The efficacy of bright light treatment for depressive symptoms

(1)

2015

E.M. Dabroek | Student ID: 0914428 Master Thesis Clinical Psychology Supervisor: Dr. M. Molendijk Institute of Psychology Universiteit Leiden

22-‐5-‐2015

The efficacy of bright light

(2)

Abstract

Objective: In this meta-analysis the evidence base for the efficacy of Bright Light Therapy

(BLT) in treating depressive symptoms was assessed.

Method: Included were all studies with any outcome scale for depressive symptoms and with

BLT as an intervention. Both symptom severity and seasonality, study design and adjunctive treatment were taken into account. We calculated Hedges’s g, heterogeneity and publication bias on the efficacy of BLT. Additionally, the Pearson correlation coefficients r between BLT efficacy and characteristics of the interventions and the participants were calculated.

Results: For randomised controlled trials g = 0.58 (0.77, 0.40), for crossover studies g =

-0.92 (-1.28, -0.56) and for pre-post studies g = -1.45 (-1.59, -1.30). The effect of BLT is larger for seasonal depressive symptoms: g = -1.24 (-0.99, -0.79), than for non-seasonal depressive symptoms: g = -0.60 (-0.73, -0.47). When BLT combined with medication, sleep deprivation or Cognitive Behavioural Therapy, was compared to BLT alone, g ranged from -0.72 to -1.14, all significant p < 0.01. Having more severe symptoms at baseline was associated with a larger efficacy of BLT (r = -0.28, p < 0.01).

Discussion: The efficacy of BLT on depressive symptoms is statistically significant and of

mediate to large effect size. Study design, seasonality of symptoms, and combination with other therapies each affect the effect size to some extent, particularly seasonality. The efficacy of BLT depends mainly on characteristics of the participants (e.g., initial severity and

(4)

1. Introduction

1.1 Bright light therapy

About thirty years ago Rosenthal conducted the first controlled trial of bright light therapy for patients with Seasonal Affective Disorder, or SAD (see 1.2 Depressive disorders and

symptoms, Rosenthal et al., 1984). The use of light as a form of treatment for depressive

disorders has been studied extensively since then. ‘Bright Light Therapy’ (BLT), also known as ‘phototherapy’, has become a popular form of treatment. This is not surprising, given that this treatment has few side effects and is relatively simple to administer (Kogan & Guilford, 1998). Although BLT is administered in a variety of ways, the guidelines for this therapy consist of daily exposure to an artificial bright light source of approximately 10,000 lux for about half an hour per day. The light source can be a light box that a patient sits in front of or it can be a headpiece that directs the light downward at the eyes of the patient. The patient is not required to keep staring at the light, but rather perform certain relaxing activities while glancing at the light every one or two minutes (Sleemi et al., 2012; Management of Seasonal Affective Disorder, 2009). Another form of BLT, called “dawn simulation” was developed by Terman et al. (1989) for the treatment of SAD and was found to be superior to placebo (Avery et al., 2001; Golden et al., 2005). While the patient is asleep, the patient is exposed to a light source with a lower light intensity than in regular BLT. The light source gradually increases in illuminance during the night. This way, patients do not have the inconvenience of finding time to sit in front of the bright light (Terman et al. 1989; Avery et al., 2001).

Several aspects of BLT have been studied over the last decades. When morning administration was compared to evening administration, after three weeks of treatment morning BLT appeared to have a slight advantage over evening administration. However, after four weeks of treatment, no difference in effectiveness of BLT was found between the two groups (Eastman, Young, Fogg, Liu & Meaden, 1998). In a review of Tuunainen, Kripke, and Endo (2009), morning administration did appear to be more effective than administration at other times of the day. The colour of the light source also influences the effectiveness of BLT. Although white light is used most often in studies, some studies suggest that blue-enriched light produces a greater effect, but also comes with greater risk on side effects, such as eye damage (Gagné, Lévesque, Gagné, & Hébert, 2011). Another risk mentioned is that BLT may cause manic or hypomanic episode in susceptible patients with bipolar depression

(5)

(see 1.2 Depressive disorders and symptoms, Chan, Lam, & Perry, 1994). However, a more recent study did not find this side effect in patients with bipolar depression and this patient group benefitted from BLT as well (Dauphinais et al., 2012). Overall, BLT is regarded as a very safe treatment for depression. The most common side effects occur early in treatment and are headaches and eye or vision problems, but these are mild, transient symptoms that do not interfere with treatment (Kogan & Guilford, 1998).

BLT may also be combined with other existing treatments for depression, such as antidepressants, exercise, sleep deprivation therapy (also called wake therapy), or cognitive behavioural therapy (CBT) (Beauchemin, & Hays, 1997; Cooney et al. 2013; Fritszsche, Heller, Hill, & Kick, 2001; Gradisar et al., 2011; Martiny et al., 2012). A number of studies suggest some efficacy for BLT alone or BLT in adjunct to these therapies. An advantage of BLT over most antidepressants is that is produces faster results. Within a week of treatment with BLT, more than half of the patients report remission of their depression (Terman & Terman, 2005). This in contrast to antidepressant medication, where it often takes 4 to 6 weeks before an antidepressant effect can be observed, (see 1.2 Depressive disorders and

symptoms, Martiny et al., 2012). Lam et al. (2006), compared BLT to the antidepressant

‘fluoxetine’, and found similar response and remission rates for both therapies. BLT had a superior effect to fluoxetine after the first week of treatment, suggesting again that it renders results faster than this antidepressant. In addition, although both treatments were generally well tolerated, fluoxetine showed a higher rate of some adverse events, such as sleep

disturbance, agitation and palpitations. BLT has also been combined with Sleep Deprivation (SD) or wake therapy. Martiny et al. (2013) combined SD, BLT and sleep time stabilisation and compared the efficacy of this combined therapy to that of exercise therapy. SD with BLT had an immediate, large, stable and significantly better effect than exercise therapy.

Pinachsov, Shurgaja, Grischin and Putilov (2000), however, compared exercise to BLT for patients with non-seasonal depression and found a significant therapeutic difference in favour of exercise therapy (see 1.2 Depressive disorders and symptoms). Although Cognitive

Behavioural Therapy (CBT) is used for the treatment of depression, not many studies have studied the combination of CBT with BLT. A study with adolescents studied a combination of both therapies to treat delayed sleep phase disorder, and found a positive effect in treating the patients with this combination of therapies (Gradisar et al., 2011).

(6)

1.2 Depressive disorders and symptoms

Although BLT has been studied in a variety of groups and disorders, this paper will focus on studies that include disorders with depressive symptoms only. Depressive symptoms, as defined by the Diagnostic and Statistical Manual of mental disorders (2000) fourth edition, text revised (DSM-IV-TR, 4th ed. rev.) for Major Depressive Disorder (MDD), are a depressed mood, decreased interest in activities, fatigue, feelings of guilt or worthlessness, thoughts about suicide, trouble concentrating, unintended change in appetite or weight, in sleep (insomnia or hypersomnia) and in activity. In bipolar disorder these depressed episodes are alternated with manic episodes (a distinctive elevated or irritable mood for more than a week). Bromet et al. (2011) found a global lifetime prevalence of MDD of 11.1 to 14.6%, with twice as many women being diagnosed than men. Among patients with depressive disorders, BLT has been especially closely studied in patients with SAD, a depressive

disorder that shows a seasonal pattern. During the winter these patients suffer from depressive symptoms such as a low mood, decrease in interest and lack of concentration. During the summer patients experience less (severe) symptoms of depression or the depression goes into remission (Praschak-Rieder, & Willeit, 2003). Depressive disorders that do not show this seasonal pattern are referred to in this article as ‘non-seasonal’ depressive disorders (non-SAD). Subsyndromal Seasonal Affective Disorder (SSAD) is a milder form of SAD

(reference). SAD is an example of an illness in which chronobiological disturbances are a key factor. Although much is still unclear, the relationship between mood and the effect of light therapy seems to be mediated by the hormone melatonin. This hormone regulates the sleep and wake pattern and plays a central role in the internal clock. When a person is exposed to bright light in the morning, the onset of melatonin secretion will occur earlier than in a person who is exposed to a placebo dim light (Gabel et al., 2013; Pail et al., 2011). Exposure to bright light in the evening, acutely suppresses and reduces melatonin secretion and result in a later timing of the circadian clock (Chang, Aeschbach, Duffy, & Czeisler, 2014). More research is needed to further explain the mechanism behind BLT.

