Abstract
Aims
To identify factors contributing to rapid axial length (AL) growth in children aged 3–9 years.
Methods
Four thousand four hundred thirty-five children were followed from 2019 to 2022. AL and corneal curvature were measured using an IOLMaster 500, while refractometry and visual acuity were also assessed. Baseline data included demographics and parental myopia status, with annual updates on height, weight and behavioural factors. Latent class growth model was used to discover AL trajectories, whereas multiple logistic regression was used to identify determinants of rapid AL elongation.
Results
For all participants, baseline age and parental myopia influenced AL growth. Specifically, children aged 3–6 years exhibited faster AL elongation when engaging in persistent excessive homework time (OR, 2.86, 95% CI 1.31 to 6.95) and near-work activities (OR, 2.13, 95% CI 1.12 to 4.10). For the 7–9-year group, the risk factors of rapid AL growth included being female (OR, 2.05, 95% CI 1.45 to 2.90) and need myopia correction at baseline (OR, 3.19, 95% CI 2.02 to 5.02). Notably, actively engaging in outdoor activities had a protective effect in the 7–9-year group (OR, 0.65, 95% CI 0.43 to 0.97).
Conclusions
This study delineates AL growth trajectories in children aged 3–9 years and highlights distinct risk factors for rapid AL growth. These findings underscore the necessity of implementing age-specific strategies for myopia prevention and control.
Keywords: Child health (paediatrics), Epidemiology, Risk Factors
WHAT IS ALREADY KNOWN ON THIS TOPIC
Axial length (AL) growth is a critical factor in the development of myopia in children, and AL growth thresholds for myopia onset have been investigated in several previous studies.
However, longitudinal studies on AL growth trajectories and associated risk factors remain limited.
WHAT THIS STUDY ADDS
For children aged 3–6 years, persistent excessive homework time and frequent near-work activities significantly increase the risk of rapid AL growth.
For children aged 7–9 years, being female and requiring myopia correction at baseline are key risk factors for rapid AL growth, while engaging in outdoor activities has a protective effect.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Personalised management plans emphasising the reduction of near-work activities and promoting outdoor activities can be developed.
Early interventions targeting children with identified risk factors may slow the progression of myopia and reduce the risk for high myopia development.
Introduction
The prevalence of myopia is increasing globally, projected to affect over half of the world’s population by 2050 if preventive measures are not implemented.1 This trend brings earlier onset, accelerated progression and heightened severity of myopia.2,4 High myopia, in particular, portends a surge in the overall global burden of the condition, including irreversible blindness.5 6 Myopia stems from excessive eye elongation, leading to blurred distance vision due to an axial length (AL) and ocular refraction system mismatch.7 Rapid, non-invasive AL measurements offer a valuable tool to identify myopia at-risk children, tracking AL growth is essential in addressing AL elongation-associated myopia development.8 9
Myopia epidemics appears to stem from environmental risk factors that are prevalent in modern, urbanised societies.10 In China, two major risk factors are inadequate time spent outdoors and increased engagement in near-work activities during childhood.11 12 Additionally, emerging behaviour-related risk factors like excessive screen time and less physical activity are also potentially associated with myopia progression.13 14 Identifying these modifiable behavioural factors would be instrumental in formulating targeted strategies to address and control myopia.
Moving forward, identifying appropriate candidates and timing for intervention will be critical in alleviating the global burden of myopia. To achieve this, extensive efforts have been directed towards understanding the myopiagenic factors and age-specific ‘prewarning value’ of AL.15,17 Nevertheless, due to limitations such as the cross-sectional nature and small sample sizes in longitudinal studies, the trajectories of AL growth and the influence of risk factors on these processes remain inconsistent and unclear.
To bridge the existing knowledge gaps, this study aimed to investigate the risk factors for rapid AL growth and myopia progression in children aged 3–9 years.
Methods
Study design
This prospective longitudinal cohort study was conducted in the Pujin Community, Minhang District, Shanghai, China, from 2019 to 2022. All children attending kindergarten and elementary school in this community are eligible for inclusion, allowing us to track children transitioning from preschool to elementary school during follow-up according to the ID number. This study was approved by the Ethics Committee of the Eye & ENT Hospital of Fudan University (approval codes 2019–022 and 2020-10-29) and adhered to the principles of the Declaration of Helsinki. Before the baseline examination, written informed consent was obtained from the parents or guardians of all participants. To obtain school-based perspective on myopia prevention and to reduce the influence of educational factors on AL elongation, we divided the cohort into two groups of 3–6 years (preschool children) and 7–9 years (low-grade pupils) in our study. Detailed inclusion and exclusion criteria are shown in online supplemental figure S1.
