Efficacy of bright light therapy for perinatal depression: A meta-analysis of a randomized controlled trial

Abstract

BACKGROUND

Pharmacological treatments are commonly used in individuals experiencing perinatal depression (PPD); however, a debate regarding the reproductive safety of antidepressants is ongoing. Many pregnant women opt to discontinue antidepressant out of concern about potential negative effects on the developing fetus, while slow and ineffective antidepressant medications hinder improved outcomes in women with PPD. In recent years, bright light therapy (BLT) has gained traction as a treatment option for PPD; however, clinical trials findings examining the efficacy of BLT in this population have been inconclusive.

AIM

To validate the feasibility and safety of BLT for the treatment of PPD.

METHODS

We performed a meta-analysis of randomized controlled trials of patients with PPD treated with BLT vs placebo following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis. We searched PubMed, Embase, the Cochrane Library, and Web of Science for randomized controlled studies published up to December 2023. The results were evaluated using the standardized mean difference of improvement for depression scores and odds ratios (ORs) for remission rate, response rate, incidence of adverse events, and dropout rate.

RESULTS

The BLT group had higher PPD response rate [50.68% vs 33.08%; OR = 2.05; 95% confidence interval (CI): 1.25-3.35; P = 0.004; I^² = 35%] and remission rate (54.10% vs 18.52%; OR = 5.00; 95%CI: 2.09-11.99; P = 0.0003; I^² = 0%) than the placebo group. Improvements in depression scores were higher in the BLT group than the placebo group for the overall efficacy (standardized mean difference = -0.47; 95%CI: -0.80 to -0.13; P = 0.007). No significant differences between the two groups in drop-outs (21.84% vs 29.63%; OR = 0.63; 95%CI: 0.31-1.29; P = 0.21; I^² = 0%) or adverse events (17.89% vs 9.68%; OR = 2.01; 95%CI: 0.95-4.25; P = 0.07; I^² = 0%) were observed.

CONCLUSION

BLT can potentially treat PPD, showing better results than the control group in this study. BLT is effective and safe and could increase the available therapeutic options for PPD.

Key Words: Bright light therapy; Randomized controlled trial; Perinatal depression; Pregnancy; Meta-analysis

Core Tip: Bright light therapy represents a promising intervention for the treatment of perinatal depression, demonstrating significant efficacy and a favorable safety profile. This non-pharmacological approach is particularly well-received among pregnant and postpartum individuals, as well as their support networks, due to its lack of systemic side effects. The integration of bright light therapy into the psychiatric armamentarium for perinatal depression may enhance therapeutic flexibility and patient adherence, thereby contributing to improved mental health outcomes in this vulnerable population.

INTRODUCTION

Perinatal depression (PPD) is a mood disorder that occurs during pregnancy and up to 12 months after childbirth[1]. The American Psychiatric Association has categorized PPD as a subtype of major depression, with studies indicating a prevalence rate of 10-15%[2], with over 25% of perinatal women experiencing symptoms of depressed mood[3]. According to the World Health Organization 2020, mental illnesses affect 15.6% of pregnant and 19.8% of postpartum women in developing countries, with depression being the most common. A recent epidemiological review in China found that the prevalence rates of perinatal, prenatal, and postpartum depression were 16.3%, 19.7%, and 14.8%, respectively[4]. The elevated rates of depression underscore its substantial impact on the health outcomes of mothers and infants, influencing the well-being of not only the affected individuals but also their family members[5].

Conventional pharmacological treatments commonly prescribed to women during pregnancy demonstrate notable effectiveness but inherent limitations[6]. Selecting antidepressant medications during the perinatal phase is complex, given the delayed onset of action and suboptimal therapeutic outcomes associated with these drugs[7]. Furthermore, extensive research has evaluated the reproductive safety of antidepressants, with a significant increase in discontinuation rates among pregnant women[8]. Furthermore, individuals with untreated PPD are at an increased risk of suicide[9] and prone to extreme behaviors, such as extended suicide[10]. The infants of women with PPD are more susceptible to adverse outcomes such as low birth weight and perinatal death[7]. Despite the potential efficacy of conventional antidepressants in alleviating PPD symptoms, one-third of patients do not recover effectively[11]. Patients with PPD have relapse rates of around 50% in subsequent pregnancies[12]. Therefore, effective treatment methods and maintenance strategies for PPD need to be identified.

