19 AM [58 minutes]). No ADVERSE EFFECTS
significant group changes over time or treatment effects
were found for habitual bedtime. Between T0 and T1, Bright light treatment and placebo were well tolerated. Their
get-up time advanced in the BLT group from 8:07 (95% adverse effect profiles did not differ (eTable 5). In the pla-
CI, 7:47-8:26) AM to 7: 34 (7:19-7:50) AM, which was a cebo group, more patients reported the emergence or in-
significantly stronger advance (7%, P ⬍ .001) than oc- crease in daytime sleepiness (36% vs 24%; ␹2 =3.95; P=.05)
curred in the placebo group (from 8:32 [8:11-8:54] AM and fatigue (34% vs 19%; ␹2 =5.11; P=.02).
to 8:04 [7:47-8:22] AM). At T2 relative to T0, get-up times
after final awakening in the BLT group (T2, 7:49 [95%
CI, 7:25-8:12] AM) were still significantly (3%, P=.001) COMMENT
more advanced than in the placebo group (T2, 8:30 [8:07-
8:54] AM). No significant group changes over time or treat- This is, to our knowledge, the first double-blind, placebo-
ment effects were found for time in bed. controlled randomized trial with a sufficient sample size
Valid actigraphy recordings were available on aver- to evaluate the effects of BLT on mood in elderly pa-
age for 217 (113) hours before T0 as baseline assess- tients with a DSM-IV diagnosis of nonseasonal MDD. The
ment, for 414 (108) hours from T0 to T1, and for 287 design appeared successful with respect to treatment ad-
(215) hours from T1 to T2. At baseline, there were no herence and balanced expectations.
statistically significant group differences with respect to Directly after 3 weeks of treatment (T1), BLT im-
actigraphic estimates of total sleep time (P = .48), sleep proved depressive symptoms better than placebo (43%
efficiency (P=.63), or sleep latency (P=.37). From T0 to vs 36%). Three weeks after treatment withdrawal (T2),

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
67

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 A 20.0 100 80
BLT BLT
17.5 90 Placebo Placebo
60
80
15.0 ∗

Mean Change From T0 in 24-h

HAM-D Score 70

UFC Levels, nmoL/24 h

40
12.5
60
10.0 50 20

7.5 40

30 0

– 20
B 12 100

90 – 40
10
HAM-D6 Core Score

80 T1 T2
8 70

60 Figure 6. Mean change from baseline (T0) in patients receiving bright light
6
treatment (BLT) and placebo by effects on 24-hour urinary free cortisol
50 (UFC) levels. Data are depicted as mean change from baseline UFC levels;
4 error bars show the 95% confidence intervals. T1 indicates after 3 weeks of
40
treatment; T2, 3 weeks after discontinuation of treatment. *P = .003.
2 30

symptoms had continued to improve in the BLT group

C 10 100
but not in the placebo group (54% vs 33%). Bright light
treatment resulted in a good responder rate (ie, ⱖ50%
Atypical Syptom Scale Score

8
80 symptom reductions) of 50% vs 41% in the placebo con-
dition at T1 and of 58% vs 36% at T2 (eTable 3 and eTable
6
60
4). These effect sizes appear comparable to those re-
ported for antidepressants (number needed to treat, 5),
4 with the noticeable difference that no adverse effect could
40 be demonstrated for BLT (eTable 5). Ancillary analyses
2 on other measures of depression severity showed com-
parable results (Table 3, eTable 4, and Figure 5).
D 35 100 In contrast to the continuing improvement after dis-
90 continuation of treatment in the present study, Martiny
30
80 et al78 found a lack of sustained effect in their study. Four
SIGH-SAD Score

25 70 weeks after their treatment period of 5 weeks, the BLT

60
and placebo groups no longer differed regarding remis-
20
50
sion rates. Martiny et al hypothesized that BLT acceler-
15 ated remission of symptoms rather than having an aug-
40
menting effect. Whereas Martiny et al supplemented BLT
10 30
with pharmacological treatment, with increasing dos-
ages after the BLT period, our study did not offer a sec-
E 30 100 ondary treatment after the BLT period. We therefore con-
90
clude that our BLT protocol induced the recovery process
25 that lasted beyond discontinuation of treatment.
80
Of interest is the finding that effects on depression, 24-
MADRS Score

