Rationale for this review

Despite these promising ascertions, results across studies remain heterogeneous. Some evidence suggests that overall illuminance, that is, the total intensity of light exposure, or temporal exposure patterns (duration, frequency, and timing of light exposure) may be more influential than spectral characteristics alone [70]. Methodological differences in study design, sample characteristics, exposure measurement techniques, and outcome metrics (such as axial length versus refractive error) further complicate the synthesis of results [14]. Consequently, while the potential role of spectral composition in myopia control is increasingly recognized, the field lacks a comprehensive synthesis that systematically evaluates the strength and direction of these associations.

Given the rapid translation of spectral-based findings into clinical and commercial applications, including red-light therapy devices, spectrally tuned indoor lighting systems, and violet-transmitting spectacle lenses, a critical, evidence-based appraisal of this emerging literature is both timely and necessary. A systematic review, with meta-analysis where appropriate, is needed to synthesize the available evidence on the relationship between the spectral properties of light and myopia progression in children and to evaluate whether specific light wavelengths are associated with clinically meaningful reductions in myopia progression. Moreover, such a review will identify key sources of heterogeneity, including differences in study populations, exposure definitions, follow-up durations, and outcome measures. Beyond its academic significance, understanding the influence of light’s spectral composition has direct implications for public eye health, industry, and technology. If particular wavelengths are shown to protect against myopia progression, this knowledge could inform practical, non-pharmacological interventions, such as optimizing classroom lighting, promoting safe outdoor exposure, development of wavelength-specific optical devices, smart lighting systems, and display technologies. These advances could support evidence-based innovations in the lighting and vision care industries while contributing to strategies aimed at slowing the global rise of childhood myopia.

To address this knowledge gap, this systematic review aims to answer the primary question: “What is the relationship between the spectral properties of light and myopia progression in children?” Specifically, the review will address the following questions: 1) Which wavelengths or spectral compositions of light are associated with reduced or increased myopia progression in myopic and premyopic children? 2) What is the magnitude of the effect of different light spectra on clinically relevant myopia outcomes (axial length elongation, spherical equivalent refractive error)? 3) How do variations in study characteristics such as population demographics, exposure definitions, follow-up duration, and outcome measures contribute to heterogeneity in reported associations? 4) What evidence exists to support the translation of spectral light exposure into clinically relevant interventions for myopia control in children? Through systematic synthesis of the available evidence, this review will characterize current knowledge, identify methodological gaps, and provide a foundation for future clinical trials, translational innovations, and preventive eye-care strategies informed by the photobiology of ocular growth.

Eligibility criteria

Exclusion criteria

Studies will be excluded if they involve adults older than 18 years only or if paediatric data cannot be extracted from mixed-age samples. Studies will be excluded if they focus exclusively on optical or pharmacological myopia-control interventions unrelated to light spectral exposure, including spectacles, contact lenses, orthokeratology, or atropine therapy, without evaluating the spectral characteristics of the light environment or exposure.Similarly, studies will be excluded if they do not report or define the spectral characteristics of light exposure or if they assess only visual acuity, accommodative lag, or choroidal thickness without corresponding measures of refractive error or axial length.. All forms of reviews, editorials, letters, opinions, and case reports will also be excluded. Additionally, preprints and conference abstracts that have not undergone peer review will be excluded.

Data sources, search terms, and search strategies for identifying studies

Electronic databases to be systematically searched will include PubMed, Embase, Web of Science Core Collection, Scopus, the CINAHL, PsycINFO, CENTRAL. The search will cover all records from database inception to the date of search execution. To further enhance coverage, manual searches will be performed on the first ten pages of Google Scholar, and reference lists of included articles and related systematic reviews will be screened to identify any additional eligible studies.

Given that the selected databases index studies published in multiple languages, no language restrictions will be applied during the search. Where non-English articles are identified, artificial intelligence–assisted translation tools, including large language model (LLM)-based translation approaches where appropriate, will be used to facilitate initial screening and data extraction. Translations and extracted data will be verified by multilingual researchers whenever feasible to ensure accuracy and minimize misinterpretation. The search will not be limited by publication date. The search strategy will be organized according to the PICOS framework – Population, Intervention, Comparison, Outcomes, and Study Design. Population terms will focus on children with myopia. Intervention or exposure terms will relate to spectral light characteristics. Outcomes will include myopia progression, changes in spherical equivalent refraction, axial elongation, and other refractive error changes. Study design terms will capture randomized controlled trials and observational designs.