Ocular measurements
To ensure consistent follow-up and annual examination, all participants underwent regular physical and eye examinations, as well as a questionnaire survey, at the beginning of each school year in early September. The standardised questionnaire is shown in online supplemental table S1. Trained technicians conducted the baseline and annual school-based ocular examinations along with the annual health monitoring. Uncorrected visual acuity (UCVA) was recorded as a Snellen value and transformed into the logarithm of the minimum angle of resolution (logMAR) for analysis. Refraction measurements were obtained using an auto-refractometer (Cannon RF10; Cannon, Tokyo, Japan), with each eye of all participants evaluated at least three times without cycloplegia. The spherical equivalent (SE) was calculated as sphere power and half of cylinder power. Ocular biometry was performed using Zeiss IOLMaster 500 (Carl Zeiss Meditec, Jena, Germany). AL was measured three times in each eye, and the mean was used for statistical analysis. Three corneal curvature measurements (K1 and K2) were obtained and averaged to obtain the mean corneal curvature. The AL to corneal radius of curvature (AL/CR) ratio, which is closely related to the SE, was calculated by dividing AL (mm) by CR (mm). Both the auto-refractometer and biometer were calibrated before each examination. Owing to missed visits, data cleaning collapsed the data into the number of follow-up visits and intervals rather than the corresponding years. To account for missed visits, we calculated annual progression by determining the difference between the current year’s measurement and the most recent available measurement, normalised by the time interval. This method was applied consistently across all data types in the study to ensure robustness and inclusion of all available data. Owing to the collinearity between the eyes of the same individuals, only measurements from the right eye were included in the statistical analysis. Definitions and descriptions of these ocular parameters related to myopia progression are shown in online supplemental table S2.
Body mass index
Height measurements were obtained while participants were barefoot, and weight (kg) was recorded using calibrated standard weighing machines. BMI was calculated as weight (kg) divided by the square of height (m). Participants with a body mass index (BMI) exceeding two SD were classified as overweight, while others were classified as normal or underweight. Over 3 years, BMI and lifestyle changed three times. BMI transitions were categorised as persistently normal (3 years of normal or underweight), intermittently overweight (1–2 years of overweight) and persistently overweight (3 years of overweight).
Lifestyle factor translation
For time-related lifestyle assessments, the following formula was used: ((hours spent on a weekday)×5+ (hours spent on a weekend day)×2)/7. This method allowed the assessment of daily outdoor physical activity, homework, sleep, academic and exercise times. Following the current recommendations of lifestyle guidelines for Chinese children,18 this study categorised lifestyle factors based on daily life activities during follow-up. Those exceeding 2 hours of outdoor activities or exercise per day were categorised as actively engaging in outdoor activities or physical exercise, respectively. Excessive homework was defined as homework that surpassed the recommended duration. Sleeping <10 hour was deemed insufficient, while prolonged academic engagement denoted excessive academic time. Conversely, others were labelled as inactive, with normal homework, sufficient sleep, normal academic and insufficient exercise times.
Over 3 years, lifestyle changes were classified into three patterns based on the observed frequencies. Outdoor time was classified as inactive (no active outdoor activity), intermittently active (1–2 years of active outdoor activity) or persistently active (3 years of active outdoor activity). This classification was similarly applied to homework, sleep, academic and exercise times. Eye usage behaviours, including near-work activities, bad posture and screen use, were assessed as ‘never’, ‘sometimes’ and ‘always’. Those responding ‘never’ for 3 years were classified as having normal behaviours, ‘always’ for 3 years as persistent behaviour, and the rest as intermittent behaviour. More detailed variable information, processing methods and reference standards are found in online supplemental table S3.
Statistical analysis
We employed latent class growth models (LCGMs) to identify AL growth trajectories over 3 years, a semiparametric analysis that delineates groups of individuals based on their posterior probabilities of following similar trajectories on an outcome over time.19 We first estimated latent trajectories of AL growth in two groups (aged 3–6 and 7–9 years) by comparing the following 1–5-class models: linear change over time without within-class intercept variance, quadratic latent class growth analysis and quadratic change over time with within-class intercept variance.20 Model fit was determined using the Vuong-Lo-MendellRubin likelihood ratio test, Akaike information criteria, Bayesian information criterion (BIC) and a measure of entropy.21 The best-fitting number of trajectories was estimated based on a minimum BIC while ensuring posterior probabilities by class exceeded 0.70 and each class comprised at least 5% of the total population.22 Trajectories were named for ease of interpretation based on their graphical patterns.
