The effect of blue-light blocking spectacle lenses on visual performance, macular health and the sleep-wake cycle: a systematic review of the literature

Rationale

Studies, in animal models1, 2 and cell culture,3, 4 have shown that wavelengths in the blue portion of the electromagnetic spectrum (400–500 nm) can induce phototoxic retinal damage. Historically, two mechanisms of photochemical damage have been recognised and eponymously named as ‘Noell damage’ and ‘Ham damage’ after the original investigators.1, 5 Noell, or Class I, damage was first observed following prolonged exposure of albino rats to fluorescent light (490–580 nm). Cellular disruption occurred initially in photoreceptors, followed by the retinal pigment epithelium (RPE). By contrast, Ham5 (Class II damage) described disruption that occurred after shorter, high intensity light exposures (between 10 s and 2 h’ duration). Shorter wavelengths were associated with more intense cellular damage, initially at the level of the RPE, with a peak of the action spectrum occurring at around 440 nm in the phakic eye. International standards have been developed based on these empirical studies6, which define exposure limits, below which adverse effects are unlikely to occur. However, driven by requirements for brighter and lower energy lighting, the last 10 years has seen significant changes in light sources for both commercial and domestic applications, with an increased use of compact fluorescent lamps (CFL) and high intensity light-emitting diodes (LEDs). Moreover, white-light LEDs (the most common type of LED) have become ubiquitous in backlit displays in smartphones and tablet computers. Although the light emitted by these LEDs appears white, their emission spectra show peak emissions at wavelengths corresponding to the peak of the blue light hazard function. It has been shown that exposure of cultured RPE cells to light equivalent to that emitted from mobile display devices causes increased free radical production and reduced cell viability.7 This has raised concerns that the cumulative exposure to blue light from such sources may induce retinal toxicity and potentially increase the risk of age-related macular degeneration.8

The rationale for the introduction of blue-blocking ophthalmic lenses was to mitigate the risk of retinal toxicity by blocking, or attenuating, short wavelength visible light, usually in the range 400 nm to 500 nm. These ophthalmic devices, which include spectacle lenses, contact lenses and intra-ocular lenses (IOLs), contain or are coated with dyes that selectively absorb blue and violet light. The choice between a conventional ultraviolet (UV) light blocking IOL and a blue-blocking IOL following cataract surgery has generated significant debate in the literature in terms of achieving a balance between photoreception and photoprotection.9-12 Possible disadvantages of blocking short-wavelength visible light transmission include disturbances of colour perception, decreased scotopic sensitivity (leading to poorer performance in dim lighting conditions) and disruption of the timing of the circadian system.13 Intrinsically photosensitive retinal ganglion cells, which provide photic input to the central circadian clock in the suprachiasmatic nucleus, express melanopsin and have an absorption peak at approximately 480 nm in the blue part of the spectrum.14

Compared to their intra-ocular counterpart, blue-blocking spectacle lenses have received relatively little scientific attention. Standard spectacle lenses generally offer protection against UV (up to wavelengths of 380 nm) and the adding of a yellow chromophore can also reduce or eliminate blue light transmission. Alternatively, anti-reflection interference coatings can be applied to both the anterior and posterior lens surfaces, to selectively attenuate parts of the blue-violet light spectrum (415 to 455 nm); this range of wavelengths includes a significant proportion of the blue light hazard function15, while the lens remains transparent to other wavelengths of visible light. In addition to their putative benefit for retinal protection, blue-blocking spectacle lenses have also been claimed to improve sleep quality following the use of electronic devices at night,16 and reduce eye fatigue and symptoms of eye strain during intensive computer tasks.17

A systematic review of the best available research evidence is essential to assess the appropriateness of marketing blue-blocking spectacle lenses at the general spectacle wearing population. This evaluation will consider both the relative benefits and potential harms of these lenses.

Objectives

Search strategy

We conducted searches using the following bibliographic databases: Ovid MEDLINE, Ovid EMBASE, PubMed and the Cochrane Library for relevant articles published before May 2017. We did not use any date or language restrictions for the bibliographic searches. An example search strategy for one of the databases (Ovid MEDLINE) is included in File S1. We also scanned the reference list of included studies and contacted experts in the field to ask if they were aware of additional published or on-going trials investigating blue-blocking lenses. We searched the PROSPERO database for relevant systematic reviews and searched clinical trials registries (Clinical trials.gov and the ISRCTN registry) for recently completed or on-going trials.

