Content uploaded by Michael Terman
Author content
All content in this area was uploaded by Michael Terman on Sep 03, 2018
Content may be subject to copyright.
Content uploaded by Michael Terman
Author content
All content in this area was uploaded by Michael Terman on Sep 03, 2018
Content may be subject to copyright.
Sleep Medicine Reviews (2007) 11, 497–507
CLINICAL REVIEW
Evolving applications of light therapy
$
Michael Terman
Department of Psychiatry, Columbia University, New York State Psychiatric Institute, 1051 Riverside Drive,
Unit 50, New York, NY 10032, USA
KEYWORDS
Light therapy;
Depression;
Bipolar disorder;
Sleep;
Circadian rhythms;
Evidence-based
medicine;
Seasonal affective
disorder;
Attention deficit hy-
peractivity disorder;
Dementia;
Parkinson’s disease;
Bulimia nervosa
Summary The psychiatric intervention, light therapy, grew from an intensive
25-year research focus on seasonal affective disorder (SAD). Dosing and timing
strategies have been honed to optimize the antidepressant effect, and efficacy
relative to placebo has provided the evidence base for widespread implementation.
A persistent question has been whether the model system for SAD has wider utility
for psychiatric disturbance, even beyond depression. The circadian phase-shifting
capacity of timed light exposure is universal, and chronobiological factors are at play
across the disease spectrum. Recent promising initiatives extend to light treatment
for nonseasonal major depressive disorder and bipolar depression, including drug-
and electroconvulsive therapy-resistant cases. With light therapy, patients with
antepartum depression may find an alternative to medication during pregnancy.
Cognitive improvement under light therapy has been noted in adult attention deficit
hyperactivity disorder. Motor function in Parkinson’s disease has improved in parallel
with the antidepressant effect of light therapy. The rest–activity disturbance of
elderly dementia has been partially allayed under light therapy. In a new initiative,
three major chronotherapeutic inventions—light therapy, sleep deprivation (wake
therapy) and sleep time displacement (sleep phase advance therapy) are being
combined to snap hospitalized patients out of deep depression and maintain long-
term improvement.
&2007 Elsevier Ltd. All rights reserved.
I have been asked to review the Future, a nice
challenge. The focus of course must be on ‘‘pre-
liminary data,’’ a term that merits some critical
comment.
Conventional clinical trials of light therapy are
difficult if not impossible to achieve. There is the
placebo control bugaboo—one cannot literally
blind a light treatment study, and naysayers will
always be able to dismiss the results of otherwise-
controlled studies. Light therapy evolved from
biologically oriented thinking. The original control
groups were not, strictly speaking, placebos.
Rather, they were active experimental manipula-
tions within or between subjects, such as the time
of day of light administration (the circadian phase-
shifting variable) and the dosing trade-off between
light intensity and exposure duration. Even though
such studies inform mechanisms of action, some
meta-analysts see in them no evidence of specific
ARTICLE IN PRESS
www.elsevier.com/locate/smrv
1087-0792/$ - see front matter &2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.smrv.2007.06.003
$
Dedicated to Anna Wirz-Justice in recognition of her
contributions to the field made during her career at the
Psychiatric University Clinics Basel.
E-mail address: mt12@columbia.edu
antidepressant activity. To conclude that such
studies are not ‘‘evidence-based’’ is no gift to
science or to patients, and it downgrades the
research initiative.
Selection of studies for meta-analyses can be
heavily biased, despite the formalistic listings of
‘‘high-quality’’ inclusion criteria. I have been asked
on occasion to comment on meta-analyses in
progress. In one case, my challenge of inclusion
criteria (which had led to negative conclusions for
antidepressant efficacy) effectively terminated the
project rather than recasting the analyses in a more
comprehensive manner. The study director told me,
‘‘I know nothing about psychiatric rating scales.’’
An obvious problem is that evidence-based meta-
analysis is, by its nature, outdated when a research
field such as ours is active. Clinical thinking and
judgment are often years ahead of evidence-based
post-hoc compilations. In an underfunded field such
as light therapy, meta-analyses are often based on
very few studies, and one or two additional trials
(completed but not included) might easily reverse
conclusions. Periodic updates—as by the Cochrane
Collaboration—are salutary, but they are still
victims of inertia.
Even when investigators forgo mechanistic stu-
dies in favor of placebo-controlled yes–no designs,
their efforts can be stymied by the mismatch
between evidence-based criteria and clinical un-
derstanding. For example, inclusion in a meta-
analysis by a national health standards agency
required a placebo control. In the same country,
researchers were prohibited by their ethics com-
mittee to use a placebo given that the efficacy of
light therapy ‘‘is already well established.’’ The
meta-analysis could therefore only include studies
from other countries.
Bright light therapy for seasonal
affective disorder (SAD)
This pioneering research field, now 25 years old,
has undergone three sets of challenges: (a)
conceptual affront to mainline antidepressant
pharmacotherapy; (b) difficulty honing convincing
placebo controls; and (c) lacking industry support
for major clinical trials, small sample sizes with
inherent outcome variability. Nonetheless, perse-
verance has finally produced consensus even in
meta-analytic terms. An American Psychiatric
Association (APA) work group of psychopharmacol-
ogists was able to glean 8 of 45 potential studies
for consideration.
1
Sample sizes varied widely
(7–85) but the designs were randomized with
‘‘acceptable’’ placebo controls including low-in-
tensity light and an inert or low-output negative air
ionizer. The work group concluded that there was a
‘‘significant reduction in depression symptom se-
verity following bright light therapyy. In other
words, when the ‘noise’ from unreliable studies is
removed, the effects of light therapy are compar-
able to those found in many antidepressant
pharmacotherapy trials’’ (p. 660).
