The effects of low-intensity narrow-band blue-light treatment compared to bright white-light treatment in sub-syndromal seasonal affective disorder

People suffering from sub-SAD in wintertime generally experience hypersomnia, a lack of energy, a craving for carbohydrates and weight gain, or a decreased interest in socializing. These symptoms sometimes are accompanied by a lowered mood, but not by an actual depression.

Kasper et al. [5, 6] have described the criteria of sub-SAD. These include a regular pattern of seasonal (winter) problems (in at least two consecutive winters for a minimum period of 4 weeks), such as decreased energy levels, less efficiency at work, decreased interest in socializing, changed eating habits (eating more carbohydrates and weight gain) and changed sleep patterns (more sleep). Subjects regard these difficulties as normal and do not see them as the symptoms of an illness. They do not seek professional help, nor do others suggest they do so. These difficulties are not recognized by people outside the subjects’ inner social circle and are easily attributed to “being overworked” or “having the flu”. The symptoms are not disturbing subjects’ lives in any major degree. Subjects have no history of major depression, nor do they suffer from any physical illness.

Although sub-SAD complaints are less severe than those of SAD, this does not imply that sub-SAD is completely harmless. Lack of energy or hypersomnia can, among other things, lead to social dysfunctioning, frustrate educational opportunities, or lead to problems in relationships or work-related problems in wintertime. Several studies have described positive effects of light in treating sub-SAD [7‐10].

Kasper et al. also formulated sub-SAD criteria based on the Seasonal Pattern Assessment Questionnaire (SPAQ, [11]). These are often used in epidemiological studies [12] to discriminate between SAD and sub-SAD. Among these criteria, the score on a subscale of the SPAQ, the Global Seasonality Score (GSS), is used as a cut-off point. The results of these studies show that the prevalence of sub-SAD is about 2-4 times higher than the prevalence of SAD [13‐15].

However, the use of the SPAQ as a discriminator between SAD and sub-SAD has been criticized [16, 17]: a high score on the GSS does not make a person depressed by definition. In view of this, in the present study, we used the Mini-International Neuropsychiatric Interview (MINI, [18]) for diagnosing/excluding depression. Potential participants who fulfilled the criteria of a mood disorder according to the MINI were excluded from the design. The Structural Interview Guide for the Hamilton Depression Rating Scale-Seasonal Affective Disorder (SIGH-SAD, [19]) was used for measuring the severity of the depressive symptoms.

The discovery of a novel photoreceptor in the retinal ganglion cells [20‐23] with a maximum sensitivity of 470–490 nm to blue light has led to research on the effects of short-wavelength light on humans. These non-image forming (NIF) photoreceptors play a role in regulating the biological clock [24‐26], but also project directly to several other areas of the brain [22, 27].

Several studies have explored the efficacy of monochromatic blue or blue-enriched light in the treatment of SAD. When more blue light was added to the spectrum of a bright-light treatment lamp (10 000 lux, Tcc = 17 000 K) the response was similar to that after exposure to bright white-light therapy (BLT, 10 000 lux, Tcc = 5 000 K) [28]. A possible explanation of this result was thought to be the saturation of the ocular receptors due to very high illuminance levels. In order to examine this, another study using the same design has compared moderate-intensity blue-enriched light (750 lux, Tcc = 17 000 K) with BLT (10 000 lux, Tcc = 5 000 K) in the treatment of SAD. This study did not show any differences in the therapeutic responses between the two light conditions [29] either.

With the invention of blue light-emitting diodes, narrow-band light sources fitting the maximum sensitivity of the NIF photoreceptors became readily available. Studies were undertaken to explore the use of LED technology in the treatment of seasonal complaints comparing the effects of light of different wavelengths. The therapeutic responses to blue light were found to be superior to those of red light [30, 31]. The effects of blue-enriched white LED light were found to be superior to placebo (deactivated ion-generator) [32] but similar to the effects of blue light [33]. In a study comparing the effects of low-intensity narrow-band blue light (BLUE) to the effects of high-intensity BLT, no differences in therapeutic outcome were found [34].

