In recent decades, there has been a significant global increase in the severity of myopia among schoolchildren, particularly in East and Southeast Asia, where the condition has reached epidemic proportions.1–3 In China, the annual incidence of myopia among schoolchildren has been reported to range from 6.3 % to 31.7 %.4–11 Myopia and high myopia can lead to permanent pathological changes and result in both direct healthcare expenditures and indirect costs such as lost productivity and reduced quality of life.12–16 Therefore, a thorough investigation of risk factors for myopia is crucial for developing effective interventions in children, especially before the onset of myopia.

Several prospective cohort studies have evaluated potential independent risk factors for myopia onset among primary schoolchildren. These studies identified risk factors including age, sex, residential area, parental myopia, daily activities, refractive error, and ocular biometric parameters such as axial length (AL) and corneal refractive power.6–9,17–21 Routine clinical examinations typically include assessments of binocular vision to optimize visual clarity and minimize asthenopia, which should be managed if necessary. However, evidence regarding the role of binocular vision in myopia development and progression remains limited.22 In 1988, Goss et al. conducted a 3-year cohort study involving 87 emmetropic schoolchildren aged 8.5 to 14.3 years in the USA.23–24 Based on univariable analysis, they found that binocular visual function parameters such as positive relative accommodation (PRA), accommodative convergence to accommodation (AC/A) ratio, near lateral heterophoria, and near fusional vergence range might be informative for predicting myopia onset. However, it remains unclear whether these parameters remain significant after adjusting for confounders. Our previous report indicated that PRA could be a potential predictor of myopia onset over a relatively short follow-up period of 2 years,25 but this observation requires verification over a longer follow-up period. Such information would be valuable for enhancing the accuracy of myopia prediction in the future.

Thus, the aim of this study was to identify independent risk factors for myopia onset, with a focus on exploring various binocular visual function parameters, and to evaluate their potential value in predicting the occurrence of myopia.

After excluding 207 schoolchildren who were lost to follow-up due to reasons such as school transfer, absence from examinations, or the use of orthokeratology lenses or atropine, a total of 866 participants (mean [SD] baseline age: 7.8 [0.7] years; 400 [46.2 %] female) with complete refraction data were included in the analysis. The characteristics of these 866 participants were compared between their baseline and graduation visits (Table 1). Compared to the baseline visit, significant increases were observed in age, AL, AC/A ratio, and time spent on near work; moreover, SER values became significantly more negative. However, at the graduation visit, there were significant decreases in CR, the magnitude of NRA and PRA, BI break point, BO break point, and time spent outdoors.

Among the 714 participants without myopia at baseline, 444 (62.2 %) developed myopia before graduation. After adjusting for confounders including baseline grade, children with the following baseline characteristics had a greater risk of myopia onset (Table 2): female sex (odds ratio [OR] = 3.03; P < .001), having two myopic parents (OR = 2.36; P = .005; compared to no myopic parents), worse UCVA (OR = 19.99; P = .007), more negative SER (OR = 0.15; P < .001), longer AL (OR = 7.28; P < .001), larger CR (OR = 2.20; P < .001), and lower magnitude of PRA (OR = 1.11; P = .02).

The correlation between baseline PRA and myopia onset before graduation remained consistent across all demographic subgroups (P-interaction > 0.05 for all; Figure 1). Consequently, the association between PRA and myopia onset did not exhibit significant variability. The ROC curves indicated that PRA had a moderate predictive capability for myopia onset before graduation, with an AUC of 0.59 (95 % CI: 0.55 to 0.63), which was less effective compared to SER (AUC = 0.65; 95 % CI: 0.61 to 0.69) but comparable to AL (AUC = 0.57; 95 % CI: 0.53 to 0.62) and CR (AUC = 0.58; 95 % CI: 0.53 to 0.62; Figure 2).

Figure 1.

Forest Plot of Positive Relative Accommodation with Various Demographic Subgroups. The dashed vertical line represents the odds ratio for the overall study population. Abbreviations: OR, odds ratio; CI, confidence interval; PRA, positive relative accommodation; SD, standard deviation.

Figure 2.

Receiver Operating Characteristic Curves of Risk Factors for Predicting Myopia Onset before Primary School Graduation. Abbreviations: ROC, receiver operating characteristic; AUC, area under the curve; UCVA, uncorrected visual acuity; SER, spherical equivalent refraction; PRA, positive relative accommodation; AL, axial length.