Had studies with biologically active controls been
included, the results would have been even more
convincing. In reviewing a trio of the largest
studies, Wirz-Justice
5
integrated the two types of
evidence: ‘‘Light is as effective as drugs, perhaps
more soy. In spite of the differences in design [of
these 3 studies], some important correspondences
emerge with respect to remission rates. The 2
placebo-controlled trials
2,3
yhave nearly identical
results: both morning and evening light are better
than placebo, and morning light is superior to
evening light. The third study
4
also demonstrates a
morning light superiority [without a placebo con-
trol]y. These comparisons of therapeutic outcome
are based on very stringent criteria for remission,
not just response, within a rather short time
(2–4 weeks). Such stringent criteria, when applied
to a 5-week multicenter trial of fluoxetine in
patients with SAD, did not differentiate between
drug (33%, n¼36) and placebo (28%, n¼32).
Fourteen patients with SAD treated with light for
5 weeks tended to remit more (50%, n¼20) than
those treated with fluoxetine (25%, n¼20;
P¼0.10).
6
’’
The timing of light relative to endogenous
melatonin onset—a basic circadian rhythm con-
cept—is a further refinement that homes in on one
of the antidepressant mechanisms of light therapy
and provides a means to optimize clinical response.
Within the favored morning interval, light adminis-
tered 7.5–9.5 h after evening melatonin onset
produces twice the remission rate (approximately
80% vs. 40%) of light presented 9.5–11 h after
melatonin onset.
7
Concurrently, the phase advance
to earlier light exposure is larger than that to later
light exposure. Clinically, this information is useless
unless one knows the patient’s circadian rhythm
phase, since melatonin onsets can vary between
patients by as much as 6 h. A practical solution is
found in the Horne-O
¨stberg Morningness–Evening-
ness Questionnaire (MEQ) score,
8
which measures
diurnal preference for activities and correlates
strongly with melatonin onset.
9
Based on the score,
the clinician can approximately infer melatonin
onset and guide the patient toward the optimum
timing of morning light. Yet these refinements
cannot be considered evidence-based.
ARTICLE IN PRESS
M. Terman498
Dawn simulation therapy for SAD
Dawn simulation presents a slow, incremental light
signal in the bedroom at the end of the sleep
interval, with maximum intensity two orders of
magnitude lower than in post-awakening bright
light therapy (e.g., 300 lux vs. 10,000 lux).
10
Until
recently, all the controlled studies were from David
Avery’s group in Seattle. Nonetheless, they have
been meta-analyzed. (How can this be done for one
center, when meta-analysis is meant to derive a
conclusion across centers?) Dawn simulation
showed efficacy relative to dimmer or briefer
signals, enough to impress the APA work group.
1
An additional, later study from my group rounds out
the analysis.
11
Here, the proportion of subjects
with depression rating scale improvement of 50% or
more was 0.62 for dawn simulation, 0.63 for bright
light therapy and 0.17 for the low-output negative
air ionizer placebo. The proportion of subjects with
depression rating scale improvement of 50% or
more was 0.86 when termination of the dawn signal
(simulated sunrise time) was scheduled 8–9 h after
melatonin onset—according to the metric of
circadian time—and 0.47 for later sunrises (Terman
JS & Terman M, in preparation). Since dawn
simulation appears to match post-awakening bright
light in efficacy, it may become the next-genera-
tion light therapy given its convenient use while the
patient sleeps. This might undermine various
industry press releases and publications
12
asserting
patients prefer a drug for SAD because of the daily
‘‘burden’’ of light therapy sessions.
Bright light therapy for nonseasonal
depression
Studies of light therapy for nonseasonal depression
have a history at least as long as studies of SAD, but
on the whole the results have been less clear-cut.
13
The APA work group found, within its selection,
positive evidence for efficacy except when light
therapy and medication were combined.
1
Studies
completed by the time of publication clearly would
have reversed that conclusion.
14–16
The adjunctive
use of light with medication is potent. The strategy
has been recommended by the Committee on
Chronotherapeutics of the International Society
for Affective Disorders
17
and in an international
response
18
to a review of new antidepressants that
overlooked light therapy, published in Science.
18
A difficulty with most nonseasonal studies has
been their inability to confront the early hypothesis
that light therapy is specifically tuned to patients
with SAD as a countermeasure to long winter
nights. Seasonality lies on a continuous dimension
from noticeable (but not disturbing) to mildly,
moderately and severely disturbing.
19
SAD falls into
the latter category, with major depressive episodes
restricted to winter. In a far larger number of cases,
recurrent or chronic depressions are exacerbated
in winter but can occur at any time of year. Such
patients provide moderate global seasonality
scores
20
in comparison to higher scores for SAD.
Thus, patients with nonseasonal depressions can
still show seasonality, which might be the key to
their response to light therapy. Subsequent to an
inconclusive meta-analysis of light therapy for
nonseasonal depression,
13
we sought to clarify this
issue with a patient sample in which depression was
chronic (at least 2 years, but most often far longer)
and without any seasonal modulation.
21
Under
morning light therapy, the proportion of subjects
with depression rating scale improvement of 50% or
more was 0.60 vs. 0.10 for the low-output negative
air ionizer placebo. We can begin to surmise that
light therapy for seasonal and nonseasonal depres-
sion is equally effective.
Why might this equivalence be? Are patients with
nonseasonal depression light deprived at any time
of year? Perhaps they are—spending most of their
time indoors, often in bed with shades drawn,
‘‘escaping the world.’’ What ensues? Exacerbation
of circadian rhythm phase delay, given the absence
of the critical early morning light signal that
synchronizes the internal clock to local time. Such
delay may be depressogenic regardless of the
season.
The mood in psychiatric practice is gradually
changing, at least among the younger generation.
Psychiatric residents and medical students are
asking for and beginning to receive training in light
therapy and chronobiology. One resident recently
reported that he now routinely prescribes light
therapy for any depressed patient who expresses
hesitation about antidepressant medication. The
growing concern about medication side effects
would seem to augur well for our alternative.
Back to the future
As we move beyond ‘‘definitive’’ randomized,
controlled trials we arrive at the terra incognita
of small, underpowered, controlled trials; open
trials; systematic case series; and one-of-a-kind
clinical observations, some with internal controls
(dose–response, discontinuation and resumption).