1.3 BLT compared with other therapies

In the last three decades BLT has been studied extensively. Since then BLT is established as an effective treatment for SAD. Studies on BLT now aim to refine and improve this form of treatment. The safety, colour, intensity, duration and distance to the light source have become an important focus of study. Other patient groups have been researched since, such as patients

(7)

with non-seasonal depressive disorders, elderly depressed patients, children and adolescents, pregnant women, patients with eating-, personality- and cognitive disorders and patients with physical illnesses, such as cancer or cardiovascular diseases. In 2005, Golden et al. (2005) reviewed 20 randomised, controlled studies applying BLT and dawn simulation. Their study showed that BLT has a significant effect in the treatment of non-SAD. BLT combined with antidepressant pharmacotherapy for patients with these disorders did not show any effect at all. Three years later Even, Schröder, Friedman and Rouillon (2008) reviewed 15 randomised, controlled studies about BLT in depressive disorders. They found that BLT has a higher efficacy in patients with SAD, compared to patients with non-SAD. When they compared BLT with a placebo (a low-intensity light condition or an air ionization device) in patients with non-SAD, they found inconsistent results; three studies found positive results for BLT alone and four studies found null results. Then, in non-SAD patients receiving

antidepressants, they compared the efficacy of BLT with a placebo light condition (dim light). They found positive results for the BLT as an adjunct to antidepressant pharmacotherapy condition. These results do not completely align with those of Golden et al. in 2005. Thus it remains unclear how much efficacy light therapy has for the treatment of patients with non-SAD. The final part of this section becomes a bit more difficult to read.

1.4 Study design

The study of Golden et al. in 2005 and of Even et al. in 2008 were limited to randomised, controlled studies (RCT’s). Using only RCT’s in a meta-analysis minimises biases caused by the placebo effect or confounding variables, but the results cannot always be generalised to all patients with depressive symptoms. Both the studies of Even et al. (2008) and Golden et al. (2005) mentioned that the placebo conditions are a limitation of their study. The issue of blindness is often poorly controlled. For some patients it is probably easy to guess that they have been assigned to a placebo condition when they are exposed to a device that only delivers a few hundred lux or less, or that they have been assigned to the active condition when they receive thousands lux per session. It is clearly very difficult to determine a placebo condition for this type of treatment, considering that in some studies the lux/time delivered in the ‘placebo condition’ was higher than in the ‘active condition’ in other studies (Even et al., 2008; Golden et al., 2005). A meta-analysis by Lee, Chad, Paterson, Janzen and Blashko (1997), that focussed on light spectrum of BLT, showed that red light (long wavelengths) were relatively ineffective compared to light of short to medium wavelengths (blue, green and

(8)

yellow light), making red light a suitable placebo condition. The short to medium

wavelengths seem to be essential for the therapeutic effect of light on SAD. A negative ion generator or red dim light are examples of control conditions that have been used in studies (Eastman, 1990; Eastman et al., 1998; Fritzsche, 2001; Goel, Terman, Terman, Macchi, & Stewart, 2005; Michalak et al., 2007). Given that the results that RCT’s report may not be generalised to all patients with depressive symptoms treated with BLT, this meta-analyses is not limited to this study design only, but will also include other study designs. It is important to compare results across designs, such as observational, crossover, RCT and

quasi-experimental pre-post test only – further referred to as ‘pre-post’ designs (Kazdin, A.E., 2010, p. 170). In the present paper, not only designs are compared, but also characteristics of the symptoms, other treatments for depressive symptoms and BLT, and the severity of the symptoms will be compared in this study. In doing so, this study aims to give a more complete description of the efficacy BLT has in the treatment of patients suffering from depressive symptoms. This leads us to the research questions and hypotheses of our study.

1.5 Research question: clinical efficacy of BLT in the treatment of depressive symptoms

The aim of this study is to assess the evidence base for the efficacy of bright light therapy in treating any type of disorder with depressive symptoms. In order to give an answer to this question, we will research the following sub questions:

What efficacy does BLT have, when different study designs are compared (RCT, crossover and pre-post)?

The studies of Even et al. (2008) and Golden et al. (2005) both only included studies with a randomised, controlled (RCT) design. Despite the issue of blindness in a RCT study design concerning BLT, the effect of BLT is measured under ‘ideal’ (controlled) circumstances, whereas with pre-post studies the effect is measured in more naturalistic, ‘real life’ circumstances and can also be due to other factors than BLT. So the latter design fails to eliminate most threats to internal validity (Leary, M.R., 2008, p.287). A higher effect size is expected in pre-post studies compared to RCT studies, because of this. A crossover study is more controlled than pre-post studies, but the effect in this study design may also because of a stacked effect of both conditions. A higher effect size is expected in crossover studies

(9)

What efficacy does BLT have in the treatment of non-seasonal depressive disorders?

The efficacy of BLT in SAD has been studied more extensively than in non-SAD. Golden et al. (2005) found a large efficacy in both patients with SAD and non-SAD. Even et. al (2008) found a higher efficacy in participants with SAD. Since the effect of BLT seems to be mediated by chronobiological factors, which are disturbed in SAD, a higher efficacy is expected in SAD (Pail et al., 2011).

What efficacy does BLT alone have compared to BLT as an adjunct to antidepressants, SD or CBT?

Research that studied the efficacy of BLT as an adjunct to these forms of therapy have yielded inconsistent results in patients with non-SAD. It is expected that BLT is effective to treat depressive symptoms on itself and combined with other therapies, since in this meta-analysis also patients with SAD are taken into account.

Does BLT show a higher efficacy in patients with more severe depressive symptoms compared to patient with less initial severity when entering the study?

As mentioned before, a mediator of the effect of BLT seems to be the hormone melatonin. The regulation of the circadian clock is regulated by this hormone. Because of this, the

amount of light that affects a patient positively is limited. If a person receives too much bright light this will affect melatonin secretion and sleep quality negatively (Chang et al., 2014). So, it is expected that patients with severe depressive symptoms benefit more from BLT than patients with mild depressive symptoms, because they have more room for improvement before they reach the this limited amount of light that affects a patient positively.

(10)

2. Methods

2.1 Study selection

All studies that cover any type of BLT and any disorder with depressive symptoms will be included in this research. We started the research with a systematic search in the database of Pubmed. To select eligible studies for our analysis, the search contained one of the following search terms: “seasonal affective disorder”, “SAD”, “depression”, “depressive”,

“non-seasonal depression”, “bipolar”, “mood disorder”, combined with one of the following search terms: “phototherapy”, “bright light therapy”, “BLT” and “Light treatment”. This search resulted in a total of 2,814 studies (see figure 1. for a visual representation of the study selection).