Our data revealed that myopia onset before elementary school graduation was associated with sex, parental myopia, UCVA, SER values, AL, CR, and notably, the magnitude of PRA at baseline. The association between PRA and myopia onset remained robust across various demographic characteristics. Moreover, PRA demonstrated a predictive capability comparable to key ocular biometric parameters such as AL and CR.

Consistent with prior studies, we found that female sex,7,17 having two myopic parents,8–9,17–18,21 more myopic refractive error,6–7,9,19,21 longer AL,20–21 and larger corneal refractive power21 at baseline were independently associated with myopia onset before primary school graduation. In our study, the myopia incidence did not differ significantly between rural and urban schools after multivariable adjustment. This finding contrasts with the one-year result reported in Taipei but aligns with the 4-year follow-up findings in Beijing.9,29 Such consistency may be attributable to urbanization driven by China's rapid economic development. Contrary to previous reports,17,19 age was not independently associated with myopia onset in our study, potentially due to the narrow age range of participants. Outdoor activities and near work were significantly associated with myopia onset in some studies,9,18,21 but not consistently across all studies.6,8 Notably, the lack of a significant association between outdoor activities and near work with myopia onset has been predominantly observed in studies involving Chinese children, consistent with our findings. This phenomenon may be explained by the increased time spent on near work, reduced outdoor exposure, and limited variability in environmental factors resulting from intense academic competition among Chinese students.

We identified that a lower magnitude of baseline PRA served as an independent risk factor for myopia onset. Our findings align with the 1996 study by Goss and colleagues, which reported that individuals who developed myopia exhibited a lower magnitude of PRA.23 This implies that a diminished capacity to sustain accommodative effort under prolonged binocular near-viewing conditions may be linked to myopia onset, potentially mediated through hyperopic retinal defocus.30–31 Notably, subgroup analyses revealed a consistent association between reduced PRA and myopia onset across various demographic and familial factors, including age, gender, school, grade, parental myopia status, and parental educational attainment. These results underscore the stability of PRA as a risk factor for myopia onset and highlight its importance in myopia prediction models and future investigative efforts.

As one of the pioneering studies investigating the independent association between binocular visual functions and myopia, our study possesses several unique strengths, including a large sample size, a relatively long follow-up period, and a high follow-up rate (4.5-year follow-up rate of 80.7 %). Nevertheless, this study also has certain limitations that warrant consideration. First, refraction measurements were conducted without cycloplegia, which might lead to an overestimation of myopic refractive error. However, we employed subjective refraction, utilizing a working lens of +2.00 D following retinoscopy performed by an experienced ophthalmologist. Compared with non-cycloplegic autorefraction, this method minimizes the impact of accommodation while maintaining operational efficiency and enhancing examination compliance. Furthermore, it facilitates the assessment of binocular visual functions, demonstrating practical value in large-scale epidemiological studies. Second, some questionnaire data were incomplete (baseline missing data: 32.8 % for near work; 23.0 % for outdoor activities). Considering the potential recall bias inherent in questionnaire-based assessments and discrepancies between reports from parents and children,32 future studies should incorporate objective and portable wearable devices to improve data accuracy.33

In conclusion, this school-based 4.5-year cohort study conducted in Wenzhou initially identifies PRA as a stable and independent risk factor for myopia onset prior to elementary school graduation, with predictive capability comparable to key ocular biometric parameters. This finding can inform future studies to enhance the accuracy of myopia prediction, thereby assisting eye care professionals and health administrators in making more informed decisions before implementing interventions in children.

This work was supported by the Scientific Research Project of Zhejiang Provincial Department of Education [grant number Y202457210], the International S&T Cooperation Program of China [grant number 2014DFA30940], and the Wenzhou Medical University-Essilor International Research Center [WEIRC, grant numbers 95013006, 95016010]. The funders had no role in the conceptualization, design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Parental written informed consent were obtained from all enrolled participants.

The datasets generated and/or analysed during the current study are not publicly available due to privacy or ethical restrictions but are available from the corresponding author on reasonable request.

All authors made substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; took part in drafting the article or revising it critically for important intellectual content; gave final approval of the version to be published; and agree to be accountable for all aspects of the work.