Although the effects of both light treatment and blue or blue-enriched light specifically have been studied repeatedly in SAD populations, only a small number of studies investigating the effects of light treatment in sub-SAD is available.

This is the first study exploring the effects of blue light on people with sub-SAD. As a control we used bright white-light therapy. We assessed effects on mood, energy, different aspects of activation, sleepiness, and sleep quality in the two conditions and looked for possible differences related to age and gender. In a previous study of an SAD population [28] no difference in efficacy was found between daily 20- and 30-minute sessions of exposure to bright light. In this study the effects of white and blue light in a 20-minute treatment have been compared for the first time. We hypothesize that both treatments are effective and that the therapeutic effects of BLUE are larger than those of BLT.

The human lens yellows with age, and yellow lenses can filter out short-wavelength light. Therefore older participants may profit less from the blue-light condition. In this study, the results of older participants have also been compared to those of younger participants.

Subjects were randomised (controlling for age and gender) to one of the two treatment modalities: either to the low-illuminance blue-light therapy (BLUE) group or to the bright white-light therapy (BLT) group. The interviewers were unaware of the treatment subjects were assigned to. Before starting the interview, participants were asked not to give any information about the light condition to the interviewers.

BLUE light (goLITE HF3320, Philips Drachten, The Netherlands) characteristics were: peak LED wavelength 470 nm (full width half-maximum 25 nm), usage distance 50 cm positioned on a table top at 45 degrees sideways, vertical photopic illuminance 100 lux at eye position, irradiance 1.0 Watt/m², equivalent melanopic illuminance 770 m-lux [38].

The BLT was a white fluorescent lighting device (EnergyLight HF3319, Philips Drachten, The Netherlands), correlated color temperature 5 000 K, vertical photopic illuminance at 20 cm usage distance 10 000 lux, irradiance 31.7 Watt/m²; equivalent melanopic illuminance of 8620 m-lux [38]. Spectral distributions of the two treatment modalities are shown in Fig. 1.

The subjects were instructed to use the lamps as an addition to their normal room illumination. In that sense, the BLUE condition was actually blue-enriched white light. Assuming a 3 000 K TL spectrum, and an average background illumination of 250 lux, the effective equivalent melanopic illuminance in both cases increased by 115 m-lux, making the BLT condition one order of magnitude higher on melanopic illuminance.

We compared the two conditions on a weekly and daily basis to assess the same issues, using self-rating scales, and on a weekly basis using standardized structured interviews. Since all assessment procedures have their own shortcomings (weekly = retrospective; daily = assessment at the moment of the day; self-report: subjective assessment biased by the opinion of the participant; standardized structured interview: biased by the opinion of the interviewers) we used different assessment procedures to strengthen the assessment procedure.

Each of the two conditions started at day 1 (Friday) with a baseline measurement consisting of a SIGH-SAD interview (with interviewers blind to the light condition), the Beck Depression Inventory, second version (BDI-II-NL, [39]), a fatigue self-rating questionnaire (Short Fatigue Questionnaire, SFQ; [40]), and a questionnaire aiming to evaluate subjects’ expectations of the effects of light therapy. 5-point scale ratings were collected with the help of this latter questionnaire to check whether subjects expected to benefit from each treatment modality (white and blue light), whether they thought each was a logical treatment and whether they would recommend one or the other to a friend. They filled out this questionnaire before they had seen the light fixtures. The SIGH-SAD can be subdivided in a section containing the Hamilton Rating Scale of Depression (HRSD) assessing depressive symptoms, and a section assessing the atypical symptoms (ATYP) that are common in SAD, such as hypersomnia, a decreased need for socializing and carbohydrate craving. Subjects who met all inclusion criteria were randomly assigned to one of the two conditions, with gender and age distributed evenly over the groups.

The SIGH-SAD, the BDI-II-NL and SFQ were repeated at day 8 (after the 5th light session), and at day 15.

From day 1 onwards, participants filled out questionnaires on a daily basis before 8.00 a.m., 30 minutes after waking up at the latest, but before light had been applied. The mean scores on the questionnaires of the first four days were considered baseline.

