Introduction

Although medical curiosity about the salubrious benefits of sunlight exposure can be traced to antiquity [1,2], contemporary clinical interest was ignited by the 20^th-century discovery of a direct link between light exposure and circadian melatonin production [3,4]. An obvious initial candidate for bright light therapy (LT) was seasonal affective disorder [5] – a form of depressive illness typically triggered by light deprivation during the short, cold days of winter – and the intervention has proven to be efficacious across a large number of randomized controlled trials [6-8]. In fact, the efficacy of LT has now been supported across a range of other mood disorders, including non-seasonal depression [9,10,6], bipolar disorder [11,12], antepartum and postpartum depression [13,14], and premenstrual dysphoric disorder [15,16].

Light therapy (LT) has also been successfully applied to the treatment of sleep disorders [17-20], as well as circadian phase sleep disorders associated with jet lag [21,22] and shift work [23,24]. More recently, LT has shown promise as an intervention for obsessive-compulsive symptoms [25], behavioral disturbances and functioning in dementia and Alzheimer’s disease [26,27], primary and secondary features of Parkinson’s disease [28], attention deficit hyperactivity disorder [29], seasonal variations in eating disturbances associated with bulimia nervosa [30,31], and some symptoms of chronic anorexia [32].

The most prominently promoted putative mechanism underlying LT’s therapeutic effect is the inhibition and shifting of melatonin production in the brain’s pineal gland, which may in turn induce a therapeutic alteration of dysregulated circadian rhythms [4]. In other words, LT can trigger a phase-shifting of the brain’s circadian clock, with the strongest therapeutic effect occurring in individuals with seasonal onset depression [6]. But retinal bright light exposure is also capable of altering neurotransmitter function in a number of cerebral circuits. For example, such exposure directly influences serotonin turnover in the brain, with the lowest rate of turnover in the winter, and more rapid turnover with exposure to increased luminosity [33]. Likewise, LT appears to enhance dopamenergic transmission in some brain regions [34,35].

Despite the established therapeutic potential of LT in seasonal depression and potential usefulness across a wide array of application, considerably less is known about the side effects and tolerability of LT. While a handful of landmark studies (e.g., [36-39]) have concluded that headache and eye/vision difficulties are the most common side effects of LT among mood-disordered patients, we can identify only one published study that has reported on the adverse effects of LT while employing a placebo control [40]. Interestingly, this study observed no significant differences in side effects between the LT and placebo groups, although it is worth noting that the dosage of light was considerably lower than that most commonly prescribed in contemporary practice. One additional study [41] examined the side effects of a head-mounted visor that delivered white light at illuminance levels of 60 lux, 600 lux, or 3500 lux, respectively. We do not regard this as an adequate test of LT effects, however, in as much as it utilized a non-standard light delivery method (visor) with considerably less empirical support.

Perhaps most importantly, to our knowledge there exist no published studies regarding the acute side effects of LT among non-psychiatric populations. On the contrary, most of what we know about LT side effects comes from the study of mood-disordered individuals, who are known to exhibit increased attentional vigilance for negative information [42], along with an increased propensity to interpret physiological changes as aversive [43], and increased sensation of anticipatory pain [44]. Such depressive interpretive biases could conceivably lead to the over-reporting of adverse effects while undergoing interventions such as LT. Accordingly, the examination of LT side effects among non-psychiatric individuals has the potential to be of some conceptual and applied utility. Moreover, we are aware of no placebo-controlled examinations of LT side effects utilizing the most commonly prescribed dosage and duration of LT -10,000 lux for 30 minutes [45].

The purpose of the present investigation, therefore, is to address the aforementioned gaps in the extant LT literature. Specifically, we evaluated the side effect profile of healthy young adults undergoing a 30-minute session of standard LT in comparison with that of a placebo control condition (of less than 500 lux red light). Since most LT side effects are known to emerge during the initial session and to diminish with repeated exposures [38], the present study evaluated the side effects associated with only one session at the currently prescribed dose and illuminance (10,000 lux).

Materials and Methods

Results

Two hundred and thirteen undergraduate students (56.3% female) met criteria for the study and were included in the final analyses. The age of participants ranged from 18 to 32 (M = 19, SD = 2) who slept an average of 6.8 hours (SD = 1.5) the night before the study. One hundred and twelve (52.6%) participants were randomized into the red light control group, and 101 into the bright light group. The percentage of participants in each group did not differ significantly by gender, X² (1, N = 213) = .06, p = .80. Likewise, univariate analyses of variance revealed no significant preexisting differences in baseline characteristics between participants in the two experimental conditions on age (t[211] = -.93, p = .35), hours of sleep (t[211] = .31, p = .76), or time spent outdoors prior to the experiment (t[209] = .01, p = .99). Similarly, no pre-treatment differences were seen for nausea (t[21] = .18, p = .86), headache (t[211] = -.03, p = .98), blurred vision (t[211] = 1.28, p = .20) or eye strain (t[211] = .16, p = .87). Table 1 presents the severity of side effects at pre-treatment and post-treatment for each light condition.

Analysis of variance (ANOVA) showed a significant main effect of time on eye strain, F(1, 211) = 32.36, p < .001, d = .47, and blurred vision, F(1, 211) = 7.60, p = .006, d = .19. Across both groups, the average post-treatment eye strain severity (M = .4, SD = 1.0) was greater than pre-treatment severity (M = .1, SD = .5); likewise, average post-treatment blurred vision severity (M = .2, SD = .5) was greater than the pre-treatment mean (M = .1, SD = .4). A significant main effect of time was also observed for nervousness, F(1, 209) = 40.01, p < .001, muscle twitching, F(1, 211) = 7.94, p = .005, and sweating, F(1, 210) = 14.27, p < .001, with severity scores lower at post-treatment on all three variables (Table 1). There was no significant main effect for headache, F(1, 211) = 2.41, p = .12, or nausea, F(1, 211) = .61, p = .43. Similarly, there was no significant main effect for total side effect intensity, F(1, 211) = .07, p = .80. There were no significant time-by-condition interaction effects and post hoc testing showed no significant effects of season.