Boolean operators (AND, OR) and truncation symbols will be used to combine related keywords, and appropriate database-specific indexing terms – such as MeSH in PubMed and Emtree in Embase – will be applied. Search terms would include (“myopia” OR “nearsightedness” OR “short-sightedness”) AND (“light exposure” OR “illumination” OR “wavelength” OR “spectral composition” OR “light spectrum” OR “blue light” OR “violet light” OR “red light” OR “natural light” OR “artificial lighting”) AND (“children” OR “adolescents” OR “school-aged” OR “pediatric”) AND (“myopia progression” OR “axial elongation” OR “AXL” OR “refractive error change” OR “spherical equivalent refraction” OR “SER”). Filters may be applied to identify both clinical and experimental studies examining myopia progression in response to specific light spectra.

All search strategies will be developed iteratively and independently reviewed by two researchers to ensure completeness and precision. The final search strings for each database will be documented in the review report to promote transparency, reproducibility, and adherence to systematic review standards.

Study selection

All search results will be exported into EndNote for compilation. The references will then be imported into the Covidence systematic review management platform for screening and review. Two independent reviewers (G.K.A. and J.A) will screen titles and abstracts to identify potentially eligible studies based on predefined inclusion and exclusion criteria. Any disagreements during this stage will be resolved through discussion, and a third reviewer (I.O.D.J) will arbitrate unresolved discrepancies. Full-text screening will then be conducted independently by the two reviewers using the eligibility criteria. Reasons for excluding studies at the full-text stage will be carefully recorded, and the overall selection process will be summarized in a PRISMA flow diagram.

Data extraction

Data from included studies will be extracted using a standardized and piloted data extraction form designed specifically for this review. Extracted variables will include general study information (author, year of publication, country, and study design); participant characteristics (sample size, mean age, sex distribution, ethnicity, and myopia baseline status); and exposure details (light source, including natural broad-spectrum sunlight versus narrow-spectrum indoor artificial lighting (e.g., LED, fluorescent, incandescent); spectral characteristics of light (wavelength or spectral composition (e.g., violet, blue, red, or full-spectrum light), spectral irradiance or spectral power distribution, illuminance intensity (lux), duration of exposure, frequency of exposure, timing of exposure (e.g., daytime versus evening); long-term safety outcomes of spectral exposure (structural and functional vision, patient-reported symptoms); near work activities (working distance, viewing duration, task type, and frequency of breaks); baseline myopia risk factors (parental history of myopia, visual task and environment, age at myopia onset, refractive status). Information regarding whether and how these potential confounding variables were measured and statistically adjusted for in the original studies will also be recorded. Outcome-related data will cover myopia progression measures such as changes in spherical equivalent refraction, axial elongation, and rate of refractive change over time, as well as follow-up duration and analytical approach. Quantitative data, including effect sizes, standard deviations, and confidence intervals, as well as any relevant qualitative findings, will be extracted. Funding sources and potential conflicts of interest will also be noted to assess study bias. Two reviewers will independently perform data extraction (G.K.A and J.A), and discrepancies will be resolved through discussion and consensus by a third reviewer (I.O.D.J).

Risk of bias

The risk of bias of all included studies will be independently assessed by two reviewers (G.K.A and J.A) using validated tools appropriate for the respective study designs. Randomized controlled trials (RCTs) will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool [77], while non-randomized or quasi-experimental studies will be evaluated with the ROBINS-I tool [78]. For observational studies, such as cohort or cross-sectional designs, the Joanna Briggs Institute (JBI) critical appraisal checklists [79] will be used. These appraisal tools would be used to assess confounding bias in the selected studies. Particular attention will be given to whether included studies adequately controlled for major environmental and behavioral confounders relevant to myopia progression. Any discrepancies between reviewers will be discussed and resolved through consensus, with arbitration by a third reviewer (I.O.D.J) when necessary.

Quality of assessment

To ensure a comprehensive evaluation of study quality, the National Heart, Lung, and Blood Institute (NHLBI) Study Quality Assessment Tools [80] will be used across relevant study designs. These tools will help assess the internal validity, methodological soundness, and potential sources of confounding bias in each study. Each included study will be rated as good, fair, or poor quality based on predefined criteria, including clarity of research objectives, adequacy of sample size, appropriateness of spectral exposure characteristics measurement (light wavelength or spectral composition, illuminance intensity, exposure duration, and spectral power distribution), near work activities, accuracy of myopia outcome assessment (e.g., axial length or refractive change), statistical robustness, and adjustment for confounding factors such as age, time spent outdoors, and temporal exposure patterns. The integration of these appraisal tools will ensure that the credibility and reliability of the evidence are rigorously evaluated. Any disagreements in quality ratings will be resolved through discussion or consultation with a third reviewer, and the final quality assessments will be reflected in the narrative synthesis and interpretation of the review findings.