After excluding all studies that didn’t report original data (i.e. reviews, meta-analyses) and case reports, 415 studies remained. All study designs that reported original data (i.e. RCT, crossover studies and quasi-experimental pre-post test study designs) were included. For these remaining studies the following five inclusion criteria were defined for a further selection: the studies 1) were written in Dutch, English, German, Greek, Italian or Spanish, 2) only included human subjects, 3) used any form of BLT as an intervention, 4) use a scale to measure

depressive symptoms (i.e. Hamilton Depression Rating Scale, Structured Interview Guide for the Hamilton Depression Rating Scale – Seasonal Affective Disorders) as an outcome and 5) only included participants that were diagnosed with any type of depressive disorder. Using these criteria, another 137 studies were excluded, and an additional 6 duplicate studies were removed, leaving 272 studies for analysis. Of duplicate studies we chose to remove the studies with the smallest number of participants. Of the 272 studies, another 118 studies were removed, of which 58 studies because they did not report usable data to calculate effect sizes of BLT or respond to our attempts via e-mail to request the missing data, before we started the analysis. Studies were included when the missing information could be calculated from the reported data, or when the authors sent the data after being contacted by us. 27 of the 118 were removed because they did not study a disorder with depressive symptoms, did not use an outcome scale to measure depressive symptoms, or the intervention did not include any form of BLT. The last 33 studies were removed because of various reasons regarding usable data, used intervention, participants or use of BLT or a combination of mentioned reasons. After this selection 154 studies were included in the quantitive synthesis of this research.

(11)

Figure 1. Flow diagram of study selection for meta-analysis

2.2 Overview of studies and their characteristics

We extracted from each study: the study design, type of control condition, if any was used, the duration of the follow up, if any, setting (home or facility, inpatient or outpatient), which

Full-text articles excluded, with reasons: - (n = 58); no usable data to calculate effect size and no response from the authors or no contact info

- (n = 27); diagnosis other than depression or treatment other than BLT or Dawn simulation - (n = 33); other reasons or combination of previous mentioned reasons

Records identified (n = 415)

Full-text articles assessed for eligibility

(n = 278)

Records excluded (n = 137)

Records screened after duplicates removed (n = 272) Studies included in quantitative synthesis (meta-analysis)) (n = 154) Records excluded (n = 6) Id en ti fi cat ion Sc re en ing E li gi b il it y In cl u d ed S ear ch Systematic search in database Pubmed (n = 2,814) Records excluded (n = 2,399)

(12)

instruments and what type of outcome were used for diagnosis and outcome measurement, and country in which the study took place. To calculate the effect sizes we extracted the baseline and outcome values on the depression scale with their standard deviations, along with the number of patients in the active treatment condition, in the control condition and the total number of studied patients. We also extracted data on the mean age of the population under study, the percentage female and the number of dropouts. We noted the diagnosis to determine seasonality of the symptoms, and if applicable any comorbid diagnosis, what individual prescribed treatment the participants received currently or in the past and, if applicable, any additional treatment as intervention during the study period and its duration. Lastly we extracted data on the characteristics of the intervention: number, frequency,

duration, light intensity (lux/hour), and light spectrum of BLT sessions, and the total duration of the treatment for both the active and control condition, if the study used one. Most of these variables were usable for data-analysis with our software, described in 2.4 Data analysis and

software.

2.3 Statistical analysis

To assess the overall efficacy of BLT on depressive symptoms, we performed four different analyses to examine several aspects of the studies, BLT and the participants. First the three most common study designs that report original data are compared. These are RCT, crossover and pre-post studies. Secondly the studies are examined on the seasonality of depressive symptoms of the participants. Studies on SAD symptoms are compared with studies on non-SAD symptoms. The third comparison focusses on the type of intervention; BLT alone is compared with BLT & medication, BLT & sleep deprivation and BLT & cognitive

behavioural therapy. The fourth aspect that is examined is the initial severity of the depressive symptoms in the participants when they receive the BLT intervention. Examined is whether there is a correlation between the efficacy of BLT and the initial severity on the depressive outcome scale. We examine the correlations between characteristics of the participants and of BLT on the one hand and the weighted effect sizes of BLT on the other hand.

2.4 Data analysis and software

The software “Comprehensive meta analysis” was used to analyse the data. With the mean scores and standard deviations on the depressive symptom outcome scale and the sample sizes per treatment condition, Hedges’s g (pooled effect-size estimate) and its 95% confidence

(13)

interval was calculated for each individual study. With these the weighted mean of the effect sizes, and its 95% confidence interval, were calculated for every meta-analysis. To test the significance of each weighted mean effect size, nondirectionally at p < 0.05 level, the z test was used. Along with these weighted mean effect sizes, we calculated per study design the heterogeneity using I2 and the Q test, and publication bias using Egger’s t (both tested nondirectionally at p < 0.05). Finally, the correlation analysis was performed, using the program “IBM SPSS Statistics 22”. The correlation analyses were run on the data from all studies and separately for the three study designs; RCT, crossover and pre-post study design groups.

(14)

3. Results

3.1 Comparison of study designs

To see if the efficacy of BLT on depressive symptoms depends on the chosen study design, the three categories randomised controlled trials (RCT), crossover, and pre-post studies were compared. Five, non-randomised, controlled trials were added to the crossover group first, after a heterogeneity test revealed that these five studies did not differ significantly from the crossover studies (Q = 2.54, df = 1, p > 0.10), but did from the RCT studies (Q = 5.46, df = 1, p < 0.05). Table 1 shows the pooled effect-size estimates, heterogeneity and publication bias for each of the three study types.

Table 1. Pooled effect-size estimates, heterogeneity and publication bias by study design.

k N Hedges’s g (95% CI) Heterogeneity Publication bias

I2 _Q _{Egger’s t}

RCT 54 2,198 -0.58 (-0.77 – 0.40) *** 74.65% 209.07 *** 1.77

Crossover 21 492 -0.92 (-1.28 – 0.56) *** 68.02% 62.53 *** 2.51 * Pre-post 78 1,457 -1.45 (-1.59 – 1.30) *** 68.93% 247.85 *** 6.26 *** * Statistical significant at P < .05; *** Statistical significance at P < .001.

As this table shows, BLT is significant at the p < 0.001 level in all three study designs. Pre-post studies showed a higher effect-size estimate for BLT than both crossover and RCT studies did (Hedges’s g were respectively -1.45, -0.92 and -0.58). Heterogeneity tests revealed that the three groups differed from each other significantly in efficacy at the p < 0.001 level. However, with crossover studies, and especially with pre-post studies, the table also shows a significant publication bias, with Egger’s t = 2.51 being significant at the p < 0.05 level and t = 6.26 being significant at the p < 0.001 level respectively. I2 ranges from 68% to 75%.

3.2 Comparison of seasonality of depressive symptoms

Does the efficacy of BLT dependent on the seasonality of the symptoms? Table 2 shows the pooled effect-size estimates, heterogeneity and publication bias for SAD, non-SAD and mixed participant groups across all study designs. These results per study design can be found in appendix 1.

(15)

Table 2. Pooled effect-size estimates and heterogeneity by seasonality of the depressive

symptoms

k N Hedges’s g (95% CI) Heterogeneity

I2 _Q

SAD 96 2,623 -1.24 (-0.99 – 0.79) *** 78.73% 446.58 ***

Non-SAD 53 1,474 -0.60 (-0.73 – 0.47) *** 49.53% 103.03 ***

Mix 4 50 -1.65 (-2.50 – 0.79) *** 74.03% 11.55 **

** Statistical significance at P < .01; *** Statistical significance at P < .001.

Again the efficacy of BLT is significant at p < 0.001 level, in this case for both SAD and non-SAD symptoms. Notice that non-SAD and non-non-SAD differ from each other significantly with regard to efficacy, with a larger clinical effect for SAD compared to non-SAD (g = -1.24 vs. g = -0.60). For mixed groups Hedges´s g is the largest (g = -1.65), but this estimate is based on a small sample size (k = 4, N = 50).

3.3 Comparison of BLT alone to BLT combined with other therapies

What is the efficacy of BLT alone compared to a combination with other therapies, such as medication, sleep deprivation and cognitive behavioural therapy? The answers to this question are shown in table 3. Again, the results for this comparison per study design can be found in appendix 2.

