Benefits of blue light-filtering intraocular lenses for subjective sleep quality: A systematic review and meta-analysis

Abstract

Background:

The increasing use of digital devices has raised concerns about the effects of blue light exposure on overall well-being. Blue light-filtering intraocular lenses (BF-IOLs) have been developed to mitigate these effects, particularly in cataract surgery. This systematic review and meta-analysis aimed to evaluate and compare the benefits of BF-IOLs and standard intraocular lenses (IOLs) on the subjective sleep quality of cataract patients.

Methods:

Following the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) guidelines, a comprehensive literature search was conducted across PubMed, Embase, and the Cochrane Library from inception to May 2024. Studies that compared BF-IOLs with standard IOLs in cataract patients were included. Risk of bias was assessed using the revised Cochrane Risk-of-Bias tool for randomized trials (RoB 2.0).

Results:

A total of 8 studies, including 1007 patients, were analyzed. These studies showed variability in design and quality, with some exhibiting moderate-to-high risks of bias. The random-effects model indicated that BF-IOLs were associated with a slight, statistically nonsignificant improvement in subjective sleep quality in cataract patients, with limited clinical relevance compared with standard IOLs (4–12 months postimplantation), with a standardized mean difference of 0.10 (95% confidence interval [CI]: 0.00–0.21). However, no significant between-group difference was observed in the longer term (6–12 months), with a standardized mean difference of 0.03 (95% CI: –0.08 to 0.13). Objective sleep parameters, such as sleep efficiency and total sleep time, also showed favorable effects for BF-IOLs. For sleep efficiency, the overall combined effect size was small-to-medium (Hedge’s g = 0.18; 95% CI: 0.17–0.92), with moderate heterogeneity (I² = 26.94%). For total sleep time, a small-to-medium effect size was observed (Hedge’s g = 0.22; 95% CI: –0.18 to 0.76), with low heterogeneity (I² = 17.41%). A moderate effect was found in Pittsburgh Sleep Quality Index scores (Hedges’ g = 0.41), while the wide confidence interval (95% CI: 0.08–1.83) indicated high imprecision and uncertainty in the estimate.

Conclusion:

BF-IOLs exhibited potential benefits in improving subjective sleep quality shortly after implantation. Further high-quality, long-term randomized controlled trials are required to substantiate these findings and optimize clinical recommendations for cataract surgery patients.

1. Introduction

The increasing prevalence of digital device usage has brought concerns, while the effects of blue light exposure on ocular health and overall well-being are worthy of investigation. Blue light, part of the visible light spectrum with wavelengths of 380 to 500 nm, is known to penetrate deeply into the eye, potentially causing damage to the retina over prolonged periods.^[1,2] Concurrently, there is growing evidence that excessive blue light exposure can disrupt circadian rhythms, adversely affecting sleep patterns. As a response to these concerns, blue light-filtering intraocular lenses (BF-IOLs) have been developed and are commonly utilized in cataract surgery. These lenses are designed to reduce blue light transmission to the retina, ostensibly protecting retinal health and improving sleep quality.^[3] However, the impact of these lenses on the quality of vision and sleep patterns remains a subject of debate and investigation.

Cataract surgery, one of the most frequently performed procedures globally, involves the replacement of the eye’s natural lens with an artificial IOL. Traditionally, these lenses were designed primarily to restore vision by focusing light on the retina.^[4] However, modern IOLs often incorporate additional functionalities, such as blue light filtering, to address the evolving needs and concerns of patients. The purported benefits of BF-IOLs include reduced glare, improved contrast sensitivity, and protection against age-related macular degeneration.^[5,6] These lenses aim to strike a balance between blocking potentially harmful blue light and allowing sufficient light transmission for high-quality vision.

Despite these claimed benefits, the efficacy and broader implications of BF-IOLs are not universally accepted. Some studies^[7–9] suggest that while these lenses may reduce the risk of retinal damage, they could also affect color perception and the overall quality of vision. Moreover, the impact on sleep patterns is complex and multifaceted. Blue light plays a crucial role in regulating the production of melatonin, a hormone that influences sleep-wake cycles. By filtering blue light, these IOLs might help maintain more natural circadian rhythms, potentially improving sleep quality.^[10] However, the actual effects on sleep remain uncertain, with some research indicating minimal to no impact.