Discussion

The present investigation represents the first placebo-controlled examination of bright light side effects at the standard recommended dosage -10,000 lux for 30 minutes [45] – in a non-clinical adult sample. We observed no significant between-group differences (bright light versus red light control) in the occurrence of any side effects, including eye strain, blurred vision, nausea and headache (see Table 2). Likewise, we found no significant pre-post increase in the study’s composite side effect scale as a function of bright light exposure.

Our findings run somewhat counter to those of several previous light therapy (LT) investigations (e.g., [36-38]). This discrepancy may be due to any of the three principal methodological differences from the aforementioned studies – a single exposure, comparison to a dim light control, and a sample of healthy young adults. However, the current results are highly consistent with the only other placebo-controlled study of adverse LT effects [41] in the extant literature. As in the present investigation, Volz and colleagues observed no significant differences in the side effects engendered by bright light versus dim red light. However, because these investigators utilized an outmoded light intensity by current practices in their LT condition -2500 lux, as opposed to the standard of 10,000 lux – the present study represents an important replication and extension of their work. Likewise, it stands as the first reported examination of LT side effects in a non-clinical sample.

It is important to note, however, that we did observe modest but statistically significant increases in the occurrence of two side effects – eye strain and blurred vision – following light exposure in the present study. But these small-magnitude increases were equally likely among participants in both treatment conditions, and were thus not specific effects of LT. Specifically, at post-treatment, we observed eye strain incidence rates of 22% and 23%, respectively, in the bright light versus dim light exposure groups; likewise, we found incidence rates of 12% versus 13%, respectively, for blurred vision (Figure 1). The great majority of affected participants in both groups rated these symptoms as only “mild” in severity. While uncontrolled LT investigations have also reported such visual side effects with bright light exposure [50], the present study suggests that they are a consequence of neither the illuminance nor the color of light. In fact, they may simply reflect an artifact of the experimental task (reading) used to occupy attention during the 30-minute light exposure procedure.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Percentage of Side Effects Reported for Dim Red and Bright White Light at Post-Treatment.

https://doi.org/10.1371/journal.pone.0075893.g001

Perhaps the most surprising study finding was the observation in both treatment groups of significant decreases in three reported side effects – nervousness, muscle twitching, and sweating – from pre-treatment to post-treatment. Rather than posit some sort of beneficial placebo-like impact of acute treatment on these three domains, we believe the finding is best attributed to mere participant habituation of the experimental setting.

The present study is characterized by a few important limitations. Notably, the generalizability of the present findings may be limited by the fact that we utilized only a single session of light exposure – a procedure that contrasts sharply with the typical administration of LT, which spans over a minimum of several days, and often over several weeks. While previous findings suggest that LT side effects tend to diminish over the course of repeated exposure sessions [38], we cannot rule out the possibility that some potential side effects were insufficiently triggered by a single session’s exposure. In fact, clarifying the pattern of temporal onset across potential LT side effect domains is a task that warrants the attention of future investigators in the area.

Another potential limitation lies in our use of a relatively young participant sample – a mean age of 19 – with a low average level of undesired physical symptoms at baseline (as reflected in pre-treatment responses to the study’s side effects measure). It would be desirable, therefore, to include in any attempted replication participants with a broader range of ages, including older individuals who may express greater variability in adverse side effects. Furthermore, our sample was exposed to bright light no earlier than 8 am, which contrasts with prior studies that initiate exposure earlier in the morning. However, college students are often phase shifted as indicated by our sample’s average length of sleep (less than 7 hours). Thus, we believe that the current investigation’s timing of exposure was fairly congruent for the recommended timing when adjusting for age.

Conclusion

The principal strength of the present study lies in its use of a placebo control group, as well as a rather large sample size, in a novel examination of LT-related side effects. The current investigation also controls for the presence of any side effects present prior to light exposure, thereby permitting an examination of the directionality of side effect severity. Commonly, studies of side effects only examine prevalence rates after intervention – i.e., only the emergence or remission of symptoms rather than their increase or decrease in severity. Furthermore, the present study utilized the currently accepted dosage of LT (10,000 lux) in a non-clinical sample, thereby addressing a significant gap in the extant LT literature.

The present findings thus make a potentially significant contribution to our understanding of LT side effects, an important area of investigation in light of the increasingly frequent use of LT in both clinical and non-clinical settings. Specifically, our results indicate that the reported prevalence of acute adverse side effects in the published clinical literature may not apply fully to non-clinical populations, as no increases in headache or nausea emerged in the present investigation, in contrast with the consensus of previously published reports. Moreover, although eye strain and blurred vision increased significantly (albeit modestly) from pre-treatment to post-treatment, such increases occurred to an equivalent extent in both the bright light and control conditions; thus, they did not appear to be a function of bright light exposure, per se. Taken together, the present study results suggest that LT may be particularly well tolerated by non-clinical populations. Nevertheless, further research is needed to clarify the extent to which these results may extend beyond young, non-clinical samples, as well as the degree to which some side effects may emerge over the course of repeated LT exposure.

Acknowledgments

The authors would like to thank Marshall Beauchamp, Rachel Bromberg, Cynthia Brown, Christine Eslinger, Tyler Kitson, and Bailey Widener for their assistance with study startup and data collection.