These questionnaires dealt with mood, the Adjective Mood Scale (AMS, [41‐43]); sleepiness, Karolinska Sleepiness Scale (KSS, [44]); and sleep quality, the Groninger Sleep Quality Scale (GSQS, [45, 46]). Also, four components of activation were measured, using the Activation Deactivation-Adjective Check List (AD-ACL, [47]): General Activation (GA; i.e. energetic, vigorous, full of pep, active and lively), Deactivation-Sleep (DS; i.e. sleepy, tired, drowsy, wide awake, and wakeful), High Activation (HA; i.e. jittery, intense, fearful, clutched-up, and tense), and General Deactivation (GD; i.e. placid, calm, at rest, still and quiet). Subjects were asked to describe their current feelings and to rate the 20 adjectives from the questionnaire on a 4-point scale.

Baseline differences between the scores of the two conditions on the SIGH-SAD, the BDI-II and the SFQ were tested by means of t-tests (continuous outcomes) and chi-square tests (dichotomous outcomes).

Effect sizes [48] were calculated for each condition. These effect sizes reflect the differences between baseline (day 1) and day 15. Results were based on the weekly assessments of the two conditions and were compared by means of repeated measures ANOVA. A responder was defined as a subject who improved by at least 50 %.

Linear Mixed Models were used to compare the two conditions on the basis of the daily self-rating questionnaires. An advantage of linear mixed models is that they allow the inclusion of random effects; i.e. parameters are allowed to vary across individuals. This may reveal heterogeneity in individual growth curves. We used models with time, condition, and the interaction between time and condition, with the baseline score as a covariate (baseline score = mean of the 4 pre-intervention scores). We fitted models with the baseline score and the 11 days after the start of the intervention as the repeated measures and allowed the slope to vary across individuals. Maximum likelihood estimation was used. We compared models with different variance-covariance matrices. Selection of the final model was based on the Bayesian Information Criterion (BIC; with lower values indicating better models). If the random effect for slope was found to be non-significant, this term was removed from the model (unless this resulted in a higher value of the BIC criterion). Performing residual diagnostics on the final models verified regression assumptions.

In a secondary analysis, the potential impact of gender and age on outcome were examined. To this end, the interaction time*condition*gender and time*condition*age were added to the models.

We also evaluated the correlations between the GSS scores and severity of complaints, both at baseline and throughout the study.

Analyses were carried out using SPSS 20. A two-tailed alpha level of 0.05 was used to determine statistical significance.

The research protocol was approved by the Medical Ethical Committee of the University Medical Center Groningen. All participants signed the informed consent form.

Sixty-four potential participants were invited for a selection interview. Eleven were excluded on the basis of an unclear diagnosis, symptoms that were not severe enough or factors having an effect on mood. 53 participants started the study. Five other participants were excluded during the design. Two due to incorrect use of the lamp, one due to failure to complete the assessment procedures after Day 1, and two because the inclusion criteria did not match the scores at the moment of intake, which implies that inclusion would have been a mistake (Fig. 2).

Seven participants (3 in the BLT and 4 in the BLUE condition) scored >18 on the SIGH-SAD on day 1. In the screening interview their scores had been below 18. This difference was caused by their a-typical complaints, not by depressive symptoms. These subjects were, therefore, included in the analysis.

Ultimately, 22 participants (18 women, mean age 38.2 + 10.2, range 24–51 yr.; 4 men mean age 39.6 + 10.5, range 24–51 yr.) received white light and 26 participants (22 women, mean age 38.1 + 11.6, range 18–58 yr. and 4 men mean age 48.3 + 12.9, range 31–62 yr.) received blue light. If the subjects had been divided into two groups using Kasper’s SPAQ diagnostic criteria of the global seasonality scores (GSS), GSS < 11 and GSS > 10, baseline scores on SIGH-SAD were found to be the same in these groups.

Both therapies were found to be highly effective in reducing the SIGH-SAD score and improving energy levels measured by the atypical symptom part (ATYP) of the SIGH-SAD (Fig. 3).

No statistically significant differences were found between light conditions on any of the weekly outcome measures. This also holds when the results are controlled for gender, age, severity of complaints (measured with the SIGH-SAD, HRSD, ATYP, or with the BDI-II, or SFQ), and expectations (measured with a self-rating questionnaire on day 1).