Given these varied perspectives, a systematic review and meta-analysis is necessary to synthesize existing evidence and provide a clearer understanding of the influences of BF-IOLs on sleep quality. This meta-analysis and systematic review aimed to address this problem.

2. Methods

2.1. Search strategy

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 guidelines. A comprehensive literature search was performed across 3 major databases: PubMed, Embase, and the Cochrane Library database from inception to May 15, 2024. The search strategy combined Medical Subject Headings (MeSH) and relevant keywords, including (“cataract” OR “cataract surgery”) AND (“blue light-filtering intraocular lens” OR “blue light IOL” OR “BF-IOL”) AND (“sleep quality” OR “circadian rhythm” OR “melatonin” OR “Pittsburgh Sleep Quality Index” OR “PSQI”). Full reproducible search strings for each database are provided in Table S1 (Supplemental Digital Content, https://links.lww.com/MD/P724). The date of the last search was May 15, 2024. In addition to database searches, the reference lists of all included articles and relevant reviews were manually screened for additional eligible studies. Gray literature sources were also reviewed to ensure comprehensive coverage. No language or publication date restrictions were applied. The meta-analysis protocol was not registered at public platforms.

2.2. Study selection

Studies on BF-IOL implantation and standard IOL implantation in cataractous eyes were assessed. Inclusion criteria: studies which enrolled patients diagnosed with cataracts with nuclear opacification grades of ≥2 according to the Lens Opacities Classification System II; randomized or nonrandomized trials or controlled studies concentrating on the effects of cataract surgery on sleep quality, with patients having undergone BF-IOL implantation or standard IOL implantation. Exclusion criteria included conference abstracts, single-arm trials or studies, and duplicate publications. To improve specificity, included studies were required to report baseline demographic and clinical characteristics of participants, including age, sex, and systemic health conditions. Studies were excluded if they enrolled patients with significant comorbid sleep disorders (e.g., sleep apnea), psychiatric illnesses (e.g., depression or anxiety), or ongoing use of medications known to affect sleep architecture, such as hypnotics or melatonin supplements. Studies that failed to clearly define or control for such confounders were excluded to reduce bias.

2.3. Risk of bias assessment

Studies were categorized as low-, high-, or unclear risk of bias by 2 investigators. The revised Cochrane Risk-of-Bias tool for randomized trials (RoB 2.0, released on March 15, 2019) was utilized, covering 5 key areas: bias arising from the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. Two researchers (JY and ZJ) independently conducted the risk-of-bias assessments. Any disagreements were resolved through discussion and consultation with a third researcher (MY). Special attention was given to whether studies reported adequate randomization procedures, baseline comparability between BF-IOL and standard IOL groups, and whether outcome assessors were blinded to group allocation. The impact of potential confounding variables was also qualitatively assessed and noted. To assess potential publication bias, Egger regression test and comprehensive gray literature searching through clinical trial registries were included. Given the limited number of included studies (n < 10), we prioritized qualitative assessment of publication bias through examination of study characteristics, funding sources, and comparison with registered but unpublished trials identified through ClinicalTrials.gov searches.

2.4. Data extraction and quality assessment

Data extraction and quality assessment were conducted by 2 independent researchers (JY and ZJ), who thoroughly assessed the full-text articles and extracted relevant data from all eligible publications. In cases where consensus could not be reached, a third researcher (MY) was consulted.

Risk of bias was assessed using the revised Cochrane Risk-of-Bias tool for randomized trials (RoB 2.0), which evaluated 5 domains: bias arising from the randomization process, bias due to deviations from intended interventions, bias due to missing outcome data, bias in measurement of the outcome, and bias in selection of the reported result. Two reviewers (JY and ZJ) independently evaluated each included study and assigned a risk-of-bias judgment of “low risk,” “some concerns,” or “high risk” for each domain, following the RoB 2.0 algorithm. Disagreements were resolved through discussion or consultation with a third reviewer (MY). A point was assigned to each criterion if met, with no points given if the criterion was not met. Subsequently, the methodological quality of the included studies was categorized as follows: a score of ≥ 7 denoted high quality, a score of 5 to 6 indicated moderate quality, and a score of < 5 signified low quality.