Here is where the action lies. Given the clear-cut
ARTICLE IN PRESS
Evolution of light therapy 499
results of light therapy for SAD, clinicians can view
this new evidence more confidently than was
possible when light therapy was novel. Since new
applications are proliferating, there is no way they
will pass the evidence-based filter for many years,
if ever. Yet some of these data oblige clinical
application now.
Refractory unipolar depression
Treatment-resistant depression remains prevalent
and vexing. In the present context, our objective
was to see if light therapy is effective—by itself or
as an adjuvant—in patients who have been
unresponsive to conventional drug treatment,
electroconvulsive therapy (ECT), or both. The
opportunity arises out of desperation, as when a
suicidal hospitalized patient does not improve.
Having heard about light therapy, the attending
psychiatrist says, ‘‘Why not try it?’’ The chron-
otherapist (M.T.) evaluates the sleep pattern with-
out objective measurement, obtains an MEQ
chronotype score if the patient can manage the
questionnaire, and immediately begins a course of
light therapy under nursing supervision. The nurses
maintain a daily log of session start and stop times,
and note any difficulties such as side effects or the
patient’s refusal to comply.
Light therapy is not introduced as an alternative,
but always as an add-on to treatments (antide-
pressants, mood stabilizers, antipsychotics and
ECT) that have not yielded adequate, if any
improvement. In some cases, hypnotic medication
is discontinued, tapered or rescheduled when the
chronotherapist suspects masking (induction) of the
sleep episode out of phase with the circadian clock.
The chronotherapist may also request that patients
not be awakened for testing of vital signs before
their normal wake-up time. In my clinical experi-
ence, early awakening followed by a ‘‘second
sleep’’ is itself depressogenic.
In one exploratory effort (collaboration Stewart
et al., M.D.
22
), treatment-resistant patients were
first given tranylcypromine—they had not pre-
viously received monoamine oxidase inhibitors
(MAOIs)—building to the highest tolerable dose
(p120 mg). About 40% of these patients responded.
After finding little or no improvement, four
patients began morning 10,000 lux light therapy.
Timing of treatment was determined on the basis
of their MEQ scores. The initial duration of
light exposure was 30 min, and was adjusted
upward individually if there was no response
within 2 weeks, or downward if there were side
effects (headache, early awakening). Patients also
maintained sleep logs and daily mood and energy
ratings, and were evaluated weekly on the Hamil-
ton Depression Rating Scale. Here are the case
summaries.
Mr. A was 37 years old, separated, with 3 years of
melancholic depression, including a suicide at-
tempt within the last year. Earlier, he had shown
an unusual seasonal pattern with relapses in March
and remissions in late summer. He had undergone
prior trials of multiple drug classes, some with
thyroid hormone or lithium augmentation, all
without response. He was also nonresponsive to
multiple unilateral and bilateral ECT. Under tra-
nylcypromine 120 mg, patient showed typical de-
pression- and MAOI-related sleep disturbance, with
early, middle and late insomnia. After starting light
therapy for 30 min at 07:15 h, the patient experi-
enced headache and early awakening. Duration was
reduced to 20 min and then edged up to 25 min.
Progress was complicated by several emergency
MAOI-induced episodes of heightened blood pres-
sure, controlled by a combination of a calcium
channel blocker and beta-blocker. During this
period he also reacted badly to a family split-up.
However, within 3 months of light therapy his
Hamilton score gradually reduced by 53% and his
sleep became coalesced, regular (23:30–07:00 h)
and satisfying. He was discharged and continued
with light + tranylcypromine at home, with a
stable, positive response.
Ms. B was 24 years old, single, with a lifetime
history of dysthymia and a past history anorexia and
social phobia. For the last 6 years, she suffered
chronic major depression. She had been nonre-
sponsive to multiple drug trials. However, in one
trial she was given a low dose of the MAOI
phenelzine with partial response. The dose should
have been raised, but this was impossible because
manufacturing of the drug had been discontinued
at the time. After terminating treatment, she
relapsed and attempted suicide. Under tranylcy-
promine 100 mg, the patient showed early, middle
and late insomnia. Light therapy was taken at
07:15 h for 30 min. Sleep promptly coalesced
(23:30–07:00 h) and within 3 weeks the patient
showed complete remission (Hamilton score ¼0)
and was discharged. She continued with light +
tranylcypromine at home, but was not compliant
with light treatment. Several times, whenever she
stopped using the light, she would experience
relapse within 2 days. On resumption of the light,
she would feel improvement within 2 days and
complete remission in 4 days.
Ms. C was 41 years old, single, with a lifetime
history of dysthymia and four prior major depres-
sive episodes. She experienced intense suicidal
ARTICLE IN PRESS
M. Terman500
ideation, but there were no attempts. She was
unresponsive to multiple drug trials, psychother-
apy, psychoanalysis and cognitive behavioral ther-
apy. Under tranylcypromine 60 mg, she experienced
initial insomnia and restless sleep. Although her
Hamilton score did not change during hospitaliza-
tion, clinically she had deteriorated. With an MEQ
score of 29, she fell into a tail of the distribution as
a ‘‘definite evening type,’’ possibly with delayed
sleep phase syndrome. Light therapy was begun at
08:00 h for 30 min, and later moved to 07:45 h.
Sleep promptly phase advanced to 24:00–07:30 h,
and morning awakening became spontaneous,
although she maintained an alarm clock backup.
She showed gradual improvement to complete
remission over 7 weeks (Hamilton score ¼3), when
she was discharged. She continued with light +
tranylcypromine at home, but was not compliant
with light treatment. Like Ms. B, she would
experience temporary relapses when discontinuing
light and quick recovery after resumption.
Ms. D, age 43, married, was more difficult. She
was chronically depressed since age 16, with
gradual worsening, and met criteria for major
depression since age 37. She had a past history of
anorexia, panic disorder and agoraphobia. She had
attempted suicide and was considered an active
suicide risk. She was unresponsive to multiple drug
trials and bilateral ECT. Her MEQ score of 76
categorized her as a ‘‘definite morning type,’’ and
she would typically sleep from 21:00 to 05:00 h. She
had no major sleep complaints, although a baseline
sleep log (under tranylcypromine 120 mg) showed
brief awakenings in the early morning hours
followed by some additional sleep, and—on the
night before light therapy began—final awakening
at 02:00 h. Light therapy began at 05:30 h for
30 min, later extended to 60 min when there was
no improvement. After failure of the morning light
trial, she was switched to evening light at 22:00 h in
an attempt to counteract her early sleep pattern.