20
70 hour UFC excretion, diurnal cortisol level, and get-up time
60 after final awakening persisted, improved, or became sig-
15 50 nificant only at T2, whereas the other sleep measures and
melatonin levels changed during BLT but returned to base-
40
10
line at T2. The finding suggests rather acute effects on mela-
30
T0 T1 T2 T0 T1 T2 tonin levels and sleep, whereas effects on clinical improve-
ment in depression symptoms and cortisol hyperactivity
Figure 5. Effects of bright light treatment (BLT) and placebo in elderly are initiated by the treatment but take longer to develop
patients with nonseasonal major depressive disorder. Data are depicted as fully. To the best of our knowledge, we are the first to re-
means; error bars show the 95% confidence intervals. Absolute values are port that, in elderly patients with MDD, 24-hour UFC and
given on the left side, and the percentage of change from baseline (T0) is
shown on the right side. Measures include the Hamilton Scale for Depression
diurnal salivary cortisol levels attenuated after BLT
(HAM-D) scores (A), the HAM-D6 (the HAM-D 6-item core version) scores (Figure 6). In contrast, placebo-treated patients contin-
(B), Atypical Symptom Scale scores (C), the Structured Interview Guide for ued to increase their 24-hour UFC levels. We hypothesize
the HAM-D–Seasonal Affective Disorder Version (SIGH-SAD) scores48,49 (D), that the burden and stress of participating in a clinical trial
and the Montgomery-Åsberg Depression Rating Scale (MADRS) scores (E).
T1 indicates after 3 weeks of treatment; T2, 3 weeks after discontinuation of with disappointing treatment effects may have further el-
treatment. evated HPA activity. Alternatively, a continued increase of