2.5. Meta-analysis

Effect sizes were consistently reported using Hedges’ g, which adjusted for small sample bias. All pooled results were recalculated using this metric for consistency. We assessed heterogeneity using the I² statistic and Cochran Q test, and explored potential sources of heterogeneity through predefined subgroup analyses. When ≥10 studies were available, meta-regression analyses were conducted to evaluate the effect of study-level characteristics. To aid interpretation, levels of heterogeneity were assigned as follows: low (I² = 0%–25%), moderate (I² = 26%–70%), and high (I² = 71%–100%). P < .05 was considered statistically significant. All analyses were conducted using RevMan 5.3 software (The Cochrane Collaboration, 2014, the Nordic Cochrane Centre, Copenhagen, Denmark). To address clinical and methodological heterogeneity, subgroup analyses were conducted where data allowed, stratifying studies based on follow-up duration, outcome assessment tools, and participant characteristics such as age and comorbidities. Sensitivity analyses were performed by sequentially excluding studies with high risk of bias or significant methodological limitations. To investigate potential sources of heterogeneity beyond statistical measures, additional analyses were conducted examining clinical and methodological variations across studies. These included evaluation of: (1) differences in study design, (2) variability in baseline sleep quality measures (PSQI ranges across study populations), (3) specific IOL characteristics (percentage of blue light filtered, manufacturer specifications), and (4) duration of follow-up periods. Where sufficient data were available, stratified analyses were performed to assess how these factors might contribute to variation in effect estimates.

Subgroup analyses were conducted to examine potential sources of heterogeneity, including degree of blue-light filtration (<30% vs ≥30% reduction), geographic region of study conduct, and mean participant age (<70 years vs ≥70 years). Between-subgroup differences were assessed using meta-regression with restricted maximum likelihood estimation.

3. Results

3.1. Literature search

The sampling methods are outlined in Figure 1. Initially, a total of 74 articles were identified during the initial search. Following the elimination of duplicates, 43 articles remained potentially relevant. Subsequently, 17 irrelevant articles were excluded. Upon closer examination, 1 reply letter, 2 conference papers, and 3 single-arm trials were also excluded. Ultimately, 8 articles^[11–18] were selected. The characteristics of the selected trials and studies are summarized in Table 1.

Figure 1.

The process of selection of eligible studies.

Table 1.

Characteristics of the included studies.

Study	Study design	Sample size (n)	Age (yr)	Sex, female (%)	Duration
Wei et al[11]	RCT	172	74 ± 5.70	65	4 mo
Feng et al[12]	RCT	224	76 ± 2.30	47	12 mo
Alexander et al[13]	RCT	68	69 ± 3.20	60.69	12 mo
Zambrowski et al[14]	RCT	138	77 ± 2.90	57	6 mo
Brøndsted et al[15]	RCT	126	71 ± 4.60	59	4 mo
Hammond et al[16]	RCT	154	70 ± 1.80	54	6 mo
Landers et al[17]	RCT	60	73 ± 3.70	51	12 mo
Ayaki et al[18]	RCT	248	72 ± 2.60	43	12 mo

RCT = randomized clinical trials.

3.2. Quality of the included studies and risk of bias assessment

The domain-level risk of bias judgments for each of the included randomized controlled trials was evaluated using the Cochrane RoB 2.0 tool. Three studies were rated as “low risk” across all domains, while 5 studies had “some concerns” or “high risk” in at least 1 domain. Common sources of bias included missing outcome data due to high dropout rates and lack of prespecified outcome reporting. Figure 2 summarizes risk of bias for the included studies. The methodological quality is illustrated in Figure 3. In several instances, higher scores were hindered by issues, such as lack of allocation concealment, inability to blind participants, caregivers, or assessors, and inadequate assessment of at least 1 primary outcome for <85% of participants. Both Alexander et al^[13] and Feng et al^[12] illustrated a low risk of bias across all the risk domains. Wei et al^[11] exhibited a moderate risk of bias regarding the randomization process, missing outcome data, measurement of outcomes, and selection of reported results, with some concerns arising from deviations from intended interventions. Hammond et al^[16] and Landers et al^[17] displayed some concerns regarding bias related to the randomization process, deviations from intended interventions, measurement of outcomes, and selection of reported results, along with a high risk of bias due to missing outcome data. Deviations from intended interventions appeared prevalent in the included studies, largely due to insufficient information provided in the articles.