She then slept well from 23:00 to 06:00 h, but
remained depressed. Nevertheless, she liked the
light therapy and was allowed additional afternoon
sessions, which gave her increased energy. She took
a light box home.
In summary, three of four strongly refractory
inpatients have shown positive response to light
therapy augmentation of MAOI. Over several years
of the tranylcypromine trial, this was the first string
of three successive successes. Followed for a year,
the three successful patients maintained positive
response after hospital discharge. We must wonder,
what was the effective agent? The observations
that patients did not respond in the hospital to
tranylcypromine alone, showed temporary relapse
when they discontinued light therapy, and showed
prompt remission after resuming treatment, sug-
gests that the effective agent was light alone.
Nonetheless, the patients’ primary providers chose
conservatively to maintain the pharmacotherapy on
the chance that the global improvement resulted
from an interaction with light.
Bipolar depression
The depressive phase of bipolar disorder has been
one of psychiatry’s toughest nuts to crack. The
addition of antidepressant drugs to mood stabili-
zers often leaves the patient depressed, and on a
statistical basis, the response to two well-estab-
lished antidepressants, paroxetine and bupropion
(in combination with a mood stabilizer) was no
better than to monotherapy with the stabilizer.
23
While it has sometimes been said that light therapy
mimics antidepressant drug action, if the response
to light therapy were successful for bipolar depres-
sion, we might posit that we have a qualitatively
different modality.
Ms. E (patient of Milica Stefanovic, M.D.), bipolar
II with psychotic features, was a 30-year-old
mother with three previous hospitalizations trig-
gered by suicidal ideation and planning. She was re-
admitted during increasing depression following
discontinuation of nortriptyline because of leg-
picking behavior. She had became hypersomnic and
aphagic, could not care for her children, and
contemplated overdosing on lorazepam. She was
unresponsive to dose increments of escitalopram,
and had experienced worsening under multiple
antidepressant and antipsychotic trials. Her mood
lightened in the summer, when she received several
atypical antipsychotic drugs to control hypomania.
After nine unsuccessful ECT sessions and two
suicide attempts while on the unit, she began daily
10,000 lux light therapy at 06:30h (according to her
MEQ score, and incremented from 30 to 45 min),
while continuing ECT 3 times per week, with
olanzapine at bedtime. Suicidal ideation disap-
peared within 2 days. After 5 days, her affect was
significantly improved (and self-acknowledged),
appetite and eating normalized, and sleep was
stable without hypersomnia. Gradually, she became
socially active on the unit, cooperative, with
normal rate and volume of speech (but with slight
latency of response), normal mood with full range
of affect, mood-congruent, brighter, with appro-
priate smiling and laughter, and linear, nondelu-
sional thinking. She was given passes for family
visits and discharged after 3 weeks. Her discharge
note states, ‘‘It was felt that light therapy was an
ARTICLE IN PRESS
Evolution of light therapy 501
adequate antidepressant given that the patient has
shown mood fluctuations on antidepressant drug
therapy.’’ At home, she continued with light
therapy, olanzapine, lithium and lorazepam, but
was later switched to lamotrigine as the sole
medication. An attempt to discontinue light ther-
apy at home failed, with immediate recurrence of
hypersomnia and depressed mood that quickly
resolved after resumption of daily treatment.
Ms. F (patient of Carolyn Douglas, M.D.), was a
44-year-old highly placed professional who suffered
schizoaffective disorder, bipolar type, with chronic
entrenched delusions. She was able to maintain
work function and satisfactory mood with a
combination of antidepressant (escitalopram) and
antipsychotic (ziprasidone) medication. However,
she was hypersomnic (sleep duration 10–12 h) with
delayed sleep phase for many years, and could not
get to the office before 13:00 h. This embarrassed
and frustrated her, and she compensated by work-
ing late into the evening. The sedating effect of
ziprasidone facilitated (‘‘masked’’) sleep onset
between 23:00 and 01:00 h, but she could not wake
up before 10:30–11:00 h. Her wish was to be able to
awaken at 07:00 h, go for a jog (she was in fine
physical shape) and arrive for work at 09:30 h. Her
MEQ score indicated an ‘‘evening type,’’ for which
awakening with light therapy at 07:30 h would be
recommended. However, the masked, hypersomnic
sleep made this infeasible. We intervened in four
ways designed to facilitate a phase advance: (a)
bright light therapy at 10:45–11:15 h, followed by a
morning jog. The timing was slowly edged earlier
over several months as earlier waking became
possible, ultimately reaching 07:00 h. The patient
showed an immediate, sustained mood lift from
ratings of ‘‘normal’’ to nearly ‘‘best ever,’’ yet
short of hypomania. Nonetheless, it was difficult to
push wake-up earlier. (b) When wake-up reached
09:45 h, we added 0.2 mg controlled released
melatonin at 18:45 h (15 h earlier), designed to
reinforce and accelerate the phase advances.
24
The
15-h time difference was maintained as wake-up
time and light therapy were gradually shifted
earlier. (c) Likewise, when wake-up reached
09:45 h, we added 90-min dawn simulation at the
end of the sleep period, also designed to reinforce
and accelerate the phase advances. (d) After taking
melatonin in late afternoon or in the evening, the
patient wore blue-blocking eyeglass fit-overs
while working under fluorescent lighting in the office
(often until 22:00 h), designed to minimize the
countervailing phase-delaying influence of evening
light. After several months, the ziprasidone dose was
reduced, and the patient was able to maintain a
stable 8 h sleep duration from approximately
23:00–07:00 h. With depression now absent and with
high energy, the patient’s delusional thinking also
abated. Thinking she was possibly cured, the patient
asked to discontinue the chronotherapeutics regi-
men; her primary provider insisted otherwise, at
least until late spring, when outdoor jogs in sunlight
might maintain the benefit.