(16)

Table 3. Pooled effect-size estimates, heterogeneity and publication bias by treatment for all

studies.

k N Hedges’s g (95% CI) Heterogeneity

I2 _Q

BLT alone 118 3,077 -1.09 (-1.22 – 0.96) *** 77.32% 515.96 *** BLT & medication 21 684 -0.72 (-0.97 – 0.46) *** 66.56% 59.81 *** BLT & SD 10 300 -0.84 (-1.36 – 0.32) ** 79.98% 44.96 ***

BLT & CBT 4 86 -1.14 (-1.36 – 0.32) ** 72.06% 10.74 *

* Statistical significant at P < .05; ** Statistical significance at P < .01; *** Statistical significance at P < .001.

Hedges’s g is significant at p < 0.001 level for BLT alone and BLT combined with

medication, being the largest for BLT alone (g = -1.09 vs. -0.72); p for difference is < 0.001. For BLT with SD and BLT with CBT g is significant at p < 0.01 level, being the largest for BLT with CBT (g = -1.14 vs. -0.84 for BLT with SD). All groups differ significantly from each other at p < 0.001 level in effect size, except for BLT combined with cognitive

behavioural therapy. The latter has an effect size significantly larger than in the other groups, at p < 0.05 level.

3.4 Correlation with characteristics of the studies, participants and BLT

In a correlation analysis the influence of several characteristics of the studies, participants and BLT interventions on the pooled effect-size estimates of Hedges’s g are examined. The results are shown in table 4. Total BLT (total received lux in therapy) was calculated by multiplying light intensity, number of sessions and duration per session. k represents the range of studies per study design that are correlated with Hedges’s g for each characteristic, as shown in the table.

(17)

Table 4. Pearson correlation of characteristics of the studies, participants and BLT

intervention with Hedges’s g

Pearson correlation with Hedges’s g RCT k = 45-54 Crossover k = 18-21 Pre-post k = 69-78 All k = 136-154 Participants Initial severity -0.312 * 0.343 -0.258 * -0.277 ** Age mean 0.178 -0.658 ** 0.128 0.082 Female percentage -0.063 0.086 -0.398 *** -0.191 * Bright light therapy

Total BLT -0.048 0.118 -0.058 0.005

Intensity -0.229 -0.067 -0.191 -0.121

Sessions -0.019 0.116 -0.029 0.137

Duration 0.155 0.081 0.210 0.129

* Statistical significant at P < .05; ** Statistical significance at P < .01; *** Statistical significance at P < .001.

More severe symptoms at baseline was associated with better efficacy (r = -0.277, p < 0.01). This correlation is found in both RCT studies and pre-post studies (r = -0.312 and -0.258, p < 0.05). A smaller number of females participating is associated with a better efficacy (r = -0.191, p < 0.05), but when the study designs are considered individually, this correlation is only found in pre-post studies (r = -0.398, p < 0.001). In crossover studies a correlation is found with age (p < 0.01), meaning that a younger age is associated with better efficacy.

Independent of study design, no correlation was found between effect size and any of the characteristics of BLT. However, a near significant trend was observed for number of sessions at p < 0.1. Among pre-post studies a similar trend was observed for intensity and duration with p < 0.1. This trend for intensity was also observed in RCT studies with p < 0.1. In table 5 the distributive statistics of these characteristics of BLT in all studies is shown.

(18)

Table 5. Median, interquartile range and range of BLT characteristics

Median Interquartile range Range

Total BLT (k = 136) 2,100,000 1,212,500 - 8,400,000 0-37,800,000

Intensity (k = 153) 2,500 2,187.5 - 10,000 0-10,000

Sessions (k = 152) 14 7 - 21 0-56

Duration (k = 152) 60 30 - 120 0-420

The median intensity of the used light in BLT (k = 153) is 2,500 lux, 50% of the intensity fell between 2,188 and 10,000 lux per hour. The median number of sessions was 14, 50% fell between 7 and 21 sessions. The median duration was 60 minutes, 50% fell between 30 and 120 minutes per session.

(19)

4. Discussion

4.1 Interpreting the results

How effective is BLT in treating depressive symptoms? We hypothesized that crossover studies would show a higher efficacy for BLT than RCT studies, but lower than pre-post studies (see 1.5 Research question: clinical efficacy in the treatment of depressive symptoms for the hypotheses). Pre-post studies match real life situations better than crossover, and especially RCT studies, but they fail to eliminate threats to the internal validity more. This results in other factors possibly contributing to a higher effect size. We found the effect size estimates to be statistically significant in all designs, but the differences between designs were also significant. The highest effect size was found in pre-post studies, the lowest in RCT studies, as was hypothesized. However, among crossover studies, and especially among pre-post studies, a significant publication bias was found, indicating that the found effect size for BLT in these groups is lower in reality.

When it comes to seasonality, BLT seems to be efficacious in the treatment of both SAD and non-SAD symptoms. It’s efficacy is significantly higher though for SAD symptoms, compared to non-SAD. This was expected since SAD shows a chronobiological disturbance, which seems to mediate the efficacy of BLT (Pail et al., 2011). BLT is also found to be more efficacious in patients with severe depressive symptoms, compared to patients with mild depressive symptoms. We argued before that patients with more severe symptoms have more room for improvement, than mildly depressed patients, before a ceiling effect limits the amount of BLT that is beneficial. Chang et al. (2014) showed that too much bright light can negatively affect sleep quality and melatonin secretion. However, another possible

explanation is that depression is caused by more factors than just the ones that are influenced by bright light. Phillips et al. (2015) describe several biomarkers and areas of the brain that are involved in depression. A combination of factors maintain depressive symptoms, of which only a few may be influenced by light. The other factors may maintain mild depressive

symptoms after receiving BLT.

In studies with a crossover design we did not find the association with the initial severity of the symptoms, but we did find that a younger age is associated with a better efficacy of BLT in this study design. The crossover design is different from the other designs in the sense that it has a control condition, but the participants are not blind to it, so

(20)

expectancies of the patients may have skewed the results. Young people are likely to have less experience with treatment, and thus have higher expectancies, than people who have tried other treatments before. Additionally, interaction effects between treatments may have occurred, since groups were not equal in sizes and not every possible sequence of treatments was administered (Kazdin, 2010). Besides younger age, we also found that in male patients BLT was more efficacious than in female patients. This may be the result of a selection bias, after all, the majority of studies in this research included mostly female participants.

However, there may be other reasons that explain this difference. Statistics do show a

difference in the prevalence among men and women when it comes to depressive symptoms, suggesting gender moderates the development and remission of depressive symptoms

(Bromet et al., 2011).

Surprisingly, no significant correlation was found between the efficacy of BLT and any of the characteristics of BLT that we examined in our study. These characteristics were the intensity of the light, the number of sessions, the duration of each session and the total amount of bright light received during the treatment. We did find a near significant correlation between the efficacy of BLT and the number of sessions. When we examined correlations per study design, we observed a near significant correlation again, with session duration in pre-post studies, and with light intensity in RCT and pre-post studies. When we examined the distribution of these characteristics, we found a small range for the duration and number of sessions. It’s possible that in these studies BLT was not administered long enough to find an association between the efficacy of BLT and duration, number of sessions or intensity. However, it is likely that the efficacy of BLT does not depend on the characteristics that we examined here. Another important characteristic seems to be timing of BLT. The efficacy of BLT appears to be mediated by the hormone that plays an essential role in the internal clock and that regulates the sleep and wake pattern (Pail et al., 2011). Secondly, exposure to bright light has shown to have a negative effect when it is administered within an hour before going to sleep (Chang et al, 2014). Another important characteristic is the colour of the light. Anderson, Glod, Dai, Cao and Lockley (2009) found blue-enriched

(white-appearing) light at a very low dose to be as effective as full spectrum white light, while Lee et al. (1997) found red light to be ineffective.