Inclusion and exclusion criteria

We included randomised controlled trials (RCTs) and pseudo-randomised controlled trials, which recruited adults, aged 18 years and above, from the general population and compared blue-blocking spectacle lenses to standard spectacles lenses, or any other comparator, where it was possible to isolate the effect of the blue-blocking lens for any of our primary or secondary outcomes. The review team decided post-hoc that this should include comparisons between high and low blue-blocking lenses. We defined blue-blocking lenses as those that block or attenuate short wavelength optical radiation between 400 nm and 500 nm.

The following outcomes were considered:

For the evaluation of visual performance and effect of the intervention on alertness and/or sleep quality, we included any measure conducted during the follow-up period of the trial. To assess the effects of blue-blocking spectacle lenses on macular health or function, studies had to be at least 6 months duration.

Data extraction and analysis

Following removal of duplicates, two reviewers (JL and CH) independently screened the titles and abstracts identified from the bibliographic searches and resolved any discrepancies by discussion and consensus. We obtained full-text copies of potentially eligible studies and these were assessed by both reviewers to decide whether they met the inclusion criteria. Reasons for exclusion were documented at this stage. We used a data extraction form that was developed and piloted for the purpose of this review. We collected data on: study design, details of participants, details of intervention, methodology, quantitative data on outcomes and funding sources. Data extraction was conducted independently by two reviewers (JL and CH) and any discrepancies resolved by discussion. The extracted numerical data was entered into Revman 518 meta-analytical software by one reviewer (JL) and this was checked by a second reviewer (CH).

Any differences of opinion in risk of bias assessments were resolved by discussion.

Our measure of treatment effect was the risk ratio (RR) for dichotomous outcomes and the mean difference (MD) for continuous outcomes, with 95% confidence intervals [CIs].

By definition, the intervention was applied to the person and therefore the unit of analysis was the same as the unit of randomisation. However, where data was presented from both eyes, we analysed the data from the right eye only to avoid a unit of analysis error. Insufficient studies were available to conduct the planned meta-analysis. However a descriptive summary of the results of the included studies has been provided. Publication bias could not be assessed, as there were an insufficient number of studies to conduct this analysis.

We assessed the certainty of the evidence using the Grades of Recommendation, Assessment and Evaluation (GRADE) Working Group approach,20 using customised software (GRADEpro GDT). One reviewer (JL) conducted the initial assessment and this was checked by the other reviewers (CH and LD). We considered risk of bias, inconsistency, indirectness, imprecision, and publication bias when judging the certainty of the evidence.

Results of the searches

The electronic searches yielded 118 references (see Figure 1 for the PRISMA flow diagram). After 19 duplicates were removed, we screened the remaining 99 references and obtained the full-text reports of 15 references for further assessment. Twelve of these17, 21-31 were eliminated (see Table of Excluded Studies in File S2 and three RCTs that met the a priori criteria for inclusion were included in the final analysis (see Characteristics of Included Studies in File S3. We did not identify any on-going studies from our searches of the clinical trials registries.

Characteristics of included studies

We included three studies in this review.32-34 Two of the studies were conducted in the USA and one in Hong Kong.

Burkhart and Phelps32 randomised 20 adult volunteers reporting sleep difficulty to wear either amber tinted glasses (blocking wavelengths <550 nm) or yellow tinted placebo glasses (blocking wavelengths <465 nm) for three hours prior to sleep. The primary outcome measure was sleep quality as determined by sleep diaries, which incorporated a 10-point Likert sleep quality scale. Sleep diaries were completed for 1 week prior to the intervention (baseline) and for 2 weeks afterwards.

Leung and co-workers33 conducted a pseudo-randomised controlled trial involving 80 computer users from two age cohorts: young adults, 18–30 years, n = 40 and middle aged adults 40–55 years, n = 40. Participants were randomised into one of three groups to assess the performance of two blue-blocking spectacle lenses (blue-blocking anti-reflection coating and a brown tinted lens) and a regular clear control lens, using a crossover design. The primary outcomes were contrast sensitivity, using the Mars contrast sensitivity letter chart under standard and glare conditions, and colour discrimination using the Farnsworth-Munsell 100-hue test. Following the visual assessment tests, participants wore each assigned lens for one month for a minimum of two hours per day. At the end of each wearing period, lens performance was subjectively assessed using a 13-item questionnaire. Each question was rated on a 1–5 scale (where 1 = very unsatisfactory and 5 = very satisfactory).

Lin and co-workers34 recruited 36 adult subjects who were randomised to one of three groups and wore either spectacles with low or high blue-blocking lenses or non-blue blocking lenses for a 2 h computer task using a laptop computer. At the end of the task, critical fusion frequency (CFF) was assessed and symptoms of eyestrain were evaluated using a 15-item questionnaire. The CFF is the lowest level of continuous flicker that is perceived as a steady source of light and a reduction in CFF was interpreted as a measure of eye fatigue.