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
68

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 24-hour UFC levels may be a characteristic of the devel- Correspondence: Ritsaert Lieverse, MD, Department of
opmental time course of MDD in elderly patients. Psychiatry, VU University Medical Center and GGZ in-
Several limitations should be discussed. First, at base- Geest, AJ Ernststraat 887, 1081 HL Amsterdam, the Neth-
line a slight randomization imbalance for outcome was seen erlands ([email protected]).
for HAM-D scores, indicating that BLT-treated patients had Author Contributions: Dr Lieverse had full access to all
slightly higher pretreatment severity ratings than placebo- the data in the study and takes responsibility for the in-
treated patients. This difference was not reflected in the other tegrity of the data and the accuracy of the data analysis.
depression severity ratings, in severity distribution, or in Financial Disclosure: None reported.
other depression characteristics. All analyses took this into Funding/Support: The study was supported by grant 014-
accountbyincludingbaselineseveritycovariatesintheanaly- 91-049 from the Successful Aging Program of the Dutch
ses. Significance of the covariate-corrected treatment effects Scientific Organization (NWO) and grant 940-37-033,
indicated that the antidepressant effects of BLT could not the AGIKO stipendium, from the Chronicity Care Pro-
be attributed to HAM-D pretreatment score differences. Sec- gram of the Dutch Scientific Organization. Philips Light-
ond, the monitoring of depression symptoms was limited ing donated 12 bright light devices for the project.
to T1 and T2. If the developmental course of improvement Online-Only Material: The eTables, eFigures, and eAp-
is the focus of interest, more frequent assessments for more pendixes are available at http://www.archgenpsychiatry
detailed analyses will be required. Moreover, with the posi- .com.
tive effect of BLT that we found, more data points would Additional Contributions: The research psychologists Rin-
have further increased the statistical significance. Third, our ske de Vries, MPsy (GGZ inGeest, the Netherlands Study
trial investigated only the immediate and 3-week delayed of Depression and Anxiety [NESDA]), Natalie Ran, MSc,
effect of a 3-week BLT treatment duration. Therefore, pro- and Janneke van Leeuwen, MSci (GGZ inGeest, Amster-
longed effects, or effects of long-term BLT, remain to be in- dam Study of Anxiety and Depression) assisted in recruit-
vestigated. A large study on long-term effects of light treat- ment and psychometry; the psychologists Zsuzsika Sjoerds,
ment on demented elderly patients without MDD suggests MSc (NESDA), and Hester Duivis, MSc (NESDA), patient
preservation of antidepressant effects rather than habitu- and technical device instructions; Tom van den Berg, PhD
ation.11 Fourth, only 89 patients were included from a total (Netherlands Institute for Neuroscience), spectrophotom-
of 444 undergoing assessment. This could have been due etry analyses; Jolanda Verhagen, BSc (Leiden University
to several factors, including (1) active case-finding efforts, Medical Center laboratory), saliva cortisol and melatonin
(2) strict inclusion criteria to fulfill the requirements for a analyses; Marie Lomecky, BSc, and Jeany Huijser, BSc (VU
diagnosis of MDD only, and (3) the criterion of absence of Medical Center laboratory), 24-hour urinary cortisol and
seasonal affective disorder. Although the findings of this creatinine analyses; and Wilma Verweij, MA (Nether-
specific study are thus limited to elderly patients with MDD, lands Institute for Neuroscience), language corrections.
efficacy of light treatment in elderly patients with a profile
of milder depression is suggested by previous work.11 REFERENCES
In conclusion, we showed that BLT had beneficial ef-
