Myopia has become a significant public health issue globally, with its incidence rising each year (1, 2). Epidemiological studies indicate that the high rates of myopia not only increase the social and economic burden but also present a considerable challenge to the public health system (3). From a pathological perspective, myopia is primarily characterized by abnormal axial elongation (1). Clinical evidence demonstrates that this pathological change can substantially elevate patients’ risk of developing severe vision-threatening conditions, such as myopic macular degeneration, retinal detachment, cataracts, and open-angle glaucoma (4). It is essential to note that the age of onset for myopia is closely associated with the risk of developing high myopia. For each year that the onset age of myopia in children is delayed, the likelihood of developing high myopia in adulthood is significantly reduced (5). This underscores the importance of early prevention and control of myopia.

Multiple studies have shown that factors such as the duration of near-work activities, work intensity, and light intensity are primary risk factors influencing the occurrence and progression of myopia in children (6). An epidemiological survey conducted during the COVID-19 pandemic revealed that the extended home-stay time and the increased use of electronic screens resulted in a 1.4- to 3-fold increase in the incidence of myopia among children aged 6–8 years (7). Long-term studies have confirmed that a lack of outdoor activities is a significant risk factor for the occurrence and development of myopia in children (8). Increasing the time spent on outdoor activities can effectively prevent or delay the onset of myopia (9–11). Even children who have been engaged in close work for a long time can have a positive impact on myopia prevention and control through intermittent high-intensity outdoor light exposure (12). Lingham et al. (13) suggested that the protective effect of outdoor light on myopia may be related to light intensity and spectral composition. Animal experiments have shown that, compared with other wavelengths of visible light, red light helps reduce the elongation of the vitreous cavity and increase choroidal thickness in rhesus monkeys, thereby delaying their emmetropization process (14).

Low-level laser therapy is a form of phototherapy that applies low-dose red and near-infrared light to induce various molecular, cellular, and tissue effects (15, 16). Researchers have proposed using repeated low-level red light (RLRL) devices to repeatedly direct light onto the retina over a short period to slow the development of myopia in children (17–19). RLRL therapy employs a phototherapy device emitting weak red laser light with a wavelength of 635 nm or 650 nm, an illuminance of 700 or 1,600 Lux, and a laser safety classification of Class 1 or Class 2 (20–22). During treatment, children are required to gaze at a fixed observation port in the instrument, focusing on a stationary red light spot within the irradiation device. The standard treatment regimen consists of two sessions per day (each lasting 3 min), with an interval of at least 4 h between sessions and a minimum frequency of 5 days per week. This treatment method has been validated in multiple previous studies, which have demonstrated that axial length (AL), spherical equivalent refraction (SER), and subfoveal choroidal thickness (SFCT) are all potential indicators of its efficacy in treating myopia in children (19, 23–25). Relevant meta-analyses have demonstrated that RLRL therapy helps slow down the elongation of the eye axis and the increase in refractive power in children with myopia (26–28). However, existing meta-analyses have the following limitations: an insufficient number of included studies, unclear sources of heterogeneity, and potential publication bias, which weaken the statistical efficacy of the study. This study aims to provide more robust meta-analysis results by including the latest RLRL-related studies. Additionally, by analyzing the sources of heterogeneity in results and evaluating the influence of publication bias, the clinical efficacy of RLRL therapy for children’s myopia intervention is systematically evaluated, providing more comprehensive evidence-based medical information for the management of children’s myopia.

This study followed the guidelines of the Cochrane Handbook and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (29, 30). This study was registered in the International Prospective Register of Systematic Reviews (PROSPERO; ID: CRD420251018947). The PRISMA 2020 checklist is provided in Supplementary File 1.

Increasing outdoor activity time is an effective measure for preventing and controlling myopia in children (8, 47, 48). Engaging in at least 11 h of outdoor activities per week with light intensity exceeding 1,000 lux can reduce the risk of rapid myopia progression by 54% (48). Even an additional 40 min of outdoor activity per day can decrease the probability of myopia development in non-myopic children by approximately 23% over the next 3 years (8). Meta-analyses further confirm that increasing outdoor activity time not only effectively prevents the onset of myopia in non-myopic children but also slows the progression of myopia in those who are already myopic (49).