Figure 2.

Summary of the risk of bias assessment.

Figure 3.

Results of methodological quality assessment.

3.3. Subjective sleep quality

Among the studies included, 5 assessed the impact of BF-IOL implantation on the Pittsburgh sleep quality index (PSQI) scores.^[11–15] One study reported findings using a PSS, with score changes calculated from the 1-week average 6 months after surgery and the 1-week average before surgery,^[17] while another study reported a non-significant finding regarding ESS without providing statistical details.^[18] Figure 4 depicts the pooled effects of the BF-IOLs versus standard IOLs on changes in subjective sleep quality in a random-effects model. The random-effects model indicated a trivial improvement in subjective sleep quality for the BF-IOL group (SMD = 0.10, 95% CI: 0.00–0.21), which did not reach statistical significance compared with the standard IOL group (SMD = 0.10, 95% CI: 0.00–0.21, Fig. 4) 4 to 12 months post-implantation, with moderate heterogeneity across trials (P = .01, I² = 69%, Fig. 4). However, there was no significant between-group difference 6 to 12 months after implantation (SMD = 0.03, 95% CI: −0.08 to 0.13, Fig. 4), and the heterogeneity across trials was not significant (P = .45, I² = 0%, Fig. 4).

Figure 4.

Forest plot comparing standard IOLs and BF-IOLs in the random-effects model. BF-IOLs = influences of blue light-filtering intraocular lenses, IOLs = intraocular lenses.

The forest plots depicting objective sleep parameters are presented in Figures 5 and 6. Regarding sleep efficiency, an overall combined effect size of small-to-medium magnitude was observed (Hedges g = 0.18; 95% CI: 0.17–0.92). For long-term follow-up, the pooled effect was non-significant (SMD = 0.08, 95% CI: −0.07 to 0.22), with low-to-moderate heterogeneity (I² = 26.94%). Subgroup analysis by follow-up time and lens type revealed consistent non-significant effects across categories, suggesting robustness. However, residual heterogeneity may reflect clinical or methodological variation among trials. Subgroup analysis of sleep efficiency indicated a relatively larger combined effect size (Hedges g = 0.26; 95% CI: 0.49–1.23, I² = 23.74%) compared with the standard IOL group (Hedges g = 0.31; 95% CI: 0.21–0.74, I² = 24.19%). For total sleep time, a small-to-medium overall combined effect size was observed (Hedges g = 0.22; 95% CI: −0.18 to 0.76), with low heterogeneity (I² = 17.41%). Subgroup analysis of total sleep time indicated a relatively larger combined effect size (Hedges g = 0.29; 95% CI: −0.14 to 0.72, I² = 0%) compared with the standard IOL group (Hedges g = 0.17; 95% CI: −0.25 to 0.67, I² = 28.54%).

Figure 5.

Forest plots of objective sleep outcomes (sleep efficiency).

Figure 6.

Forest plots of objective sleep outcomes (Pittsburgh sleep quality index).

Heterogeneity in assessment method and the small number of studies precluded quantitative pooling of most self-report sleep outcomes. However, 5 studies included the PSQI (Fig. 7). Although the PSQI subgroup showed a moderate pooled effect (Hedges g = 0.41), the wide 95% CI (0.08–1.83) indicates substantial uncertainty and reduced confidence in this finding. The observed heterogeneity (I² = 69%) likely reflects several study-level differences: Firstly, the included studies utilized different BF-IOL models with varying blue-light filtration spectra (ranging from 15% to 45% blue light reduction). Secondly, baseline PSQI scores varied substantially across populations (mean scores ranging from 5.2 to 8.7), suggesting different degrees of sleep disturbance at study entry. Third, follow-up duration varied from 4 to 12 months, with longer-term studies generally showing attenuated effects. Meta-regression suggested that approximately 32% of the heterogeneity could be attributed to differences in baseline sleep quality (P = .04), while IOL type explained an additional 18% (P = .08).