fects in elderly patients with nonseasonal MDD and found 1. Bunney WE, Bunney BG. Molecular clock genes in man and lower animals: possible
implications for circadian abnormalities in depression. Neuropsychopharmacology.
indirect support for the contention that therapeutic ef- 2000;22(4):335-345.
fects may in part be mediated by enhancements of circa- 2. Boivin DB. Influence of sleep-wake and circadian rhythm disturbances in psy-
dian system functioning. These results support inclusion chiatric disorders. J Psychiatry Neurosci. 2000;25(5):446-458.
3. Grandin LD, Alloy LB, Abramson LY. The social zeitgeber theory, circadian rhythms,
of chronotherapeutic strategies in the treatment options for and mood disorders: review and evaluation. Clin Psychol Rev. 2006;26(6):
nonseasonal MDD in elderly patients. Bright light treat- 679-694.
4. Rusting CL, Larsen RJ. Diurnal patterns of unpleasant mood: associations with
ment may provide a viable alternative for patients who neuroticism, depression, and anxiety. J Pers. 1998;66(1):85-103.
refuse, resist, or do not tolerate antidepressant treatment. 5. Deuschle M, Schweiger U, Weber B, Gotthardt U, Körner A, Schmider J, Stand-
hardt H, Lammers CH, Heuser I. Diurnal activity and pulsatility of the hypothalamus-
pituitary-adrenal system in male depressed patients and healthy controls. J Clin
Submitted for Publication: November 30, 2009; final re- Endocrinol Metab. 1997;82(1):234-238.
vision received June 3, 2010; accepted June 3, 2010. 6. Hoogendijk WJ, van Someren EJ, Mirmiran M, Hofman MA, Lucassen PJ, Zhou
Author Affiliations: Departments of Psychiatry (Drs Liev- JN, Swaab DF. Circadian rhythm–related behavioral disturbances and structural
hypothalamic changes in Alzheimer’s disease. Int Psychogeriatr. 1996;8(suppl
erse, Nielen, Smit, and Hoogendijk), Integrative Neuro- 3):245-252, 269-272.
physiology (Dr Van Someren), and Neurology (Dr Uitde- 7. Zhou JN, Riemersma RF, Unmehopa UA, Hoogendijk WJ, van Heerikhuize JJ, Hof-
man MA, Swaab DF. Alterations in arginine vasopressin neurons in the suprachi-
haag) and GGZ inGeest (Drs Lieverse, Nielen, Smit, and asmatic nucleus in depression. Arch Gen Psychiatry. 2001;58(7):655-662.
Hoogendijk), Neuroscience Campus, and Department of 8. Gold PW, Chrousos GP. Organization of the stress system and its dysregulation
Epidemiology and Biostatistics (Dr Uitdehaag), VU Uni- in melancholic and atypical depression: high vs low CRH/NE states. Mol Psychiatry.
2002;7(3):254-275.
versity Medical Center, Netherlands Institute for Neuro- 9. Terman M, Terman JS. Light therapy for seasonal and nonseasonal depression:
science, Royal Netherlands Academy of Arts and Sciences efficacy, protocol, safety, and side effects. CNS Spectr. 2005;10(8):647-663, 672.
(Dr Van Someren), and Center for Neurogenomics and Cog- 10. Wirz-Justice A, Benedetti F, Berger M, Lam RW, Martiny K, Terman M, Wu JC.
Chronotherapeutics (light and wake therapy) in affective disorders. Psychol Med.
nitive Research, VU University (Dr Hoogendijk), Amster- 2005;35(7):939-944.
dam; AmaCura, Limburg (Dr Lieverse); Leiden Institute 11. Riemersma-van der Lek RF, Swaab DF, Twisk J, Hol EM, Hoogendijk WJ, Van
Someren EJ. Effect of bright light and melatonin on cognitive and noncognitive
for the Clinical and Experimental Neuroscience of Sleep, function in elderly residents of group care facilities: a randomized controlled trial.
Department of Neurology, Leiden University Medical Cen- JAMA. 2008;299(22):2642-2655.
ter, Leiden (Dr Van Someren); and Department of Psychia- 12. McEnany GW, Lee KA. Effects of light therapy on sleep, mood, and temperature
in women with nonseasonal major depression. Issues Ment Health Nurs. 2005;
try, Erasmus University Medical Center, Rotterdam 26(7):781-794.
(Dr Hoogendijk), the Netherlands. 13. Thalén BE, Mørkrid L, Kjellman BF, Wetterberg L. Cortisol in light treatment of