Bright light therapy (BLT) is a type of physical therapy that uses light rays. A patient is placed in a bright light environment and completes light therapy according to the established time and cycle. Its mechanism may be through regulating the neural circuits of brain function, which plays an important role in the patient’s clinical symptoms. While the neural mechanisms by which BLT improves mood are currently unknown[13], BLT has gradually been used to treat mental illnesses in recent years, demonstrating therapeutic effects. However, few studies have examined BLT as an adjunctive treatment for patients with PPD. BLT is safe, effective, and relatively inexpensive compared to other treatments, such as pharmacological and psychological treatments[14]. However, studies have been inconsistent, and evidence of its effectiveness in treating PPD needs to be substantiated. Li et al[15] conducted a systematic review of depression scale scores after receiving light therapy to treat depression and sleep in pregnant and postpartum women, concluding that light therapy was effective in relieving depression and sleep; however, the results displayed considerable heterogeneity. The purpose of this review was to systematically synthesize existing studies to further assess the efficacy of BLT in treating PPD in terms of remission rate, response rate, and improvement in depression scores before and after the intervention to identify the evidence-based medical basis for BLT in clinical treatment decisions.

MATERIALS AND METHODS

Protocol and registration

This study was registered in PROSPERO (https: //www.crd.york.ac.uk/PROSPERO/; Registration Number CRD42024560280).

Information sources and search

We systematically searched PubMed, Embase, the Cochrane Library, and Web of Science using a combination of Medical Subject Headings and free-word methods. Studies were published between the database’s inception and December 2023. The search terms used were perinatal, pregnancy, depression, postpartum depression, and light therapy (Supplementary Table 1).

Literature eligibility criteria

Inclusion criteria: (1) Randomized control trials (RCTs) written in English; (2) A sample that included perinatal women diagnosed with depression at baseline. All studies excluded patients with other systemic diseases; (3) Experimental groups were treated with BLT, and control groups received dim light therapy (DLT); (4) Outcomes were assessed using the Edinburgh postnatal depression scale, structured interview guide for the Hamilton depression rating scale, and Hamilton depression scale; and (5) The primary outcome was the response to treatment. The secondary outcomes include: Depression remission, improved depression score, adverse events, and treatment drop-out.

Exclusion criteria: (1) Repeated publications; (2) Studies not including outcome indicators; (3) Articles that were case reports, abstracts, comments, letters, and reviews; (4) Studies that used cross-sectional, cohort, case-control, single-arm, and observational designs; and (5) Animal trials.

Literature quality evaluation

We used the Cochrane risk-of-bias tool for randomized controlled studies to evaluate the possibility of bias in specific studies. The risk of bias tool evaluates selection, performance, detection, attrition, and reporting indicators[16]. The risk of bias assessment resulted in each article were classified as “low risk of bias”, “unclear”, or “high risk of bias”, and labeled as “+”, “-”, and “?”. Two reviewers assessed the risk of bias separately. When disagreements emerged between the reviewers from the included studies, a consensus was reached through discussion.

Data extraction and consolidation

Two researchers independently retrieved the first author, sample size, publication date, study site, average age, diagnosis, frequency and intensity of intervention, drug dosage, duration of intervention, depression scale scores, remission rate, response rate, adverse events, and dropouts. Response rate was defined as a drop of 50% or more in depression score from baseline to the completion of therapy, and remission rate was defined for each study.

Statistical methods

We utilized Review Manager 5.3 and Stata 12.0 software for statistical analysis. We used odds ratio (OR) and 95% confidence interval (CI) as the effect size for dichotomous variables. We chose the standardized mean difference and 95%CI as the effect size for the continuous variables because of the differences in the scales used to assess PPD. The meta-analysis indicated significant differences at P < 0.05. The heterogeneity of the entries was evaluated using I² statistics. I² < 50% was not considered significantly heterogeneous using a fixed-effects model. I² > 50% indicated significant heterogeneity; therefore, we used a random-effects model[17]. The sources of significant heterogeneity were investigated using subgroup analysis and univariate meta-regression. We performed a sensitivity analysis to test the stability of the results. Significant publication bias was indicated by P < 0.05.