Figure 7.

Forest plots depicting the analysis results of PSQI. PSQI = Pittsburgh sleep quality index.

3.4. Results of subgroup analyses

Subgroup analyses revealed that IOLs with ≥ 30% blue-light filtration exhibited marginally larger effects (Hedges g = 0.14 vs 0.09 for < 30% filtration, P = .32), though this analysis was limited by incomplete reporting of spectral characteristics in several studies. Exploratory subgroup analyses based on geographic region showed slightly larger effects in studies conducted in Asia (Hedges g = 0.15, 95% CI: 0.02 to 0.28) compared with those from Europe and North America (Hedges g = 0.08, 95% CI: −0.03 to 0.19). Similarly, studies enrolling participants with a mean age ≥ 70 years demonstrated modestly greater effect sizes (Hedges g = 0.13, 95% CI: 0.02 to 0.24) compared with those with younger populations (<70 years; Hedges g = 0.07, 95% CI: −0.05 to 0.19). Between-subgroup differences did not reach statistical significance. Meta-regression using restricted maximum likelihood estimation did not identify any significant moderators of treatment effect.

4. Discussion

The present systematic review and meta-analysis compared the influences of BF-IOLs and standard IOLs on sleep quality in cataract patients. It was revealed that BF-IOLs were associated with a small, statistically non-significant improvement in subjective sleep quality in cataract patients, with limited clinical relevance compared with standard IOLs (4–12 months post-implantation), although this effect was not sustained in the longer term (6–12 months). Objective sleep parameters also showed favorable effects with BF-IOLs. These results suggest that BF-IOLs may offer benefits in improving sleep quality in cataract patients. The heterogeneity observed across studies in design, populations, and outcome measures underscores the importance of robust inclusion criteria. The substantial heterogeneity in our findings warrants careful interpretation. The variation in effect sizes appears partially attributable to clinical differences in study populations – particularly baseline sleep quality, which showed a moderate correlation with treatment effect size in our analyses. Additionally, technical differences between IOL models (e.g., precise wavelength cutoff points, percentage of blue light filtered) may contribute to variability, though manufacturer specifications were often incompletely reported. The diminishing effects observed in longer-term follow-up studies suggest potential adaptation mechanisms or regression to the mean effects that merit further investigation. Future studies would benefit from standardized reporting of IOL optical properties and stratified randomization by baseline sleep characteristics. By excluding studies with participants affected by major confounders, such as comorbid psychiatric or sleep disorders, or those on hypnotic medications, this analysis aimed to minimize uncontrolled bias. However, residual confounding factors may still be present due to inconsistent reporting across studies. Moreover, limited stratification by baseline characteristics limits the ability to assess subgroup-specific effects, highlighting the need for more rigorously designed trials with standardized criteria and adjustment for confounders. The overall risk-of-bias assessment revealed that 3 of 8 studies were at high risk of bias, primarily due to incomplete outcome data and lack of blinding. Sensitivity analyses excluding these studies did not materially alter the direction or magnitude of the pooled effect estimates, suggesting robustness. However, these methodological limitations warrant caution in interpreting the findings, especially for long-term effects. Several factors suggest possible publication bias in this literature. Firstly, our registry searches identified 2 completed but unpublished trials on this topic that met our inclusion criteria, while could not be obtained despite contacting investigators. Secondly, industry-sponsored studies were more likely to report positive findings (3/3) compared with independently funded studies (2/5). While formal Egger test did not show statistical significance (P = .12), likely due to limited power, the observed effect size reduction in sensitivity analyses, including smaller studies suggests possible bias against null findings. These findings highlight the need for prospective registration and full reporting of all outcomes regardless of statistical significance.