The role of this light environment regulation in myopia prevention and control could be achieved through biological mechanisms such as the retinal dopaminergic system and hemodynamic changes. Animal experimental studies (50–52) have shown that light, as a key environmental factor affecting myopia progression, has a mechanism of action that is not only related to intensity and duration but also closely tied to the biological effects of specific light wavelengths. In myopia models of rhesus monkeys and tree shrews, narrowband long-wave light has been shown to significantly inhibit the increase in AL caused by form deprivation or hyperopic defocus (14, 53). This protective effect may be related to the photoregulation mechanism of the retinal dopaminergic system, where different wavelengths of light can differentially regulate dopamine secretion and metabolism (54). As a key neurotransmitter in eye growth regulation (55), dopamine not only directly participates in the emmetropization process but also promotes choroidal thickening by regulating the release of other transmitters, such as nitric oxide, thereby counteracting abnormal axial elongation (56–58). Oxidative stress and inflammatory responses may contribute to the pathological process of myopia by disrupting the aforementioned neurotransmitter regulation pathways (59). Especially under hypoxic conditions, oxidative damage can significantly impair the normal regulatory function of the nitric oxide and dopamine systems (60, 61).

From a hemodynamic perspective, dopamine can exert a protective effect by enhancing retinal perfusion and choroidal blood flow (62). Given that the choroid has dual functions, providing nutrition to the retina and regulating refraction (63, 64), the reduction in its blood flow is considered an important inducer of choroidal and retinal thinning and axial elongation during the development of myopia (65). This hypothesis is supported by clinical observations, which often reveal reduced ocular blood flow in myopic patients (66–68). In contrast, researchers have observed that low-concentration atropine (69), defocus incorporated multiple-segment spectacle lenses (70), and orthokeratology (71, 72) can induce choroidal thickening and changes in blood flow in the intervention of childhood myopia progression. Although the causal relationship between reduced blood flow and tissue thinning remains controversial (2), animal experiments have confirmed that narrowband long-wavelength light can induce vitreous cavity depth reduction and refractive changes related to choroidal thickening in tree shrews (73, 74). These studies provide a theoretical basis for RLRL intervention in the progression of childhood myopia. Some researchers (34, 42) have proposed a hypothesis that RLRL may decrease scleral hypoxia by increasing choroidal blood flow, reducing oxidative stress and inflammation, thereby controlling myopia progression (75–77).

Multiple RLRL clinical studies (33, 34, 37, 39, 44) have shown that it can induce choroidal thickening and slow axial elongation in myopic children. On the other hand, in terms of meta-analyses, Youssef et al. (26) verified the effectiveness of RLRL treatment in delaying childhood myopia. However, the meta-analysis was unable to further investigate the sources of heterogeneity in the studies, and the stability of the results and potential publication bias require further verification due to the limited number of studies included (a total of 5 studies). Similar problems also exist in previously published related meta-analyses (each meta-analysis included less than 10 studies), which, to some extent, weaken the statistical power of these meta-analyses (27, 28, 78).

This study evaluated the effect of RLRL treatment on the prevention and control of childhood myopia through a meta-analysis. The results showed that, compared with the control group, the RLRL group had a significant advantage in delaying axial elongation, with MDs of −0.21 mm, −0.30 mm, and −0.61 mm at 6, 12, and 24 months of follow-up, respectively; the MDs of SE changes were 0.39 D, 0.61 D, and 1.33 D, respectively; the MDs of SFCT increased by 28.12 μm and 29.72 μm at 6 and 12 months of follow-up, respectively. The results indicate that RLRL intervention can effectively delay the progression of myopia in children, and this effect may be related to the SFCT thickening induced by RLRL (65). It is worth noting that, given the heterogeneity in the current RLRL meta-analysis results, this study, benefiting from the large number of included studies (a total of 20), employed meta-regression and subgroup analysis for the first time to explore the possible sources of heterogeneity. The results showed that the mean of baseline SER and the treatment frequency were the primary sources of heterogeneity in this study. Children with more severe baseline myopia may achieve better AL and SER control effects through RLRL; higher treatment frequency may also induce thicker SFCT changes in myopic children. This conclusion is similar to the findings of Liu et al. (21), whose study showed that RLRL was more effective in myopic children than in pre-myopic children.

Notably, in the Low Myopia subgroup, there was still significant heterogeneity in AL, SE, and SFCT at 6 months of follow-up and in AL and SE at 12 months of follow-up, suggesting that this phenomenon may be related to the subgroup grouping strategy. Although the association between baseline SER and treatment efficacy has been established in the meta-regression, the subgroup grouping method based on the mean of baseline SER can only partially explain the sources of heterogeneity. We speculate that this phenomenon may be related to the lack of sample stratification based on baseline SER in the original studies. The differences in sample characteristics among the original studies may be a potential source of residual heterogeneity in the Low Myopia subgroup.

Based on the analysis of existing literature, this study first reported the presence of publication bias in RLRL treatment studies affecting SFCT. This bias was confirmed by the asymmetric funnel plot of the SFCT indicator and Egger’s test (p = 0.005). Despite the presence of publication bias, the non-parametric trim-and-fill analysis revealed that its impact on the results was minimal. Before and after the trim-and-fill analysis, the changes in SFCT at 12 months of follow-up were 29.542 μm (95% CI: 19.519, 39.565) and 27.333 μm (95% CI: 17.503, 37.163), respectively. Combined with the results of the sensitivity analysis, these findings indicate that the results of this study have good stability.