Data synthesis

For heterogeneity in study designs, population characteristics, exposure definitions, and myopia outcome measures across included studies, a narrative synthesis will primarily be employed to summarize the evidence and, where feasible, through subgroup analyses. The synthesis will be structured around key dimensions such as study design (randomized, observational, or experimental), population characteristics (age, baseline myopia level, geographic region, and ethnicity), spectral exposure characteristics (light wavelength or spectral composition, illuminance intensity, spectral power, duration and timing of light exposure and light source type), and outcome measures (change in spherical equivalent refraction, axial elongation, or refractive error progression rate). Patterns of association will be explored, and methodological factors contributing to heterogeneity, such as measurement techniques, exposure quantification, and follow-up duration, will be highlighted.

Where the data characteristics of retrieved studies are sufficiently comparable (i.e., at least two studies with similar exposure definitions and outcome measures), a meta-analysis will be conducted using a random-effects model to account for between-study variability. Effect sizes (e.g., mean difference or standardized mean difference for continuous outcomes) and their 95% confidence intervals (CIs) will be calculated. Statistical heterogeneity will be quantified using the I² statistic, with values of 25%, 50%, and 75% representing low, moderate, and high heterogeneity, respectively. If adequate data are available, subgroup analyses and contextual interpretation will be performed based on wavelength range (e.g., full-spectrum, narrow-spectrum), light source (natural vs. artificial), geographic region, ethnicity or study design and behavioral and environmental myopia risk factors to assess the consistency and generalizability of the evidence. Interpretation of spectral effects will therefore consider the extent to which confounding variables were measured and adjusted for within individual studies. Sensitivity analyses will be carried out by excluding studies rated as high risk of confounding bias. Contrary, synthesis without meta-analysis (SWiM) will be adopted supposed retrieved data are heterogeneous [81].

Publication bias

If a meta-analysis includes at least ten studies, publication bias will be assessed visually using funnel plots and statistically using Egger’s regression test or Begg’s rank correlation test to detect potential small-study effects.

Confidence in cumulative evidence

The overall certainty of evidence across included studies will be assessed using the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) framework. This approach will evaluate the strength and reliability of the evidence for each outcome related to myopia progression (e.g., change in spherical equivalent refraction and axial elongation) in relation to spectral light exposure. The evidence will be rated as high, moderate, low, or very low certainty based on factors such as risk of environmental and behavioral confounding bias, consistency of results across studies, directness of evidence, precision of estimates, and potential publication bias. The GRADE assessment will provide a transparent summary of the confidence in cumulative evidence, thereby guiding the interpretation of the findings and their relevance for clinical and public health recommendations.

Handling missing data

When studies present incomplete, unclear, or missing data, the review team will attempt to contact corresponding authors to obtain clarification or missing information. If these data cannot be retrieved, the extent and type of missing information will be clearly documented. For quantitative analyses, the potential influence of missing data on effect estimates will be carefully considered. Where possible, sensitivity analyses will be performed to determine how assumptions about missing data affect the overall results. Studies with substantial missing data that may compromise validity will either be rated as lower quality or excluded from quantitative synthesis, while their findings will still contribute to the narrative synthesis. Missing data considerations will also inform the risk of bias assessment using tools such as the RoB 2, ROBINS-I, JBI, SYRCLE and NHLBI quality appraisal tools. This structured approach will help minimize bias and ensure conclusions are based on the most complete and credible evidence available.

Ethical consideration and dissemination

This study is a systematic review protocol that involves the synthesis of data obtained exclusively from previously published studies and publicly available sources. As no new human participants or animals are directly involved, and no identifiable personal data are collected or analyzed, ethical approval is not required. All data included in the review will be extracted from studies that have received prior ethical clearance as stated by their respective authors. The review will be conducted in accordance with established ethical standards for secondary research and reporting guidelines such as PRISMA. Of note, the results that emanates from this review will be disseminated through peer-reviewed journals, conference presentations, and academic symposia focused on myopia, vision science, and ocular epidemiology. In addition, findings will be shared with researchers, clinicians, and public health stakeholders to inform future experimental studies, clinical guidelines, and environmental recommendations related to light exposure and myopia control in children.

Timeline and status of review

This study protocol was developed in November 2025 with the optimization of search strategy and its registration in PROSPERO (CRD420251171027). We aim to complete this review in 8 months with literature screening (title/abstract and full-text screening) to take place in August 2026 with respect to feedback from the peer reviewers. Data extraction is expected to be completed by October 2026, with the timeline dependent on the number of included studies. Evidence synthesis, interpretation of findings, and manuscript drafting are expected to be completed by December 2026 (S2 Table).