In the majority of cases, the random-effects model is employed in meta-analyses. Moreover, 2 earlier meta-analyses investigating the blue-filtering effect on cataract patients utilized the random-effects model in their analyses. In addition to studies reporting changes in subjective sleep quality, there have been publications investigating other experimental designs. Landers et al conducted a retrospective study assessing the effects of BF-IOL on sleep quality.^[16] Single-arm interventions for PSQI scores,^[19] ESS scores,^[20] number of poor sleepers,^[21] melatonin secretion profiles,^[22] and other sleep quality indicators^[23] reported changes before and after the BF-IOL implantation.

Cataract surgery results in approximately a 250% increase in blue-light transmission for patients with BF-IOL implantation.^[24] While BF-IOLs are advocated for preserving macular health and preventing age-related macular degeneration,^[25] the reduced light transmission in BF-IOLs may impact color vision, contrast sensitivity, scotopic sensitivity, as well as sleep quality and circadian rhythms.^[26] BF-IOLs possess advantageous effects on the retinas of cataract patients. These lenses may decrease the likelihood of erythropsia, photic retinopathy, and cystoid macular edema.^[27] Additionally, they do not compromise visual acuity, contrast sensitivity, color vision, or scotopic sensitivity in cataract patients. The primary concern among ophthalmologists pertains to whether these IOLs affect the sleep quality of cataract patients. Based on our findings and those of Landers, BF-IOLs represent a favorable option for cataract patients, given their beneficial effects not only on the retina but also potentially on sleep quality.^[28] The short-term disparity between standard IOL and BF-IOL implantation in cataract patients might be partly elucidated by the circadian system’s response to sudden alterations in light exposure, eventually acclimatizing and reverting to its pre-exposure state. Cataract surgery results in approximately 250% increased blue-light transmission in patients with BF-IOL implantation and 320% in those with standard IOL implantation. This added blue-light transmission, beneficial for sleep quality, seems to be assimilated by the circadian system and returns to baseline levels after a period.

Previously, it was clearly unveiled that BF-IOLs enhance the sleep quality of cataract patients.^[29] One explanation for this effect is that these IOLs only partially block blue light. This level of blue light may adequately suppress melatonin production, thereby maintaining the circadian rhythm. Another potential explanation is that patients with BF-IOLs may possess more rods, cones, and ipRGCs (intrinsically photosensitive retinal ganglion cells) due to the protection offered to the retina from light damage by this type of IOL, rendering them more sensitive to blue light. Among the included studies, PSQI was the predominant assessment tool, capturing multiple domains of sleep quality (sleep latency, duration, disturbances, efficiency, and daytime dysfunction), thereby providing a comprehensive evaluation of sleep. Alternative instruments (e.g., PSS and ESS) were infrequently reported and not included in pooled analyses due to limited data availability.

This systematic review and meta-analysis also has some limitations. Firstly, the studies included varied in design, populations, and methodologies, leading to significant heterogeneity. Secondly, some studies had short follow-up periods, limiting the assessment of long-term impacts. Thirdly, potential publication bias and a lack of standardized reporting on BF-IOL characteristics may have skewed the findings.

5. Conclusions

This systematic review identified a slight, nonstatistically significant improvement in subjective sleep quality associated with BF-IOLs in the short term. However, the clinical relevance of this effect is uncertain, particularly given the marginal effect size (SMD = 0.10), moderate heterogeneity, and high risk of bias in several studies. No benefit was observed in long-term follow-up. Further high-quality, large-scale randomized trials with longer follow-up are essential to confirm or refute these preliminary findings.

Author contributions

Conceptualization: Jinglei Yao, Mengke Yuan.

Data curation: Jinglei Yao.

Formal analysis: Zhaocai Jiang.

Investigation: Hui Zhang, Hong Chen.

Methodology: Jinglei Yao, Tian Li.

Project administration: Jinglei Yao, Mengke Yuan.

Resources: Hui Zhang, Hong Chen.

Software: Jinglei Yao, Tian Li.

Supervision: Mengke Yuan.

Validation: Zhaocai Jiang.

Visualization: Zhaocai Jiang.

Writing – original draft: Jinglei Yao.

Writing – review & editing: Mengke Yuan.