Mr. G (patient of Carolyn Douglas, M.D.), age 75,
lived in an elder care institution. He had experi-
enced severe, recurrent, agitated, and sometime
delusional depressions since age 18, interspersed
with irritable hypomanic episodes. For the last 7
years the depression was relentless. He was
unresponsive to numerous medication trials, and
although ECT had provided temporary benefit,
cardiac complications during treatment 3 years
earlier, which necessitated pacemaker/defibrilla-
tor placement, prevented resumption. He showed a
distinct pattern of diurnal variation related to
nighttime carbohydrate binges. He would fall
asleep at 21:00 h, awaken at 23:00 h for a binge,
feel better (‘‘almost normal’’) for 1–2 h, and
resume intermittent sleep with final awakening at
04:00 h. He was exhausted in the morning, and
would refuse to leave bed or have breakfast. His
phase-advanced sleep suggested use of evening
light, and I selected dusk simulation (300 lux from
21:00 to 23:00 h followed by a 90 min fade to
0.001 lux) over bright light, to avoid bright light-
induced agitation. The patient immediately re-
sponded with coalesced sleep starting during the
dusk fade, eliminating the nighttime binge. Within
a week he reported, ‘‘The [dusk simulation] is
slowly making me better.’’ He awakened at 06:00 h,
left bed, ate breakfast in the common room, and
for the first time spent time walking outdoors after
breakfast. Clinically, his anxiety continued to
lessen and he began to enjoy things (e.g., vivid
reminiscences of Marilyn Monroe), short of hypo-
mania. After several weeks, brief awakening
recurred after 2 h sleep and he resumed eating
‘‘a couple of cookies,’’ but sleep, mood and anxiety
remained improved. Despite two intervening hos-
pitalizations during bouts of aggressive and delu-
sional behavior, he has continued the lighting
regimen for nearly a year. He complained immedi-
ately when power failures aborted dusk presenta-
tion on three occasions. Although dusk + dawn
simulation for a patient with SAD was reported in a
case study many years ago,
10
this may be the first
clinical application of dusk-alone therapy. In gen-
eral, elderly patients given evening bright light
therapy have not shown convincing response or
adhered to the schedule.
25
Clinical trials of light therapy for bipolar de-
pression are in their infancy, but show promise.
ARTICLE IN PRESS
M. Terman502
Sit et al.
26
studied nine women with longstanding
nonseasonal bipolar I or II disorder in which mood
stabilizers controlled manic phases, but anti-
depressants did not relieve depressed phases. A
placebo lead-in with dim red light administered
upon awakening led to minor improvement, with no
clinically successful outcome. The first four pa-
tients then received 7000 lux morning bright light
therapy in a flexible dose-ranging protocol with
exposure durations of 7–30 min. One patient
showed a full, sustained response at 30 min, while
the three others experienced onset of disruptive
mixed states characterized by irritability, elevated
energy, infeasible multitasking, creativity, aggres-
sion, racing thoughts and pressured speech.
Light treatment had to be terminated. The time
of day of exposure was then switched to early
afternoon for five additional patients on the
surmise that mixed states could be avoided,
based on earlier observations of patients with
rapid cycling bipolar disorder.
27
Two of these
patients showed major improvement at 60 min,
two were unsuccessful, and one patient with
partial response at 45 min achieved remission when
switched to morning light at 37.5 min. The lesson
drawn from this study is that light is a potent
intervention in nonseasonal bipolar depression, and
we will need a more sensitive dosing strategy than
has been used for SAD. This is one of the evolving
applications of light therapy that surely can win
support for major, controlled clinical trials, perhaps
even to the point of meta-analytic confirmation of
effect.
Benedetti et al. have had remarkable success in
reversal of bipolar depression by combining lithium
with sleep deprivation and light therapy.
28,29
The
specific contribution of light exposure to this
effect, however, is unclear. Importantly, responders
with histories of drug resistance were significantly
more likely than the others to show relapse over a
9-month follow-up period. Maintenance of the light
therapy regimen after hospital discharge might
reduce or eliminate this risk of relapse.
Antepartum depression
Rarely is a new line of clinical research sparked
directly by critical comments in grant peer review,
but here is one happy example. As with standard
antidepressant trials, pregnant women routinely
have been excluded from SAD light therapy trials
out of caution for the fetus. In a politically (and
scientifically) correct critique, the protocol was
considered discriminatory toward women, espe-
cially as the side-effect profile for light therapy was
deemed innocuous. The criticism taken to heart, a
group of colleagues at Yale, Case Western Reserve
and Columbia sought the pilot data that would
justify a funded trial. Our rationale was that no
pharmacotherapy is specifically approved for ante-
partum depression, and a successful alternative to
medication would be welcome by many women and
their doctors.
In the initial group, 16 women with major
depression received 3–5 weeks of 10,000 lux
morning light therapy and showed orderly improve-
ment, more than half with Hamilton scores in-
dicative of remission.
30
The response was no
greater among women who reported histories of
seasonality and those who did not. Two women
reported nausea under light therapy (which is the
most common side effect in patients with SAD
31
),
ameliorated in one case by dose reduction to 45 min
light exposure.
On this promising basis, we applied for federal
funding of a controlled trial but were refused for
two primary reasons: (a) we had not demonstrated
superiority relative to placebo; and (b) since
standard antidepressants already were known to
relieve antepartum depression, it was not a priority
to establish a non-drug alternative. So, with seed
funding from a private foundation we conducted a
randomized, controlled pilot trial that compared
response to 7000 lux vs. 500 lux light in 10
women.
32
Although 5 weeks were insufficient to
demonstrate a significant effect, by 10 weeks—
which included dose adjustments—the superiority
of bright light was clear. Nonetheless, a second
application for federal funding failed for two
primary reasons: (a) the placebo light in the pilot
study may have been active (low dose vs. inert) and
we had not established a result for the proposed 50-
lux substitute; and (b) this work did not merit a
multicenter effort. By contrast, a parallel applica-
tion by Anna Wirz-Justice to the Swiss National
Science Foundation received a top-rated review
and a 5-year randomized, controlled clinical trial is
actually underway.