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
69

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023
 seasonal and non-seasonal depression: relationship between melatonin and cortisol. 47. Hamilton M. Development of a rating scale for primary depressive illness. Br J
Acta Psychiatr Scand. 1997;96(5):385-394. Soc Clin Psychol. 1967;6(4):278-296.
14. Scheer FA, Buijs RM. Light affects morning salivary cortisol in humans. J Clin 48. Terman M, Williams JBW. SAD assessment tools revisited. Light Treat Biol Rhythms.
Endocrinol Metab. 1999;84(9):3395-3398. 1994;7:23.
15. Berson DM, Dunn FA, Takao M. Phototransduction by retinal ganglion cells that 49. Williams JBW, Link MJ, Rosenthal NE, Amira L, Terman M. Structured Inter-
set the circadian clock. Science. 2002;295(5557):1070-1073. view Guide for the Hamilton Depression Rating Scale–Seasonal Affective Disor-
16. Hannibal J, Fahrenkrug J. Melanopsin: a novel photopigment involved in the pho- der Version (SIGH-SAD). New York: New York State Psychiatric Institute; 1992.
toentrainment of the brain’s biological clock? Ann Med. 2002;34(5):401-407. 50. O’Sullivan RL, Fava M, Agustin C, Baer L, Rosenbaum JF. Sensitivity of the six-item
17. Hattar S, Liao HW, Takao M, Berson DM, Yau KW. Melanopsin-containing reti- Hamilton Depression Rating Scale. Acta Psychiatr Scand. 1997;95(5):379-384.
nal ganglion cells: architecture, projections, and intrinsic photosensitivity. Science. 51. Martiny K. Adjunctive bright light in non-seasonal major depression. Acta Psy-
2002;295(5557):1065-1070. chiatr Scand Suppl. 2004;(425):7-28.
18. Benedetti F, Barbini B, Colombo C, Smeraldi E. Chronotherapeutics in a psychi- 52. Terman M, Amira L, Terman JS, Ross DC. Predictors of response and nonre-
atric ward. Sleep Med Rev. 2007;11(6):509-522. sponse to light treatment for winter depression. Am J Psychiatry. 1996;153
19. Wirz-Justice A. From the basic neuroscience of circadian clock function to light (11):1423-1429.
therapy for depression: on the emergence of chronotherapeutics. J Affect Disord. 53. Rosenthal NE, Genhart M, Sack DA, Skwerer RG, Wehr TA. Seasonal affective dis-
2009;116(3):159-160. order: relevance for treatment and research of bulimia. In: Hudson JL, Pope HG,
20. Frazão R, Pinato L, da Silva AV, Britto LR, Oliveira JA, Nogueira MI. Evidence of eds. Psychobiology of Bulimia. Washington, DC: American Psychiatric Press; 1987.
reciprocal connections between the dorsal raphe nucleus and the retina in the 54. LieverseR,NielenMM,VeltmanDJ,UitdehaagBM,vanSomerenEJ,SmitJH,Hoogendijk
monkey Cebus apella. Neurosci Lett. 2008;430(2):119-123. WJ. Bright light in elderly subjects with nonseasonal major depressive disorder: a
21. Kawano H, Decker K, Reuss S. Is there a direct retina-raphe-suprachiasmatic nucleus double blind randomised clinical trial using early morning bright blue light compar-
pathway in the rat? Neurosci Lett. 1996;212(2):143-146. ing dim red light treatment. Trials. July 31, 2008;9:48. doi:10.1186/1745-6215-9-48.
22. Campbell SS, Kripke DF, Gillin JC, Hrubovcak JC. Exposure to light in healthy 55. Parmelee PA, Thuras PD, Katz IR, Lawton MP. Validation of the Cumulative Ill-
elderly subjects and Alzheimer’s patients. Physiol Behav. 1988;42(2):141-144. ness Rating Scale in a geriatric residential population. J Am Geriatr Soc. 1995;
23. Espiritu RC, Kripke DF, Ancoli-Israel S, Mowen MA, Mason WJ, Fell RL, Klauber 43(2):130-137.
MR, Kaplan OJ. Low illumination experienced by San Diego adults: association 56. First MB, Spitzer RL, Gibbon M, Williams JB. Structured Clinical Interview for
with atypical depressive symptoms. Biol Psychiatry. 1994;35(6):403-407. DSM-IV Axis I Disorders–Patient Edition. SCID-I/P, Version 2.0. New York: New
24. Turner PL, Mainster MA. Circadian photoreception: ageing and the eye’s impor- York Psychiatric Institute, Biometric Research Dept; 1996.