In both conditions the complaints assessed with the different instruments decreased during the 15-day period (Fig. 3 and Table 1): SIGH-SAD, 24 items, main effect ‘time’ F (2,45) = 53.9, p < 0.001), with no significant differences between conditions (main effect “condition” F (1,46) = 1.13, ns), nor over time between conditions (interaction effect “time*condition” F (2,45) = 0.06, ns). When dividing the SIGH-SAD in “typical” (HRSD) and “atypical” (ATYP) symptoms, the following is found: for HRSD (17-items Table 1) a main effect “time” F (2,45) = 22.04, ns; main effect “condition” F (1,46) = 0.35, ns; main effect “time*condition” F (2,45) = 0.28, ns and for ATYP (7 items, Table 1) a main effect “time” F (2,45) = 45.9, p = < 0.001; main effect “condition” F (1,46) = 1.09, ns; main effect “time*condition” F (2,45) = 0.05, ns. Based on the scores of the weekly assessed self-rating instruments the results for the BDI-II that were found had a main effect “time” F(2,45) = 35.13, p < 0.001; main effect “condition” F (1,46) = 0.11, ns; main effect “time*condition” F (2,45) = 1.77, ns. For the SFQ were found a main effect “time” F (2,45) = 16.77, p < 0.001; main effect “condition” F (1,46) = 0.20,ns; main effect “time*condition” F (2,45) = 0.1, ns.

In the absence of a clinical mood disorder, the atypical symptoms play a more prominent role in the seasonal difficulties as measured with the SIGH-SAD. At baseline, the subjects scored highest on fatigability and hypersomnia.

The results based on the scores of the daily self-rating questionnaires are shown in Table 2. One participant of the BLT condition was excluded from the analysis due to failure to fill out daily questionnaires during the baseline period and the majority of the following days. Consequently, in these analyses data of 47 subjects were used. The time effect was significant in all models. Both groups ended with equal scores on the daily questionnaires. For the KSS and General Deactivation subscale of the AD-ACL a significant effect of the interaction time*condition was found. Subjects receiving white light showed a larger decline in scores on the KSS and GD subscale of the AD-ACL compared to the blue light condition.

For the questionnaires concerning Mood, Sleep and the subscales Deactivation-Sleep, General Activation and High Activation of the AD-ACL the interaction time*condition was not significant.

We also examined the influence of gender and age on the outcomes. No interaction effects were found between gender or age on the one hand, and time or time*condition on the other. Adjustments for gender and age did not cause any substantial changes in the results either. Therefore, gender and age have not been included in the final models.

The severity of the GSS score of the SPAQ is not related to the severity of baseline SIGH-SAD, HRSD or ATYP scores when two groups are created on the basis of a GSS cut-off score of 11 (Table 3). In the BLT group 8 out of 22 had a GSS score of 11 or higher (36.4 %), 1 subject had a GSS score of 17. In the BLUE group 7 out of 26 had a GSS score of 11 or higher (27.9 %). When comparing the two conditions (high vs. low GSS scores) based on the SIGH-SAD scores, we find the main effect ‘time’ F (2,45) = 50.55, p < 0.001, with no significant differences between conditions (main effect “condition” F (1,46) = 0.081, ns) or over time between conditions (interaction effect “time*condition” F (2,45) = 3.14, ns).

Participants spontaneously reported some side effects. 9 % of the participants in the BLT condition reported headaches, 9 % experienced headaches and nausea, and 9 % had headaches and felt hyper during the treatment. In the BLUE condition 8 % of the participants reported headaches, 4 % experienced headaches and nausea and another 4 % reported headaches and palpitations during the treatment. Also, in the BLUE condition, 4 % reported dry eyes and yet another 4 % reported diarrhoea during treatment. No statistically significant difference in the number of side effects between the conditions was found, though.

Participants seem to equally like the colours white or blue. Participants in the conditions would also like to get the same treatment the following year, as well as the remaining part of the season. The participants who received blue light were a little happier with the treatment than those receiving white light. However, there are no significant differences between the two conditions.