Beyond depression
A prescient 1998 collection was entitled Seasonal
Affective Disorder and Beyond: Light Treatment for
SAD and Non-SAD Conditions,
33
yet its novelty lay
mainly in extension to nonseasonal depressions
(e.g., premenstrual dysphoric disorder) and comor-
bid conditions (e.g., bulimia nervosa, insomnia).
In principle, circadian rhythm disturbance and
chronotherapeutics should not be restricted to
depressive illness, and there have been continual
ARTICLE IN PRESS
Evolution of light therapy 503
attempts to expand the purview, at least within the
psychiatric domain.
Attention deficit hyperactivity disorder
(ADHD)
As with earlier work on bulimia nervosa, which
began with observations of winter exacerbation,
34
light therapy for adult ADHD was initially construed
in conjunction with SAD and the circadian rhythm
phase delay characteristic in winter.
35
However, 29
patients were selected on the basis of ADHD
criteria regardless of seasonal mood pattern.
A 3-week open trial of morning bright light
therapy produced clinically meaningful, statisti-
cally significant improvement in core ADHD symp-
toms irrespective of depression status. Importantly,
increased score on the MEQ—a proxy measure of
circadian rhythm phase advance—was a strong
correlate of ADHD improvement, while depression
status was not.
The underlying circadian concept is reinforced
by direct measurement of endogenous melatonin
and sleep patterns in children with ADHD, ages
6–12, with or without chronic idiopathic sleep
onset insomnia.
36
Both melatonin and sleep onset
were approximately 1 h phase-delayed in the
children with insomnia. Although morning light
therapy for the insomniacs, with special attention
to ADHD symptoms, is an obvious next step, it
would be worthwhile also to test the children with
normal sleep. In SAD, for example, light-evoked
circadian rhythm phase advances appear salutary
even for patients without delayed wintertime
sleep.
7
Dementia
The disrupted rest–activity pattern that accompa-
nies dementia has long been a target of light
therapy studies, but results have been weak or
negative.
37
Not only is this disruption a major
reason for institutionalization, but the stability and
amplitude of actigraphic data are significantly
correlated with cognitive and functional capacity
and emotional state.
38
There are hints that further
work will pay off. In a recent controlled study of 46
patients with Alzheimer’s disease, 1 h of morning
light exposure (2500 lux and above) for 10 weeks
did not provide superior outcome to normal room
light exposure, though there was improvement in a
subgroup of patients with the most disrupted
rest–activity cycles.
39
Anna Wirz-Justice’s group
administered dusk and dawn simulation (white light
peaking at 200 lux or a dim red placebo) at the
bedside of 13 demented patients for 3 weeks.
40
Globally, there was no improvement in circadian
rhythm stability or amplitude. However, the dusk-
to-dawn signal advanced nocturnal sleep onset by
more than 1 h, with commensurate increases in
sleep duration and nocturnal quiescence.
A recent controlled study of 189 patients pre-
sented a stronger intervention—daylong bright light
exposure (1000 lux vs. 375 lux) coupled with
high-dose melatonin (2.5 mg) before bedtime in
both groups. The bright light group showed en-
hanced stability and amplitude of the rest–activity
rhythm, improved mood and reduced cognitive
deterioration.
41,42
By contrast, mood deteriorated
under the normal room light control, which might
suggest a risk of melatonin monotherapy in chron-
otherapy for dementia.
For light therapy to be more generally effective
in elderly dementia—or old age in general—we
may need: (a) substantially enhanced light expo-
sure protocols that compensate for reduced ocular
transmission and retinal sensitivity; (b) individua-
lized timing anchored to circadian rhythm phase;
and (c) strategies that enhance rhythmic ampli-
tude, above and beyond phase shifting.
Parkinson’s disease
Though it might be hypothesized that light therapy
would reduce the depressive symptoms often seen
in Parkinson’s disease, it would be surprising also to
obtain improvement in core motor symptoms, the
targets of dopamine replacement therapy. A 12-
patient case series, using late evening light
exposure of 1000–1500 lux, found a majority of
patients who, within 2 weeks, showed clinically
significant reduction in bradykinesia and rigidity
(but not tremor) in parallel with the antidepressant
effect.
43
The choice of evening light exposure was
based on the presumption of advanced melatonin
onset and peak phases, which are especially
prominent under dopamine replacement therapy.
44
Despite such phase advances, initial insomnia is
prevalent in these patients, and evening light
surprisingly served to reduce sleep onset latency.
Of note, under light therapy many patients were
able to sustain up to 50% dose reduction of
dopaminergic medication without loss of symptom
control, which offers promise for alleviating treat-
ment-emergent levadopa-related motor complica-
tions. Open questions are: (a) the effect of evening
light therapy on circadian rhythm phase; (b) the
comparative efficacy of morning light; and, of
course, (c) the outcome of a randomized, placebo-
controlled trial.
ARTICLE IN PRESS
M. Terman504
Toward integrated chronotherapeutics
The observation of remarkable, often instanta-
neous remissions from major depression after one
night of sleep deprivation
45
holds a key to a new
therapeutic strategy unmatched by medication.
The downside has been the often-instantaneous
relapse following the next night’s recovery sleep.
The challenge has been to maintain the improve-
ment, if only as a booster toward eventual
medication response. Promising strategies include
multiple sleep deprivations with interspersed re-
covery sleep,
28,46
earlier bedtime and wake-up
following sleep deprivation (sleep phase ad-
vance),
47
and initiation of daily morning light
therapy following sleep deprivation.
48
Such proto-
cols for depressed inpatients are under active
investigation at centers in Copenhagen, Irvine,
Milan and New York.
In a consensus formulation (Benedetti F, Wirz-
Justice A, Terman M, in preparation), we are
developing a chronotherapeutic combination
strategy for general clinical use, initially for
inpatients under close observation. Overnight
sleep deprivation (rechristened wake therapy),
morning light therapy and sleep schedule advances
are introduced sequentially over 1–2 weeks, de-
pending on the patient’s day-to-day response.