tant role in systemic health. Br J Ophthalmol. 2008;92(11):1439-1444. 57. Martiny K, Lunde M, Undén M, Dam H, Bech P. Adjunctive bright light in non-
25. Van Someren EJ, Riemersma-Van Der Lek RF. Live to the rhythm, slave to the seasonal major depression: results from clinician-rated depression scales. Acta
rhythm. Sleep Med Rev. 2007;11(6):465-484. Psychiatr Scand. 2005;112(2):117-125.
26. Golden RN, Gaynes BN, Ekstrom RD, Hamer RM, Jacobsen FM, Suppes T, Wis- 58. Brainard GC, Sliney D, Hanifin JP, Glickman G, Byrne B, Greeson JM, Jasser S,
ner KL, Nemeroff CB. The efficacy of light therapy in the treatment of mood dis- Gerner E, Rollag MD. Sensitivity of the human circadian system to short-
orders: a review and meta-analysis of the evidence. Am J Psychiatry. 2005; wavelength (420-nm) light. J Biol Rhythms. 2008;23(5):379-386.
162(4):656-662. 59. Beauchemin KM, Hays P. Sunny hospital rooms expedite recovery from severe
27. Martiny K, Lunde M, Simonsen C, Clemmensen L, Poulsen DL, Solstad K, Bech P. and refractory depressions. J Affect Disord. 1996;40(1-2):49-51.
RelapsepreventionbycitalopraminSADpatientsrespondingto1weekoflighttherapy: 60. Deltito JA, Moline M, Pollak C, Martin LY, Maremmani I. Effects of phototherapy
a placebo-controlled study. Acta Psychiatr Scand. 2004;109(3):230-234. on non-seasonal unipolar and bipolar depressive spectrum disorders. J Affect
28. Labbate LA, Lafer B, Thibault A, Sachs GS. Side effects induced by bright light treat- Disord. 1991;23(4):231-237.
ment for seasonal affective disorder. J Clin Psychiatry. 1994;55(5):189-191. 61. Kripke DF. Therapeutic effects of bright light in depressed patients. Ann N Y Acad
29. Kripke DF, Tuunainen A, Endo T. Benefits of light treatment for depression. Am Sci. 1985;453:270-281.
J Psychiatry. 2006;163(1):162-163. 62. Kripke DF. Light treatment for nonseasonal depression: speed, efficacy, and com-
30. Even C, Schröder CM, Friedman S, Rouillon F. Efficacy of light therapy in nonsea- bined treatment. J Affect Disord. 1998;49(2):109-117.
sonal depression: a systematic review. J Affect Disord. 2008;108(1-2):11-23. 63. Kripke DF, Risch SC, Janowsky D. Bright white light alleviates depression. Psy-
31. Tuunainen A, Kripke DF, Endo T. Light therapy for non-seasonal depression. Coch- chiatry Res. 1983;10(2):105-112.
rane Database Syst Rev. 2004;(2):CD004050. 64. Loving RT, Kripke DF, Shuchter SR. Bright light augments antidepressant ef-
32. Forbes D, Morgan DG, Bangma J, Peacock S, Pelletier N, Adamson J. Light therapy fects of medication and wake therapy. Depress Anxiety. 2002;16(1):1-3.
for managing sleep, behaviour, and mood disturbances in dementia. Cochrane 65. Mackert A, Volz HP, Stieglitz RD, Müller-Oerlinghausen B. Phototherapy in non-
Database Syst Rev. 2004;(2):CD003946. seasonal depression. Biol Psychiatry. 1991;30(3):257-268.
33. Kim S, Song HH, Yoo SJ. The effect of bright light on sleep and behavior in de- 66. Neumeister A, Goessler R, Lucht M, Kapitany T, Bamas C, Kasper S. Bright light
mentia: an analytic review. Geriatr Nurs. 2003;24(4):239-243. therapy stabilizes the antidepressant effect of partial sleep deprivation. Biol
34. Loving RT, Kripke DF, Elliott JA, Knickerbocker NC, Grandner MA. Bright light Psychiatry. 1996;39(1):16-21.
treatment of depression for older adults [ISRCTN55452501]. BMC Psychiatry. 67. Yamada N, Martin-Iverson MT, Daimon K, Tsujimoto T, Takahashi S. Clinical and
November 9, 2005;5:41. doi:10.1186/1471-244X-5-41. chronobiological effects of light therapy on nonseasonal affective disorders. Biol
35. Loving RT, Kripke DF, Knickerbocker NC, Grandner MA. Bright green light treat- Psychiatry. 1995;37(12):866-873.
ment of depression for older adults [ISRCTN69400161]. BMC Psychiatry. No- 68. Colombo C, Lucca A, Benedetti F, Barbini B, Campori E, Smeraldi E. Total sleep
vember 9, 2005;5:42. doi:10.1186/1471-244X-5-42. deprivation combined with lithium and light therapy in the treatment of bipolar