RESULTS

Study selection

Totally 2262 records were retrieved through the search strategy, including PubMed (134), Embase (522), Cochrane Library (354), and Web of Science (1252; Figure 1). We removed 583 duplicate entries using Zotero. Then, the remaining articles were screened for titles and abstracts, and 1630 irrelevant entries were removed. After reading the remaining articles, 43 studies were eliminated because they did not meet the predetermined inclusion and exclusion criteria. Ultimately, six RCTs were chosen to synthesize data.

Figure 1 Flowchart for searching and screening literature. RCT: Randomized control trial.

Characteristics of the study

This study included 168 patients with PPD, 87 in the BLT group and 81 in the control group, incorporating data from a total of six RCTs[18-23] published between 2004 and 2022. In this study, the test group was defined based on their use of the BLT for PPD. The light treatments ranged in intensity from 7000 to 10000 Lux, and all studies were performed in the morning. Daily treatments ranged from 30 minutes to 60 minutes, with treatment duration ranging 3 weeks to 6 weeks. The control group received DLT using dim lights < 1000 Lux. All studies used lightboxes. Tables 1 and 2 show the main features of the included trials.

Table 1 Essential features of the included research.

Ref.	Location	Population		Age (years), mean ± SD	T/C (n)
Bais et al[18], 2020	Netherlands	Pregnant	DSM-V	T: 31.9 ± 4.4	33/34
				C: 31.9 ± 5.3
Corral et al[20], 2007	Spain	Postnatal	DSM-IV; SIGH-SAD ≥ 15	T: 34.6 ± 4.0	5/10
				C: 33.6 ± 2.1
Donmez et al[19], 2022	Marmara	Perinatal	DSM-V; EPDS ≥ 12	T: 29.73 ± 6.57	12/11
				C: 28 ± 3.8
Epperson et al[21], 2004	United States	Pregnant	DSM-IV; SIGH-SAD ≥ 20	T: 32.1 ± 3.9	5/5
				C: 32.1 ± 3.9
Garbazza et al[22], 2022	Multicenter	Perinatal	EPDS ≥ 12	T: 33	11/11
				C: 32
Wirz-Justice et al[23], 2011	Italy and Switzerland	Pregnant	DSM-IV; EPDS ≥ 10	T: 31 ± 4.7	16/11
				C: 32.7 ± 5.4

T: Treatment group; C: Control group; DSM: Diagnostic and statistical manual of mental disorders, fifth edition; SIGH-SAD: Structured interview guide for the Hamilton depression rating scale; EPDS: Edinburgh postnatal depression scale.

Table 2 Essential features of the included research.

Ref.	Test group	Control group	Duration	Main evaluation scales
Bais et al[18]	9000 Lux	100 Lux	30 minutes/day, 6 weeks	EPDS; HAMD
Corral et al[20]	10000 Lux	600 Lux	30 minutes/day, 6 weeks	EPDS; SIGH-SAD
Donmez et al[19]	10000 Lux	< 500 Lux	45 minutes/day, 3 weeks	EPDS; HAMD; SIGH-SAD
Epperson et al[21]	7000 Lux	500 Lux	60 minutes/day, 5 weeks	EPDS; HAMD; SIGH-SAD
Garbazza et al[22]	10000 Lux	19 Lux	30 minutes/day, 6 weeks	EPDS
Wirz-Justice et al[23]	7000 Lux	70 Lux	60 minutes/day, 5 weeks	SIGH-SAD; HAMD

EPDS: Edinburgh postnatal depression scale; HAMD: Hamilton depression scale; SIGH-SAD: Structured interview guide for the Hamilton depression rating scale.