At entry, the patient is allowed to sleep at his or
her habitual hours, in contrast to enforced ward
schedules. Timing of the chronotherapies is an-
chored to habitual sleep time, the MEQ score or
both. Treatment begins with a night of wake
therapy followed by the first light therapy session.
A decision tree guides the sequence of procedures
based on remission, partial improvement, nonre-
sponse, or worsening or relapse following transient
remission. In cases of instantaneous remission,
light therapy is repeated each morning to maintain
the effect. With partial or no improvement,
bedtime the next day is 5 h earlier than normally,
with wake-up for the second half of the night
and continued light therapy in the morning.
The patient may receive a second night of
wake therapy followed by sleep beginning 3 h
before habitual bedtime. If necessary, the
patient receives a third night of wake therapy
followed by sleep beginning 1 h before habitual
bedtime. Light exposure duration is increased in
15 min steps every 3 days, up to 60 min, if response
continues to fall short of remission. The chron-
otherapeutics ensemble is compatible with con-
current standard medication; indeed, lithium
appears to magnify the success rate in patients
with bipolar depression.
49
The consensus group
is preparing guidelines for clinical administration,
to be published as Psychiatric Chronotherapeutics:
A Treatment Manual.
Conclusions
Depending on one’s research background or
experience in clinical practice, the finding that
daily doses of light exposure can rapidly and
profoundly alleviate mood, sleep and cognitive
disturbances will seem either remarkable or self-
evident. Contraindications for light therapy are
few, and its compatibility with medication and
other modes of psychiatric treatment augurs
well for its incorporation into standard practice.
However, successful administration of light
therapy requires grounding in the principles of
chronobiology, with individual tailoring of the
regimen.
Research agenda
Given the probable effectiveness of light
therapy for a range of depressive as well as
other neuropsychiatric disorders, we need to
determine whether:
1. The basic dosing and timing regimens
established for seasonal affective disorder
must be reconsidered for each new applica-
tion,
2. Specific circadian rhythm parameters (e.g.,
phase relative to local time or sleep)
characterize each disorder, and whether
these parameters respond specifically to
light therapy,
3. Light therapy should be used to comple-
ment the effect of medication regimens or
can serve to replace them.
Practice points
In order to meaningfully supervise light therapy:
1. Clinicians need to gauge its effect on
circadian rhythm phase re-adjustment.
2. Patients need to maintain daily records of
treatment sessions, sleep time, mood and
energy ratings, and medication doses and
timing.
3. The interplay with medication needs to be
continually monitored.
ARTICLE IN PRESS
Evolution of light therapy 505
References
1. Golden RN, Gaynes BN, Ekstrom RD, et al. The efficacy of
light therapy in the treatment of mood disorders: a review
and meta-analysis of the evidence. Am J Psychiatry
2005;162:656–62.
2. Eastman CI, Young MA, Fogg LF, et al. Bright light treatment
for winter depression: a placebo-controlled trial. Arch Gen
Psychiatry 1998;55:883–9.
3. Terman M, Terman JS, Ross DC. A controlled trial of timed
bright light and negative air ionization for treatment of
winter depression. Arch Gen Psychiatry 1998;55:875–82.
4. Lewy AJ, Bauer VK, Cutler NL, et al. Morning vs. evening
light treatment of patients with winter depression. Arch
Gen Psychiatry 1998;55:890–6.
5. Wirz-Justice A. Beginning to see the light. Arch Gen
Psychiatry 1998;55:861–2.
6. Lam RW, Gorman CP, Michalon M, et al. Multicenter,
placebo-controlled study of fluoxetine in SAD. Am J
Psychiatry 1995;152:1765–70.
7. Terman JS, Terman M, Lo ES, et al. Circadian time of
morning light administration and therapeutic response in
winter depression. Arch Gen Psychiatry 2001;58:69–75.
8. Horne JA, O
¨stberg O. A self-assessment questionnaire to
determine morningness-eveningness in human circadian
rhythms. Int J Chronobiol 1976;4:97–110.
9. Terman M, Terman JS. Light therapy for seasonal and
nonseasonal depression: efficacy, protocol, safety and side
effects. CNS Spectr 2005;10:647–63.
10. Terman M, Schlager D, Fairhurst S, et al. Dawn and dusk
simulation as a therapeutic intervention. Biol Psychiatry
1989;25:966–70.
11. Terman M, Terman JS. Controlled trial of naturalistic dawn
simulation and negative air ionization for seasonal affec-
tive disorder. Am J Psychiatry 2006;163:2126–33.
12. Modell JG, Rosenthal NE, Harriett AE, et al. Seasonal
affective disorder and its prevention by anticipatory
treatment with bupropion XL. Biol Psychiatry 2005;58:
658–67.
13. Tuunainen A, Kripke DF, Endo T. Light therapy for non-
seasonal depression. Cochrane Database Sys Rev 2005;2:
CD004050.
14. Benedetti F, Colombo C, Pontiggia A, et al. Morning light
treatment hastens the antidepressant effect of citalopram: a
placebo-controlled trial. J Clin Psychiatry 2003;64:648–53.
15. Martiny K. Adjunctive bright light in non-seasonal major
depression. Acta Psychiatr Scand Suppl 2004;425:7–28.
16. Terman M. Review: light therapy is an effective treatment
for seasonal affective disorder. Evid Based Ment Health
2006;9:21.
17. Wirz-Justice A, Benedetti F, Berger M, et al. Chronother-
apeutics (light and wake therapy) in affective disorders.
Psychol Med 2005;25:939–44.
18. Wirz-Justice A, Terman M, Oren DA, et al. Brightening
depression [letter]. Science 2004;303:467–9.
19. Terman M. On the question of mechanism in phototherapy
for seasonal affective disorder: considerations of clinical
efficacy and epidemiology. J Biol Rhythms 1988;3:155–72.