(21)

4.2 Previous research

The meta-analyses of Golden et al. (2005) and Even et al. (2008) found contradictive results. When they compared the seasonality of the symptoms, Golden et al. (2005) found a higher efficacy for BLT in non-SAD, compared to SAD, whereas Even et al. (2008) found a higher efficacy of BLT for SAD. The current meta-analysis found a positive result for the efficacy of BLT in both SAD and non-SAD. The results do seem to be more similar to those of Even et al. (2008), since BLT appeared to have a statistically significant higher efficacy for SAD than for non-SAD.

The combination of BLT with antidepressants has also been studied by both these meta-analyses. Even et al. (2008) found positive results for BLT combined with

antidepressants, compared to a placebo (dim light) condition, which Golden et al. (2005) did not find. Again, the current study shows more similar results to the study of Even et al. (2008). We found a positive result for BLT combined with antidepressants (and other therapies). The effect sizes of BLT, found in this analysis, were much higher than the effect sizes of several antidepressants, found by Llorca et al. (2014). Thus, BLT is a very effective alternative to antidepressants in the treatment of disorders with depressive symptoms, especially for those with SAD.

4.3 Shortcomings of this research

The aim of this meta-analysis was to assess the efficacy of BLT for depressive symptoms. It is not clear whether the efficacy for BLT found in this study is actually accounted for by the therapy itself and not by other variables. Surprisingly no correlations were found between the efficacy of BLT and the characteristics of BLT itself, in none of the study designs. More research is needed to explain the mechanisms behind BLT, and to identify the mediators of BLT. BLT works in a very controlled setting (RCT studies) but how well does it work in real life? The publication bias in pre-post studies makes it more difficult to give an exact answer to this question. Being aware of participating in the research, and the novelty aspect of this type of treatment may have contributed to the participants improving. In this meta-analysis we didn’t examine other characteristics of BLT, such as time of administration during the day, setting where participants received BLT (at home or at a facility, as inpatient or outpatient) or characteristics of the study, such as number of times that depressive symptoms were measured or the expectations of the participants before they entered the study. The number of times symptoms are measured and the setting influence how much attention is given to the patient

(22)

and to his or her suffering from the symptoms, which may influence the effect of therapy in general.

Another shortcoming is that there may have been a selection bias of studies, since a number of studies did not report data needed to calculate effect sizes. The authors were

emailed, but in many cases they did not e-mail back the required data before the analysis were started. This was especially the case for older studies, which also often lacked contact info. Based on language, a number of studies were also excluded, resulting in a selection of studies that mostly originated from western countries. Associated with this could be certain latitude, which may be associated to the amount of light people receive per day. We also found that most studies reported on SAD, MDD or Bipolar depression. Other disorders with depressive symptoms were underrepresented. As mentioned earlier, the majority of participants in most studies were female, possibly skewing the results of this meta-analysis. It is difficult to just generalise the results to all people with depressive symptoms for these reasons mentioned above.

4.4 Summary and recommendations

In summery this meta-analysis proved the efficacy for BLT as provided by RCT, crossover and pre-post studies for SAD and non-SAD symptoms. BLT was very effective in the

treatment of depressive symptoms, whether given as a treatment on its own, or in combination with other therapies. This study also shows that the efficacy of BLT depends mainly on characteristics of the studies and of the participants, and not on the characteristics of BLT that were examined in this meta-analysis. Still, much remains unclear about BLT and the

mechanisms behind it, which this study does not explain. More research on the mediators of BLT will also help to get a better understanding of which characteristics of the therapy are responsible for its effect. This may lead to improving this relatively new therapy. Secondly, the results from this research are very broad, and cannot be generalised to specific patient groups, especially regarding non-SAD. It is recommended that the efficacy of BLT will be studied for specific depressive disorders. BLT seems to be very effective, but can it be administered to all patients with depressive symptoms? Not only the efficacy, but also the risks are not clear yet. As mentioned earlier, studies are inconclusive about the risk BLT may have for triggering a manic episode in bipolar patients. On the other hand, BLT may be a safe alternative to medication for certain groups; pregnant women with depressive symptoms for

(23)

instance. More research on side effects in specific patient groups is needed for the

improvement and safe administration of this therapy in the future. Evidence for the efficacy of BLT is clearly present, but how does this efficacy relate to the efficacy of other therapies prescribed for depression currently?

(24)

References

American Psychiatric Association. (2000). Diagnostic and statistical manual of mental

disorders (4th ed., text rev.). Washington, DC: Author.

Anderson, J.L., Glod, C.A., Dai, J., Cao, Y., & Lockley, S.W. (2009). Lux vs. wavelength in light treatment of seasonal affective disorder. Acta Psychiatrica Scandinavica, 120, 203-212. doi: 10.1111/j.1600-0447.2009.01345.x.

Anderson, J.L., Vasile, R.G., Mooney, J.J., Bloomingdale, K.L., Samson, J.A.., &

Schildkraut, J.J. (1992). Changes in norepinephrine output following light therapy for fall/winter seasonal depression. Biological Psychiatry, 32, 700-704.

Aquino Miranda-Scippa, A.M. de, Nogueira Pires, M.L., Handfas, B.W., Marie, S.K.N., & Calil, H.M. (2008). Pituitary volume and the effects of phototherapy in patients with seasonal winter depression: a controlled study. Revista Brasileira de Psiquiatria,

30(1), 50-54.

Avery, D., Bolte, M.A., & Millet, M. (1992). Bright dawn simulation compared with bright morning light in the treatment of winter depression. Acta Psychiatrica Scandinavica,

85, 430-434.

Avery, D.H., Bolte, M.A., Dager, S.R., Wilson, L.G., Weyer, M., Cox, G.B., & Dunner, D.L. (1993). Dawn simulation treatment of winter depression: a controlled study. American

Journal of Psychiatry, 150, 113-117.

Avery, D.H., Bolte, M.A., Wolfson, J.K., & Kazaras, A.L. (1994). Dawn simulation compared with a dim red signal in the treatment of winter depression. Biological

Psychiatry, 36, 181-188.

Avery, D. H., Eder, D. N., Bolte, M. A., Hellekson, C. J., Dunner, D. L., Vitiello, M. V., & Prinz, P. N. (2001). Dawn simulation and bright light in the treatment of SAD: a controlled study. Biological Psychiatry, 50, 205-216.

(25)

Avery, D.H., Khan, A., Dager, S.R., Cohen, S., Cox, G.B., & Dunner, D.L. (1991). Morning or evening bright light treatment of winter depression? The significance of

hypersomnia. Biological Psychiatry, 29, 117-126.

Avery, D.H., Khan, A., Dager, S.R., Cox, G.B., Dunner, D.L. (1990). bright light treatment of winter depression: morning versus evening light. Acta Psychiatrica Scandinavica, 82, 335-338.

Avery, D.H., Kizer, D., Bolte, M.A., & Hellekson, C. (2001). Bright light therapy of subsyndromal seasonal affective disorder in the workplace: morning vs. afternoon exposure. Acta Psychiatrica Scandinavica, 103, 267-274.

Avissar, S., Schreiber, G., Nechamkin, Y., Neuhaus, I., Lam, G.K., Schwartz, P., Turner, E., Matthews, J., Naim, S., & Rosenthal, N.E. (1999). The effects of seasons and light therapy on G protein levels in mononuclear leukocytes of patients with seasonal affective disorder. Archives of General Psychiatry, 56, 178-183.

Bauer, M.S., Kurtz, J.W., Rubin, L.B., & Marcus, J.G. (1994). Mood and behavioral effects of four-week light treatment in winter depressives and controls. Journal of Psychiatric

Research, 28(2), 135-145.

Baumgartner, A., Volz, H.P., Campos-Barros, A., Stieglitz, R.D., Mansmann, U., & Mackert, A. (1996). Serum concentrations of thyroid hormones in patients with nonseasonal affective disorders treatment with bright and dim light. Biological Psychiatry, 40, 899-907.