Furthermore, although this study has confirmed the efficacy of RLRL in myopia intervention, several aspects of the current RLRL therapy remain to be further explored, including the dose–response relationship, the myopia rebound effect, and treatment safety. First of all, regarding the dose–response relationship, Jiang et al.’s study (42) demonstrated that, in the RLRL group, as treatment compliance increased from <50 to >75%, the control effects on AL and SER increased from 44.6 and 41.7% to 76.8 and 87.7%, respectively. Dong et al. (35) found that an RLRL device with a power of 0.29 mW had better myopia control effects than an RLRL treatment device with a power of 0.03 mW. Zhou et al.’s study (39) did not reveal differences in the efficacy of three RLRL powers (0.37 mW, 0.60 mW, and 1.20 mW) on controlling childhood myopia. However, the trend indicated that higher power might have better efficacy. Our results also showed that SFCT was affected by the frequency of treatment. These research results collectively demonstrate the potential dose–effect relationship of RLRL treatment.

Second, RLRL therapy may induce a myopia rebound effect, wherein the progression of myopia accelerates significantly after the discontinuation of intervention compared to the treatment period. Chen et al. (34) first reported a mild rebound effect following RLRL therapy, showing that, during treatment, AL and SER increased by 0.01 mm and 0.05 D, respectively, whereas, within 3 months post-treatment, AL and SER increased by 0.16 mm and −0.20 D, respectively. Xiong et al. (17) further confirmed this phenomenon, revealing that the RLRL-single vision spectacle (SVS) group exhibited greater AL and SER increases (0.42 mm and −0.91 D, respectively) in the second year after treatment cessation compared to the SVS-SVS group (0.28 mm and −0.54 D, respectively). Notably, this rebound effect is not unique to RLRL therapy, as other myopia interventions, such as atropine (79, 80) and orthokeratology (81), also demonstrate similar rebound characteristics.

Finally, concerning treatment safety, early studies (34, 37, 40, 41, 45) primarily relied on questionnaires to assess subjective symptoms and employed fundus photography, ultrasonography, and optical coherence tomography to detect retinal structural damage. Multiple systematic reviews (28, 78, 82) have shown no evidence of irreversible structural changes or vision loss in the short term. However, existing evidence remains insufficient to confirm the long-term safety of RLRL fully. Ostrin et al. (83) believe that current adverse event metrics lack sensitivity to subtle light-induced damage. Even exposure within international safety thresholds may induce subclinical damage to the retinal pigment epithelium and photoreceptor layers (84–86). A recent high-resolution adaptive optics scanning laser ophthalmoscopy study (84) further revealed that myopic children receiving RLRL therapy for at least 1 year exhibited reduced parafoveal cone density and a 7.23-fold increased risk of abnormal signals. Therefore, a comprehensive assessment of the long-term safety of RLRL therapy remains an indispensable component of future research.

This study has the following limitations: (1) The included literature is limited to studies published in Chinese and English, and the subjects are all Chinese children, which may introduce language bias and limit the generalizability of the research results to other regions and ethnic groups; (2) due to the specificity of the intervention measures, except for the study by Dong et al. (35), which used a sham device group, the remaining RCTs did not implement a double-blind design; (3) in terms of follow-up time, most studies had a follow-up period of 6 or 12 months, and only two studies (17, 20) completed a 24-month follow-up. Therefore, it is necessary to further evaluate the long-term effects of RLRL therapy on children’s AL, SER, and SFCT; (4) due to the lack of stratified analysis based on baseline characteristics in the original studies, there may be potential residual heterogeneity in subgroup analysis, indicating the need for further evaluation of the efficacy differences of RLRL in children with different degrees of myopia; and (5) although meta-regression clarified the sources of heterogeneity, the number of SFCT-related studies in subgroup analysis was limited, which may have weakened the statistical power. Therefore, it is recommended to expand the sample size in the future to improve the reliability of the conclusions.

In summary, this meta-analysis demonstrates that RLRL treatment has significant clinical effects on pre-myopic and myopic Chinese children, including increasing SFCT thickness, delaying AL elongation, and reducing SER progression. The treatment effect is related to the degree of myopia in children. However, existing studies lack long-term safety assessments, and their treatment effects may be influenced by dose-effect relationships, carrying the risk of myopia rebound after discontinuation of the intervention. Therefore, future research should conduct larger-scale, longer-term clinical studies, especially stratified studies targeting children with varying degrees of myopia, to determine the optimal prevention and control effect of RLRL, while ensuring safety through more rigorous experimental designs.