Risk of bias and certainty of the evidence

We evaluated the risk of bias in the included studies using the Cochrane risk of bias tool.19 Figures 2 and 3 present a graph and summary of the risk of bias for the included studies. Overall the studies were at an unclear or high risk of bias. We rated two studies32, 34 as having an unclear risk of selection bias, since they did not describe the method for random sequence generation or how this was concealed. Leung and colleagues33 allocated participants to different sequences of lens wear by date of admission and therefore the sequence was non-random and at a high risk of selection bias. Given that two of the included studies randomised small numbers of participants,32, 34 there were baseline differences in the outcome of interest, which may have affected the results. Although attempts were made to mask outcome assessors to the intervention received, it was not possible to mask participants due to differences in appearance between the lenses being tested. We judged one study34 to be at a high risk of selective reporting bias, due to a failure to report on 2/15 of the questions from the symptom questionnaire and no protocol or trial registration was available. Two studies32, 33 were judged to be at an unclear risk of selective reporting since either no protocol or trial registry entry was available, or in one case the trial was retrospectively registered.33

We rated the certainty of evidence for each outcome using GRADE (see Table 1).

Effects of the intervention

Primary outcome measures

Two studies33, 34 randomising 116 participants, provided data on differences in visual performance with blue-blocking lenses compared to a clear control lens. Leung et al33 investigated the effect of blue-blocking lenses on contrast sensitivity and colour vision using a crossover design. There was no evidence of a difference in log contrast sensitivity or total error score on the FM 100-hue test between the intervention and control lenses (Table 1). Lin et al34 measured CFF (a proxy measure of eye fatigue) before and after a two-hour computer task. There was no observed difference between the low-blocking and no-blocking (clear) lens groups, but there was evidence of a less negative change in CFF between the high and low-blocking lens groups indicating less fatigue with computer use for the high-block group (Figure 4).

These studies also compared symptoms of eyestrain for the intervention and control lenses using Likert rating scales.33, 34 Leung et al.33 measured symptoms of eyestrain on a 5-point scale after 1 month of wearing low blue-blocking (blue-filtering anti-reflection coating), high blue-blocking (brown-tinted) or control (non blue-blocking) lenses. There was no significant difference between the intervention and control lenses for either the low blue-blocking lens (Mean difference (MD) = 0.00 [−0.22, 0.22]) or the high blue-blocking lens (MD = −0.05 [−0.31, 0.21]). Lin et al34 compared symptoms related to eye fatigue or eye strain before and after a two hour computer task for participants wearing clear (control) lenses or low or high blue-blocking lenses using a 15-item questionnaire. Since there was no statistical difference between the low blue-blocking and clear lens groups, the study authors pooled the data for the low blue-blocking and clear lens participants and compared the symptom scores, after the task, for each question. Statistical differences between groups, for each questionnaire item, were then investigated using the Mann–Whitney U test. For the current review, we analysed the ordinal data from the 13 questionnaire items reported and calculated the proportion of subjects in each group showing a post-task symptomatic improvement for each question. The risk ratio (RR) with 95% confidence intervals was calculated for each question using Revman18 (Table 2). A significant symptomatic improvement was found for only one question ‘My eyes feel itchy’ (RR 2.68 [1.32, 5.44]).

Table 2. Analysis of symptom questionnaire from Lin et al34 comparing subjects wearing high blue blocking lenses to those wearing low blue-blocking or clear lenses

Question	RR (95%CI)
I feel pain around or inside my eyes	2.60 [0.85, 7.98]
My eyes feel heavy	2.50 [0.95, 6.57]
My eyes feel itchy	2.68 [1.32, 5.44]
My eyes feel tired	3.33 [0.95, 11.66]
I find it hard to focus my eyesight	1.75 [0.83, 3.67]
I see written or computer text as blurry	1.67 [0.54, 5.11]
My computer monitor looks too bright	1.28 [0.44, 3.67]
I feel tired when doing work	2.08 [0.74, 5.84]
My neck shoulders, back and lower back hurt	0.52 [0.13, 2.09]
My fingers hurt	0.52 [0.07, 4.17]
I feel mentally stressed	1.30 [0.54, 3.14]
The suns glare affects my eyes when outdoors	1.37 [0.55, 3.40]
I find fluorescent office lighting to be bothersome to my eyes	7.00 [0.88, 55.66]

• RR, Risk Ratio.