36. Terman M. Evolving applications of light therapy. Sleep Med Rev. 2007;11(6):497- depression: replication of main effects and interaction. Psychiatry Res. 2000;
507. 95(1):43-53.
37. Sumaya IC, Rienzi BM, Deegan JF II, Moss DE. Bright light treatment decreases 69. Bloching B, Dechêne C, Täschner KL. Outlasting antidepressant effect of late par-
depression in institutionalized older adults: a placebo-controlled crossover study. tial sleep deprivation by bright light therapy. J Sleep Res. 2000;9(suppl 1):21.
J Gerontol A Biol Sci Med Sci. 2001;56(6):M356-M360. 70. Newell-Price J, Trainer P, Besser M, Grossman A. The diagnosis and differential
38. Tsai YF, Wong TK, Juang YY, Tsai HH. The effects of light therapy on depressed diagnosis of Cushing’s syndrome and pseudo-Cushing’s states. Endocr Rev. 1998;
elders. Int J Geriatr Psychiatry. 2004;19(6):545-548. 19(5):647-672.
39. World Medical Association. Declaration of Helsinki: Ethical Principles for Medi- 71. Schatzberg AF, Samson JA, Bloomingdale KL, Orsulak PJ, Gerson B, Kizuka PP,
cal Research Involving Human Subjects. 5th rev. Edinburgh, Scotland: World Medi- Cole JO, Schildkraut JJ. Toward a biochemical classification of depressive dis-
cal Association; 2000. orders, X: urinary catecholamines, their metabolites, and D-type scores in sub-
40. Volz HP, Mackert A, Stieglitz RD. Side-effects of phototherapy in nonseasonal groups of depressive disorders. Arch Gen Psychiatry. 1989;46(3):260-268.
depressive disorder. Pharmacopsychiatry. 1991;24(4):141-143. 72. Van Someren EJ, Nagtegaal E. Improving melatonin circadian phase estimates.
41. Kripke DF, Mullaney DJ, Klauber MR, Risch SC, Gillin JC. Controlled trial of bright Sleep Med. 2007;8(6):590-601.
light for nonseasonal major depressive disorders. Biol Psychiatry. 1992;31 73. Ancoli-Israel S, Clopton P, Klauber MR, Fell R, Mason W. Use of wrist activity
(2):119-134. for monitoring sleep/wake in demented nursing-home patients. Sleep. 1997;
42. Baumgartner A, Volz HP, Campos-Barros A, Stieglitz RD, Mansmann U, Mack- 20(1):24-27.
ert A. Serum concentrations of thyroid hormones in patients with nonseasonal 74. Kushida CA, Chang A, Gadkary C, Guilleminault C, Carrillo O, Dement WC. Com-
affective disorders during treatment with bright and dim light. Biol Psychiatry. parison of actigraphic, polysomnographic, and subjective assessment of sleep
1996;40(9):899-907. parameters in sleep-disordered patients. Sleep Med. 2001;2(5):389-396.
43. Cohen J, ed. Statistical Power Analysis for the Behavioral Sciences. 2nd ed. Hillsdale, 75. Sacket DL, Straus SE, Richardson WS, Rosenberg W, Haynes RB. Evidence-
NJ: Lawrence Erlbaum Associates Inc; 1998. Based Medicine: How to Practice and Teach EBM. 2nd ed. New York, NY: Churchill
44. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Dis- Livingstone; 2000.
orders, 4th ed, text revision (DSM-IV-TR). Washington, DC: American Psychiatric 76. Altman DG. Confidence intervals for the number needed to treat. BMJ. 1998;317
Press; 2000. (7168):1309-1312.
45. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, Leirer VO. Develop- 77. Laupacis A, Sackett DL, Roberts RS. An assessment of clinically useful measures
ment and validation of a geriatric depression screening scale: a preliminary report. of the consequences of treatment. N Engl J Med. 1988;318(26):1728-1733.
J Psychiatr Res. 1982-1983;17(1):37-49. 78. Martiny K, Lunde M, Undén M, Dam H, Bech P. The lack of sustained effect of
46. Montgomery SA, Åsberg M. A new depression scale designed to be sensitive to bright light, after discontinuation, in non-seasonal major depression. Psychol
change. Br J Psychiatry. April 1979;134:382-389. Med. 2006;36(9):1247-1252.

(REPRINTED) ARCH GEN PSYCHIATRY/ VOL 68 (NO. 1), JAN 2011 WWW.ARCHGENPSYCHIATRY.COM
70

©2011 American Medical Association. All rights reserved.

Downloaded From: https://jamanetwork.com/ on 03/18/2023