20. Rosenthal NE, Bradt GH, Wehr TA. Seasonal Pattern
Assessment Questionnaire. Bethesda, MD: National Insti-
tute of Mental Health; 1987.
21. Goel N, Terman M, Terman JS, et al. Controlled trial of
bright light and negative air ions for chronic depression.
Psychol Med 2005;35:945–55.
22. Stewart JW, Terman M, Mangurian C, et al. Light therapy
augmentation of MAOI treatment in refractory depression.
Chronobiol Int 2005;23:740–1.
23. Sachs GS, Nierenberg AA, Calabrese JR, et al. Effectiveness
of adjunctive antidepressant treatment for bipolar depres-
sion. N Engl J Med 2007 Mar 28 [Epub ahead of print].
24. Lewy AJ, Lefler BJ, Emens JS, et al. The circadian basis of
winter depression. Proc Natl Acad Sci USA 2006;103:
7414–9.
25. Campbell SS, Terman M, Lewy AJ, et al. Light treatment for
sleep disorders: consensus report. V. Age-related distur-
bances. J Biol Rhythms 1995;10:151–4.
26. Sit D, Wisner KL, Hanusa BH, et al. Light therapy for bipolar
disorder: a case series in women. Bipolar Disord, in press.
27. Leibenluft E, Turner EH, Feldman-Naim S, et al. Light therapy
in patients with rapid cycling bipolar disorder: preliminary
results. Psychopharmacol Bull 1995;31:707–10.
28. Benedetti F, Barbini B, Fulgosi MC, et al. Combined total
sleep deprivation and light therapy in the treatment of
drug-resistant bipolar depression: acute response and long-
term remission rates. J Clin Psychiatry 2005;66:1535–40.
29. Benedetti F, Barbini B, Colombo C, et al. Chronotherapeu-
tics in a psychiatric ward. Sleep Med Rev 2007;11:509–22.
30. Oren DA, Wisner KL, Spinelli M, et al. An open trial of
morning light therapy for treatment of antepartum
depression. Am J Psychiatry 2002;159:666–9.
31. Terman M, Terman JS. Bright light therapy: side effects and
benefits across the symptom spectrum. J Clin Psychiatry
1999;60:799–808.
32. Epperson CN, Terman M, Terman JS, et al. Randomized
clinical trial of bright light therapy for antepartum depres-
sion: preliminary findings. J Clin Psychiatry 2004;65:421–5.
33. Lam RW, editor. Seasonal affective disorder and beyond:
light treatment of SAD and non-SAD conditions. Washing-
ton, DC: American Psychiatric Press; 1998.
34. Lam RW, Goldner EM, Solyom L, et al. A controlled study of
light therapy for bulimia nervosa. Am J Psychiatry 1994;
151:744–50.
35. Rybak YE, McNeely HE, Mackenzie BE, et al. An open trial of
light therapy in adult attention-deficit/hyperactivity dis-
order. J Clin Psychiatry 2006;67:1527–35.
36. van der Heijden KB, Smits MG, van Someren EJ, et al.
Effect of melatonin on sleep, behavior, and cognition in
ADHD and chronic sleep-onset insomnia. J Am Acad Child
Adolesc Psychiatry 2007;46:233–41.
37. Ancoli-Israel S, Martin JL, Kripke DF, et al. Effect of light
treatment on sleep and circadian rhythms in demented
nursing home patients. J Am Geriatr Soc 2002;50:282–9.
38. Carvalho-Bos SS, Riemersma-van der Lek RF, Waterhouse J,
et al. Strong association of the rest-activity rhythm with
well-being in demented elderly women. Am J Geriatr
Psychiatry 2007;15:92–100.
39. Dowling GA, Hubbard EM, Mastick J, et al. Effect of
morning bright light treatment for rest-activity disruption
in institutionalized patients with severe Alzheimer’s
disease. Int Psychogeriatr 2005;17:221–36.
40. Fontana-Gasio P, Kra¨uchi K, Cajochen C, et al. Dawn–dusk
simulation light therapy of disturbed circadian rest–
activity cycles in demented elderly. Exp Gerontol 2003;
38:207–16.
41. Riemersma RF, Swaab DF, Hoogendijk WJG, et al. Light and
melatonin: effect on sleep, mood and cognition in
demented elderly. Neurobiol Aging 2004;25(S2):194.
42. van Someren EJ, Riemersma-van Der Lek RF. Live to the
rhythm, slave to the rhythm. Sleep Med Rev 2007;11:
465–84.
ARTICLE IN PRESS
The most important references are denoted by an asterisk.
M. Terman506
43. Willis GL, Turner EJD. Primary and secondary features of
Parkinson’s disease improve with strategic exposeren to bright
light: a case series study. Chronobiol Int 2007;24:521–37.
44. Bordet R, Devos D, Brique S, et al. Study of circadian
melatonin secretion pattern at different stages of Parkin-
son’s disease. Clin Neuropharmacol 2003;26:65–72.
45. Wirz-Justice A, van den Hoofdakker RH. Sleep deprivation
in depression: what do we know, where do we go? Biol
Psychiatry 1999;46:445–53.
46. Kuhs H, Farber D, Borgstadt S, et al. Amitriptyline in
combination with repeated late sleep deprivation versus
amitriptyline alone in major depression. A randomised
study. J Affect Disord 1996;37:31–41.
47. Voderholzer U, Valerius G, Schaerer L, et al. Is the
antidepressive effect of sleep deprivation stabilized
by a three day phase advance of the sleep period? A
pilot study. Eur Arch Psychiatry Clin Neurosci 2003;253:
68–72.
48. Neumeister A, Goessler R, Lucht M, et al. Bright light
therapy stabilizes the antidepressant effect of partial sleep
deprivation. Biol Psychiatry 1996;39:16–21.
49. Benedetti F, Barbini B, Campori E, et al. Sleep phase
advance and lithium to sustain the antidepressant effect of
total sleep deprivation in bipolar depression: new findings
supporting the internal coincidence model? J Psychiatr Res
2001;35:323–9.
ARTICLE IN PRESS
Evolution of light therapy 507