Evaluation of RCT quality

Figure 2 shows the risk of bias evaluation for the six RCTs, which had a low probability of bias overall. Six studies incorporated randomization into their design, and three outlined their randomization method (i.e., random number table or computer randomization). Two provided detailed information on distribution concealment (i.e., the use of envelopes or methods for keeping the shape of the machine identical in appearance). One study had its blinding disrupted, and the remaining 5 studies used double-blinding to avoid bias. Follow-up bias was observed in one study, and a low risk was observed in the other investigations. Four RCTs were broadly in line with the individual design of the study and had a low chance of reporting bias. Two did not mention if any bias was associated with the experimental design in question. All six studies were considered at low risk for other types of bias.

Figure 2 An overview of the methodological quality for included randomized control trials.

Meta-analysis results

Response rate: We investigated the six studies’ treatment response rates, which were defined as a 50% or greater reduction in depression scores from baseline to treatment completion. The pooled data showed a statistically significant difference between the test and control groups (50.68% vs 33.08%; OR = 2.05, 95%CI: 1.25-3.35, P = 0.004, I^² = 35%; Figure 3A).

Figure 3 Meta-analysis results. A: Forest plot of response rate; B: Forest plot of remission rate; C: Forest plot of improvement in depression score. The a, b, and c represent the results of one test for Edinburgh postnatal depression scale, structured interview guide for the Hamilton depression rating scale, and Hamilton depression scale, respectively. CI: Confidence interval.

Remission rate: Four RCTs reported the remission rates of depressive symptoms after the intervention, with remission being defined differently depending on each study’s criteria. The results were statistically significant, with higher remission rates in the test group than in the control group (54.10% vs 18.52%; OR = 5.00, 95%CI: 2.09-11.99, P = 0.0003; I^² = 0%; Figure 3B).

Improvements in depression score: Five studies investigated the efficacy of neuromodulation therapies for PPD, which were measured before and after treatment using the structured interview guide for the Hamilton depression rating scale, Edinburgh postnatal depression scale, or Hamilton depression scale. According to the pooled results (standardized mean difference = -0.47, 95%CI: -0.80 to -0.13, P = 0.007, I^² = 0%), we used a fixed effects model (Figure 3C). The depression scores showed better overall efficacy against PPD in the neuromodulation group than did the control group.

Incidence of adverse events: Two studies reported detailed data on the adverse effects (AEs). Four studies reported no AEs, including headaches, dizziness, sleep problems, nausea, and vomiting, with dizziness being the most common. The two groups in each article were analyzed by adding various adverse reaction times for each article. The combined findings of these two studies found no significant difference in AEs between the trial and control groups (17.89% vs 9.68%; OR = 2.01, 95%CI: 0.95-4.25, P = 0.07; I^² = 0%; Supplementary Figure 1).

Drop-outs: We observed no significant difference in drop-outs between the two groups (21.84% vs 29.63%; OR = 0.63, 95%CI: 0.31-1.29, P = 0.21; I^² = 0%; n = 168; Supplementary Figure 2).

Sensitivity analysis: We performed sensitivity analyses for the response rates, remission rates, and depression score improvement and found that the effect sizes for all studies were within the 95%CI of the total compound effect size, with the overall study results being reliable (Supplementary Figure 3). Our analysis of the PPD remission and response rates using a one-by-one exclusion method also showed that the overall study results were reliable.

Publication bias: Funnel plots were used to determine the remission rate, response rate, and depression score improvement (Supplementary Figure 4). Since this study included fewer than ten papers, Egger’s analysis was not performed.

DISCUSSION

In the present meta-analysis, we found that BLT was effective in treating PPD. As we hypothesized, the response rates, remission rates, and depression score improvements in the test group significantly differed from the control group. The present study confirmed the efficacy of 7000-10000 Lux light therapy for PPD, while the corresponding low-intensity light treatment in the control groups showed a non-significant decrease in PPD-related depression scale scores. The 1250 Lux blue-enriched white light has also been shown to be effective in improving depression in postpartum women[24] in several single-arm evaluations[25-28]. However, owing to the limited number of studies and the small sample size in our meta-analysis, further collection of unpublished data is required to provide further support.