Beauchemin, K. M., & Hays, P. (1997). Phototherapy is a useful adjunct in the treatment of depressed in-patients. Acta Psychiatrica Scandinavica, 95, 424-427.

Benedetti, F., Calabrese, G., Bernasconi, A., Cadioli, M., Colombo, C., Dallaspezia, S., Falini, A., Radaelli, D., Scotti, G., & Smeraldi, E. (2009). Spectroscopic correlates of antidepressant response to sleep deprivation and light therapy. A 3.0 Tesla study of bipolar depression. Psychiatric Research: Neuroimaging, 173, 238-242.

(26)

Benedetti, F., Colombo, C., Pontiggia, A., Bernasconi, A., Florita, M., & Smeraldi, E. (2003). Morning light treatment hastens the antidepressant effect of citalopram: a placebo-controlled trial. Journal of Clinical Psychiatry, 64, 648-653.

Benedetti, F., Dallaspezia, S., Fulgosi, M.C., Barbini, B., Colombo, C., & Smeraldi, E. (2007). Phase advance is an actimetric correlate of antidepressant response to sleep deprivation and light therapy in bipolar depression. Chronobiology International,

24(5), 921-937.

Bielski, R.J., Mayor, J., & Rice, J. (1992). Phototherapy with broad spectrum white fluorescent light: a comparative study. Psychiatry Research, 43, 167-175. Bromet, E., Andrade, L.H., Hwang, I., Sampson, N.A., Alonso, J., Girolamo, G. de, ...

Kessler, R.C. (2011). Cross-national epidemiology of DSM-IV major depressive episode. BioMed Central Medicine, 9, 90. doi:10.1186/1741-7015-9-90.

Brunner, D.P., Kräuchi, K., Dijk, D.J., Leonhardt, G., Haug, H.J., & Wirz-Justice, A. (1996). Sleep electroencephalogram in seasonal disorder and in control women: effects of midday light treatment and sleep deprivation. Biological Psychiatry, 40, 485-496. Chan, P. K., Lam, R. W., & Perry, K. F. (1994). Mania precipitated by light therapy for

patients with SAD. Journal of Clinical Psychiatry, 55(10), 454.

Chang, A.M., Aeschbach, D., Duffy, J.F., Czeisler, C.A. (2014). Evening use of light-emitting eReaders negatively affects sleep, circadian timing and next-morning alertness.

Proceedings of the National Academy of Sciences, 112(4), 1232-1237.

doi:10.1073/pnas.1418490112/-/DCSupplemental.

Cooney, G. M., Dwan, K., Greig, C. A., Lawlor, D. A., Rimer, J., Waugh, F. R., McMurdo, M., & Mead, G. E. (2013). Exercise for depression. The Cochrane Collaboration. doi: 10.1002/14651858.CD004366.pub6.

Corral, M., Wardrop, A.A., Zhang, H., Grewal, A.K., & Patton, S. (2007). Morning light therapy for postpartum depression. Archives of Women’s Mental Health, 10, 221-224. doi:10.1007/s00737-007-0200-1.

(27)

Dauphinais, D. R., Rosenthal, J. Z., Terman, M., DiFebo, H. M., Tuggle, C., & Rosenthal, N. E. (2012). Controlled trial of safety and efficacy of bright light therapy vs. negative air ions in patients with bipolar depression. Psychiatry Research, 196, 57-61.

doi:10.1016/j.psychres.2012.01.015

Desan, P.H., Weinstein, A.J., Michalak, E.E., Tam, E.M., Meesters, Y., Ruiter, M.J., Horn, E., Telner, J., Iskandar, H., Boivin, D.B., & Lam, R.W. (2007). A controlled trial of the Litebook light-emitting diode (LED) light therapy device for treatment of seasonal affective disorder (SAD). BMC Psychiatry, 7(38). doi:10.1186/1471-244X-7-38. Dietzel, M., Saletu, B., Lesch, O.M., Sieghart, W., & Schjerve, M. (1986). Light treatment in

depressive illness. Polysomnographic, psychometric and neuroendocrinological findings. European Neurology, 25(2), 93-103.

Doghramji, K., Gaddy, J.R., Stewart, K.T., Rosenthal N.E., & Brainard, G.C. (1990). 2- versus 4-hour evening phototherapy of seasonal affective disorder. The Journal of

Nervous and Mental Disease, 178(4), 257-260.

Eastman, C. I. (1990). What the placebo literature can tell us about light therapy for SAD.

Psychopharmacology Bulletin, 26, 495-504.

Eastman, C.I., Lahmeyer, H.W., Watell, L.G., Good, G.D., & Young, M.A. (1992). A placebo-controlled trial of light treatment for winter depression. Journal of Affective

Disorders, 26, 211-222.

Eastman, C. I., Young, M. A., Fogg, L. F., Liu, L., & Meaden, P. M. (1998). Bright light treatment of winter depression: a placebo controlled trial. Archives of General

Psychiatry, 55, 883-889.

Echizenya, M., Suda, H., Takeshima, M., Inomata, Y., & Shimizu, T. (2013). Total sleep deprivation followed by sleep phase advance and bright light therapy in drug-resistant mood disorders. Journal of Affective Disorders, 144, 28-33. doi:

(28)

Epperson, C.N., Terman, M., Terman, J.S., Hanusa, B.H., Oren, D.A., Peindl, K.S., & Wisner, K.L. (2004). Randomized clinical trial of bright light therapy for antepartum depression: preliminary findings. Journal of Clinical Psychiatry, 65, 421-425. Even, C., Schröder, C. M., Friedman, S., & Rouillon, F. (2008). Efficacy of light therapy in

nonseasonal depression: A systematic review. Journal of Affective Disorders, 108, 11-23.

Flory, R., Ametepe, J., & Bowers, B. (2010). A randomized, placebo-controlled trial of bright light and high-density negative air ions for treatment of seasonal affective disorder.

Psychiatry Research, 177, 101-108. doi: 10.1016/j.psychres.2008.08.011.

Fritszsche, M., Heller, R., Hill, H., & Kick, H. (2001). Sleep deprivation as a predictor of response to light therapy in major depression. Journal of Affective Disorders, 62, 207-215.

Gabel, V., Maire, M., Reichert, C. F., Chellappa, S. L., Schmidt, C., Hommes, V., Viola, A. U., & Cajochen, C. (2013). Effects of artificial dawn and morning blue light on daytime cognitive performance, well-being, cortisol and melatonin levels.

Chronobiology International, 30(8), 988-997. doi: 10.3109/07420528.2013.793196.

Gagné, A. M., Lévesque, F., Gagné, P., & Hébert, M. (2011). Impact of blue vs red light on retinal response of patients with seasonal affective disorder and healthy controls.

Progress in Neuro-Psychopharmacology & Biological Psychiatry, 35, 227-231.

Gallin, P.F., Terman, M., Remé, C.E., Rafferty, B., Terman, J.S., Burde, R.M. (1995). Ophthalmologic examination of patients with seasonal affective disorder, before and after bright light therapy. American Journal of Ophthalmology, 119, 202-210. Geerts, E., Bouhuys, N., Meesters, Y., & Jansen, J. (1995). Observed behavior of patients

with seasonal affective disorder and an interviewer predicts response to light treatment. Psychiatry Research, 57, 223-230.

Geerts, E., Kouwert, E., Bouhuys, N., Meesters, Y., & Jansen, J. (2000). Nonverbal

interpersonal attunement and extravert personality predict outcome of light treatment in seasonal affective disorder. Journal of Affective Disorders, 59, 193-204.

(29)

Glickman, G., Byrne, B., Pineda, C., Hauck, W.W., & Brainard, G.C. (2006). Light therapy for seasonal affective disorder with blue narrow-band light emitting diodes (LEDs).

Biological Psychiatry, 59, 502-507. doi: 10.1016/j.biopsych.2005.07.006.