BLT for PPD remains a clinical challenge. Current guidelines recommend psychotherapy (e.g., cognitive-behavioral therapy and interpersonal psychotherapy) for mild-to-moderate PPD and medication for severe PPD; however, these treatments have not achieved the desired results. Although some emerging interventions (e.g., transcranial magnetic stimulation, transcranial direct current stimulation, and vagus nerve stimulation) are controversial, research has demonstrated the effectiveness of light therapy in treating PPD. However, some study results were inconsistent. One critical reason for this inconsistency is the small sample sizes of the studies or that they were conducted at a single research site, limiting the credibility of the statistical analysis. Many researchers have reported challenges in enrolling patients, resulting in the studies failing to achieve the expected sample size. However, no statistical differences between the two groups were found in terms of AEs, which were transient and disappeared shortly after treatment. We found no significant differences in the test groups’ detachment rates compared to the rates in the control group, suggesting lower AEs or no effect on detachment in the test groups in women with PPD, in response to the better safety and acceptability of BLT.

Follow-up data and literature on prolonged interventions are not currently available. The longest BLT intervention among the studies in this meta-analysis was 10 weeks, and in this intervention, one participant developed hypomania, recovering after decreasing the daily intervention time. Some patients showed improvement without symptomatic relief after increasing their daily irradiation duration, suggesting that individualized treatment for light therapy improves the likelihood of its success[21]. Subgroup analysis of the length of intervention showed better efficacy for interventions lasting three and five weeks than six weeks, similar to previous results. Previous meta-analyses reported greater efficacy among interventions less than 6 weeks[29] or 2-3 weeks[15]. However, a regression analysis of existing literature reported a significant effect for BLT interventions lasting 10 weeks compared to 5 weeks on the mood of patients with PPD[21]. Subgroup analyses for intervention intensity showed that 7000 Lux interventions were more effective than 10000 Lux interventions, which was unexpected (Supplementary Table 2). The relationship between the length of each intervention, the timing of the intervention, and efficacy remains controversial, which may be related to the mechanism of light therapy also affecting mood through hormones due to its alteration of circadian rhythms via the eye-brain hormone regulatory pathway[30]. The mechanism of BLT has not been clarified, and further studies on the light information transduction pathways associated with depression are needed.

Light therapy is a non-invasive neuromodulation treatment showing promising results; however, other neuromodulation therapies have been used to treat PPD, with some showing positive effects. Transcranial magnetic stimulation treatment, which stimulates both the right and left dorsolateral prefrontal cortex, effectively improves depressive symptoms in patients with PPD[31-33]. Vigod et al[34] concluded that a randomized clinical trial of transcranial direct current stimulation treatment for depression during pregnancy was feasible. Similarly, a case-control study demonstrated the efficacy of electroconvulsive therapy to treat PPD[35], and several case reports and related studies have also reported the feasibility and safety of vagus nerve stimulation in treating PPD[36-39]. Unfortunately, few reliable RCTs are currently available. In addition, several clinical studies on neuromodulation therapies are actively in progress, and we plan to conduct large randomized controlled trials to comprehensively study the efficacy and safety of neuromodulation therapies, including phototherapy, for treating PPD.

Our study had a few limitations, we only included studies that recruited participants diagnosed with PPD who received either BLT or DLT interventions; therefore, the sample size for the meta-analysis was small. Further collection of relevant unpublished results is required to consolidate the results of the current analysis. Although we excluded higher-risk bias studies, other biases may still be present. Additionally, some studies included participants who were also being treated using pharmacotherapy, potentially affecting the results. Third, differences in the timing of BLT existed, but all the included studies showed low heterogeneity in the results; therefore, we did not analyze heterogeneity in this meta-analysis. Our sensitivity analysis suggested stable results.

CONCLUSION

BLT showed significant efficacy in treating PPD. While it was effective in reducing depression symptoms among women with PPD, BLT’s long-term outcomes remain unclear. Currently, more randomized controlled trials are required to assess the effectiveness and safety of PPD interventions. In summary, BLT has great potential for treating PPD. Better results are available, which could expand the therapeutic options available to psychiatrists.