Goel N., Terman, M. Terman, J. S., Macchi, M. M., & Stewart, A. W. (2005). Controlled trial of bright light and negative ions for chronic depression. Psychological Medicine, 35, 945-955. doi:10.1017/S0033291705005027.

Golden, R. N., Gaynes, B. N., Ekstrom, R. D., Hamer, R. M., Jacobsen, F. M., Suppes, T., Wisner, K. L., & Nemeroff, C. B. (2005). The efficacy of light therapy in the

treatment of mood disorders: A review and meta-analysis of the evidence. American

Journal of Psychiatry, 162, 656-662.

Gordijn, M.C.M., Mannetje, D. ‘t, & Meesters, Y. (2012). The effects of blue-enriched light treatment compared to standard light treatment in seasonal affective disorder. Journal

of Affective Disorder, 136, 72-80. doi: 10.1016/j.jad.2011.08.016.

Gradisar, M., Dohnt, H., Gardner, G., Paine, S., Starkey, K., Menne, A., Slater, A., Wright, H., Hudson, J.L., Weaver, E., & Trenowden, S. (2011). A randomized controlled trial of cognitive-behavior therapy plus bright light therapy for adolescent delayed sleep phase disorder. Sleep, 34(12), 1671-1680. doi: 10.5665/sleep.1432.

Grota, L.J., Yerevanian, B.I., Gupta, K., Kruse, J., & Zborowski, L. (1989). Phototherapy for seasonal major depressive disorder: effectiveness of bright light of high or low intensity. Psychiatry Research, 29, 29-35.

Hoekstra, R., Fekkes, D., Wetering, B.J.M. van de, Pepplinkhuizen, L., & Verhoeven, W.M.A. (2003). Effect of light therapy on biopterin, neopterin and tryptophan in patients with seasonal affective disorder. Psychiatry Reseach, 120, 37-42. doi: 10.1016/S0165-1781(03)00167-7.

Holsboer-Trachsler, E., Hemmeter, U., Hatzinger, M., Seifritz, E., Gerhard, U., & Hobi, V. (1994). Sleep deprivation and bright light as potential augmenters of antidepressant drug treatment – neurobiological and psychometric assessment of course. Journal of

(30)

Jacobsen, F.M., Mueller, E.A., Rosenthal, N.E., Rogers, S., Hill, J.L. & Murphy, D.L. (1994). Behavioral responses to intravenous meta-chlorophenyl-piperazine in patients with seasonal affective disorder and control subjects before and after phototherapy.

Psychiatry Research, 52, 181-197.

Janas-Kozik, M., Krzystanek, M., Stachowicz, M., Krupka-Matuszczyk, I., Janas, A., & Rybakowski, J.K. (2011). Bright light treatment of depressive symptoms in patients with restrictive type of anorexia nervosa. Journal of Affective Disorders, 130, 462-465.

Joffe, R.T., Moul, D.E., Lam, R.W., Levitt, A.J., Teicher, M.H., Lebegue, B., Oren, D.A., Buchanan, A., Glod, C.A., Murray, M.G., Brown, J., & Schwartz, P. (1993). Light visor treatment for seasonal affective disorder: a multicenter study. Psychiatry

Research, 46, 29-39.

Kazdin, A. E. (2010). Research Design in Clinical Psychology. Boston, MA: Pearson Education.

Kogan, A. O., & Guilford, P. M. (1998). Side effects of short-term 10.000 lux light therapy.

American Journal of Psychiatry, 155, 293-294.

Kripke, D.F., Craig Risch, S., & Janowsky, D. (1983). Bright white light alleviates depression. Psychiatry Research, 10, 105-112.

Labbate, L.A., Lafer, B., Thibault, A., Rosenbaum, J.F., & Sachs, G.S. (1995). Influence of phototherapy treatment duration for seasonal affective disorder: outcome at one vs. two weeks. Biological Psychiatry, 38, 747-750.

Lafer, B., Sachs, G.S., Labbate, L.A., Thibault, A., & Rosenbaum, J.F. (1994). Phototherapy for seasonal affective disorder: a blind comparison of three different schedules. The

Lam, R.W. (1994). Morning light therapy for winter depression: predictors of response. Acta

Psychiatrica Scandinavica, 89, 97-101.

Lam, R.W., Buchanan, A., Mador, J.A., & Corral, M.R. (1992). Hypersomnia and morning light therapy for winter depression. Biological Psychiatry, 31, 1062-1064.

(31)

Lam, R.W., Buchanan, A., Mador, J.A., Corral, M.R., Remick, R.A. (1992). The effects of ultraviolet-A wavelengths in light therapy for seasonal depression. Journal of Affective

Disorders, 24, 237-244.

Lam, R.W., Carter, D., Misri, S., Kuan, A.J., Yatham, L.N., & Zis, A.P. (1999). A controlled study of light therapy in women with late luteal phase dysphoric disorder. Psychiatry

Research, 86, 185-192.

Lam, R. W., Levitt, A. J., Levitan, R. D., Enns, M. W., Morehouse, R., Michalak, E. E., & Tam, E. M. (2006). The Can-SAD study: a randomized controlled trial of the

effectiveness of light therapy and fluoxetine in patients with winter seasonal affective disorder. American Journal of Psychiatry, 163, 805-812.

Lam, R.W., Tam, E.M., Shiah, I.S., Yatham, L.N., & Zis, A.P. (2000). Effects of light therapy on suicidal ideation in patients with winter depression. Journal of Clinical Psychiatry,

61, 30-32.

Leary, M.R. (2008). Introduction to behavioral research methods. Boston, MA: Pearson Education.

Lee, T. M. C., Chad, C. C. H., Paterson, J. G., Janzen, H., & Blashko, C. A. (1997). Spectral properties of phototherapy for seasonal affective disorder: a meta-analysis. Acta

Psychiatrica Scandinavica, 96, 117-121.

Levitt, A.J., Joffe, R.T., & King, E. (1994). Dim versus bright red (light-emitting diode) light in the treatment of seasonal affective disorder. Acta Psychiatrica Scandinavica, 89, 341-345.

Levitt, A.J., Lam, R.W., & Levitan, R. (2002). A comparison of open treatment of seasonal major and minor depression with light therapy. Journal of Affective Disorders, 71, 243-248.

Lewy, A.J., Bauer, V.K., Cutler, N.L., Sack, R.L., Ahmed, S., Thomas, K.H., Blood, M.L., & Latham Jackson, J.M. (1998). Morning vs evening light treatment of patients with winter depression. Archives of General Psychiatry, 55, 890-896.

(32)

Lewy, A.J., Sack, R.L., Miller, S., & Hoban, T.M. (1987). Antidepressants and circadian phase-shifting effects of light. Science, 235, 352-354.

Lieverse, R., Someren, E.J.W. van, Nielen, M.M.A., Uitdehaag, B.M.J., Smit, J.H., &

Hoogendijk, W.J.G. (2011). Bright light treatment in elderly patients with nonseasonal major depressive disorder. A randomized placebo-controlled trial. Archives of General

Psychiatry, 68(1), 61-70.

Lingjærde, O., Reichborn-Kjennerud, T., Haggag, A., Gärtner, I., Berg, E.M., & Narud, K. (1993). Treatment of winter depression in Norway. Short- and long-term effects of 1500-lux white light for 6 days. Acta Psychiatrica Scandinavica, 88, 292-299. Llorca, P.M., Lançon, C., Brignone, M., Rive, B., Salah, S., Ereshefsky, L., & Francois, C.

(2014). Relative efficacy and tolerability of vortioxetine versus selected

antidepressants by indirect comparisons of similar clinical studies. Current Medical

Research and Opinion 30(12), 2589-2606. doi:10.1185/03007995.2014.969566.

Loving, R.T., Kripke, D.F., & Shuchter, S.R. (2002). Bright light augments antidepressant effects of medication and wake therapy. Depression and Anxiety, 16, 1-3. doi: 10.1002/da.10036.

Loving, R.T., Kripke, D.F., Elliott, J.A., Knickerbocker, N.C., & Grandner, M.A. Bright light treatment of depression for older adults. BMC Psychiatry, 5(41). doi: 10.1186/1471-244X-5-41.

Loving, R.T., Kripke, D.F., Knickerbocker, N.C., & Grandner, M.A. (2005). Bright green light treatment of depression for older adults. BMC Psychiatry, 5(42).

doi:10.1186/1471-244X-5-42.

Mackert, A., Volz, H.P., Stieflitz, R.D., & Müller-Oerlinghausen, B. (1991). Phototherapy in nonseasonal depression. Biological Psychiatry, 30, 257-268.

Management of seasonal affective disorder. (2009). Drug Therapy Bulletin, 47, 128-132. Martinez, B., Kasper, S., Ruhrmann, S., & Möller, H.J. (1994). Hypericum in the treatment of

seasonal affective disorders. Journal of Geriatric Psychiatry and Neurology, 7(1), 29-33. doi: 10.1177/089198879400700109.

(33)

Martiny, K. (2004). Adjunctive bright light in non-seasonal major depression. Acta

Psychiatrica Scandinavica, 110(425), 7-28.

Martiny, K., Lunde, M., Undén, M., Dam, H., & Bech, P. Acta Psychiatrica Scandinavica,

112, 117-125. doi: 10.1111/j.1600-0447.2005.00574.x.

Martiny, K., Refsgaard, E., Lund, V., Lunde, M., Sørensen, L., Thougaard, B., Lindberg, L., & Bech, P. (2012). A 9-week randomized trial comparing a chronotherapeutic

intervention (wake and light therapy) to exercise in major depressive disorder patients treated with Duloxetine. Journal of Clinical Psychiatry, 73(9), 1234-1242. doi: 10.4088/JCP.11m07625.

Martiny, K., Refsgaard, E., Lund, V., Lunde, M., Sørensen, L., Thougaard, B., Lindberg, L., & Bech, P. (2013). The day-to-day acute effect of wake therapy in patients with major depression using the HAM-D as primary outcome measure: Results from a

randomised controlled trial. PLOS ONE, 8(6), doi: 10.1371/journal.pone.0067264 Matthew, E., Vasile, R.G., Sachs, G., Anderson, J., Lafer, B., & Hill, T. (1996). Regional

cerebral blood flow changes after light therapy in seasonal affective disorder. Nuclear

Medicine Communications, 17, 475-479.

McEnany, G.W., & Lee, K.A. (2005). Effects of light therapy on sleep, mood, and

temperature in women with nonseasonal major depression. Issues in Mental Health

Nursing, 26, 781-794. doi: 10.1080/01612840591008410.

McIntyre, I.M., Armstrong, S.M., Norman, T.R., & Burrows, G.D. (1989). Treatment of seasonal affective disorder with light: preliminary australian experience. Australian

and New Zealand Journal of Psychiatry, 23, 369-372.

Meesters, Y., Dekker, V., Schlangen, L.J.M., Bos, E.H., & Ruiter, M.J. (2011). Low-intensity blue-enriched white light (750 lux) and standard bright light (10 000 lux) are equally effective in treating SAD. A randomized controlled study. BMC Psychiatry, 11(17). doi: 10.1186/1471-244X-11-17.

(34)

Meesters, Y., Jansen, J.H.C., Beersma, D.G.M., Bouhuys, A.L., & Hoofdakker, R.H. van den. (1993). Early light treatment can prevent an emerging winter depression from

developing into a full-blown depression. Journal of Affective Disorders, 29, 41-47. Michalak, E. E., Murray, G., Levitt, A. J., Levitan, R. D., Enns, M. W., Morehouse, R., Tam,

W. M., Cheung, A., & Lam, R. W. (2007). Quality of life as an outcome indicator in patients with seasonal affective disorder: Results from the Can-SAD study.

Psychological Medicine, 37, 727-736. doi: 10.1017/S0033291706009378.

Moscovici, L., & Kotler, M. (2009). A multistage chronobiologic intervention for the treatment of depression: a pilot study. Journal of Affective Disorders, 116, 201-207. doi: 10.1016/j.jad.2009.01.015.

Müller, M.J., Seifritz, E., Hatzinger, M., Hemmeter, U., & Holsboer-Trachsler, E. (1997). Side effects of adjunct light therapy in patients with major depression. European

Archives of Psychiatry and Clinical Neuroscience, 247, 252-258.

Nagayama, H., Sasaki, M., Ichii, S., Hanada, K., Okawa, M., Ohta, T., Asano, Y., Sugita, Y, Yamazaki, J., Koshaka, M., Kotorii, T., Maeda, K., Okamoto, N., Ishizuka, Y., Takahashi, K., Honda, Y., & Takahashi, S. (1991). Atypical depressive symptoms possible predict responsiveness to phototherapy in seasonal affective disorder. Journal

of Affective Disorders, 23, 185-189.

Naus, T., Burger, A., Malkoc, A., Molendijk, M., & Haffmans, J. (2013). Is there a difference in clinical efficacy of bright light therapy for different types of depression? A pilot study. Journal of Affective Disorders, 151, 1135-1137. doi: 10.1016/j.jad.2013.07.017. Neumeister, A., Goessler, R., Lucht, M., Kapitany, T., Bamas, C., & Kasper, S. (1996). Bright light therapy stabilizes the antidepressant effect of partial sleep deprivation. Biological

Psychiatry, 39, 16-21.

Neumeister, A., Turner, E.H., Matthews, J.R., Postolache, T.T., Barnett, R.L., Rauh, M., Vetticad, R.G., Kasper, S., & Rosenthal, N.E. (1998). Effects of tryptophan depletion vs catecholamine depletion in patients with seasonal affective disorder in remission with light therapy. Archives of General Psychiatry, 55, 524-530.

(35)

Oren, D.A., Berman, R.M., Anand, A., & Charney, D.S. (2000). No effect of light on basal glucagon levels in winter seasonal depressives and comparison subjects. Psychiatry

Research, 94, 263-266.

Oren, D.A., Brainard, G.C., Johnston, S.H., Joseph-Vanderpool, J.R., Sorek, E., & Rosenthal, N.E. (1991). Treatment of seasonal affective disorder with green light and red light.

Oren, D.A., Desan, P.H., Boutros, N., Anand, A., & Charney, D.S. (2002). Effects of light on low nocturnal bilirubin in winter depression: a preliminary report. Biological

Psychiatry, 51, 422-425.

Oren, D.A., Wisner, K.L., Spinelli, M., Epperson, C.N., Peindl, K.S., Terman, J.S., &

Terman, M. (2002). An open trial of morning light therapy for treatment of antepartum depression. American Journal of Psychiatry, 159, 666-669.

Pail, G., Huf, W., Pjrek, E., Winkler, D., Willeit, M., Praschak-Rieder, N., & Kasper, S. (2011). Bright-light therapy in the treatment of mood disorders. Neuropsychobiology,

64, 152-162. doi: 10.1159/000328950.

Papatheodorou, G., Kutcher, S. (1995). The effect of adjunctive light therapy on ameliorating breakthrough depressive symptoms in adolescent-onset bipolar disorder. Journal of

Psychiatry & Neuroscience, 20(3), 226-232.

Parry, B.L., Berga, S.L., Mostofi, N., Sependa, P.A., Kripke, D.F., & Gillin, J.C. (1989). Morning versus evening bright light treatment of late luteal phase dysphoric disorder.

Partonen, T. (1994). Effects of morning light treatment on subjective sleepiness and mood in winter depression. Journal of Affective Disorders, 30, 99-108.

Partonen, T., Appelberg, B., & Partinen, M. (1993). Effects of light treatment on sleep

structure in seasonal affective disorder. European Archives of Psychiatry and Clinical