In the analysis phase, we performed multiple interpolations for cases where AL measurements were recorded but lifestyle variables were missing. Variance inflation factor method was performed to check the collinearity between the variables (online supplemental table S4). χ2 and Student’s t-tests were used to analyse the differences in participant characteristics between the two age groups. Multivariate logistic regression analyses were performed to assess the associations between lifestyle factors and AL growth trajectories. ORs and corresponding 95% CIs were used to summarise the associations between participant characteristics, lifestyle changing patterns and AL growth trajectories. Repeated-measures analysis of variance (ANOVA) was used to assess UCVA, SE, corneal curvature, and AL/CR growth trends in different AL growth patterns over time. All statistical analyses were conducted using R software V.4.3.2. Statistical significance was set at p<0.05 (two-tailed).
Results
Participants’ characteristics and predictor variables
Table 1 shows the baseline survey data, including demographics and key variables, for 4435 individuals (average age: 6.32±1.71) included in this study. Among them, 51.6% were male, and the average age of initial pencil use was 4.66±1.13 years (range: 2–6 years). Notably, 87.3% did not wear corrective spectacles at baseline, and instances of low birth weight and preterm delivery were rare (3.8% and 6.0%, respectively). Approximately 45.2% reported having both parents with myopia. Significant differences were observed in age and need for corrective spectacles at baseline, age of pencil use and parental myopia status between the two age groups (all p<0.05). Online supplemental table S5 compared baseline information for included and excluded children, and while it showed significant differences on several variables, the main reason was the significantly older age of the excluded children, 9.16±3.31 years, and the absence of more than 70% of lifestyle variables. We believe that the main focus of our analysis—identifying risk factors for rapid AL growth—was unaffected.
Table 1. Children characteristics.
| All children | 3–6 years | 7–9 years | P value* | |
|---|---|---|---|---|
| Number (%) | 4435 (100) | 2622 (59.1) | 1813 (40.9) | |
| Age at baseline (mean±SD) | 6.32±1.71 | 5.18±1.13 |
7.98±0.83 |
<0.001 |
| Gender | 0.764 | |||
| Male | 2287 (51.6) | 1357 (51.9) | 930 (51.4) | |
| Female | 2148 (48.4) | 1265 (48.2) | 883 (48.7) | |
| Myopia correction need at baseline | <0.001 | |||
| Spectacles wearing | 256 (5.8) | 68 (2.6) | 188 (10.4) | |
| No need | 3871 (87.3) | 2404 (91.7) | 1467 (80.9) | |
| Unknown | 308 (6.9) | 150 (5.7) | 158 (8.7) | |
| Pencil age (mean±SD) | 4.66±1.13 | 4.25±0.96 | 5.26±1.10 | <0.001 |
| Birth weight | 0.556 | |||
| Low | 168 (3.8) | 103 (3.9) | 65 (3.6) | |
| Normal | 4267 (96.2) | 2519 (96.1) | 1748 (96.4) | |
| Preterm birth | 0.283 | |||
| Preterm delivery | 265 (6.0) | 169 (6.5) | 96 (5.2) | |
| Term delivery | 4170 (94.0) | 2453 (93.6) | 1717 (94.7) | |
| Parental myopia | 0.006 | |||
| Neither | 938 (21.1) | 521 (19.9) | 417 (23.0) | |
| Mother only | 844 (19.0) | 499 (19.0) | 345 (19.0) | |
| Father only | 649 (14.6) | 366 (14.0) | 283 (15.6) | |
| Both | 2004 (45.2) | 1236 (47.1) | 768 (42.4) | |
P values were calculated by Student's t-tests and χ2 analysis, and P < 0.05 was considered significant (marked in bold).
AL growth trajectories
Figure 1A illustrates the upward trajectory of AL growth from 2019 to 2022. Importantly, this trajectory displayed distinct patterns between participants aged 3–6 years and 7–9 years, as confirmed by repeated-measures ANOVA (p<0.001). A similar trend was observed for the AL/CR ratio (p<0.001; figure 1B). One to five class models were generated, all of which converged, and the best log-likelihood was replicated for each solution. Detailed information on the model estimation, selection criteria and concise summaries is found in online supplemental tables S6 and 7. Consequently, the two-trajectory model was selected for subsequent analyses across both age groups. Among 2622 participants aged 3–6 years, two distinct classes were identified based on their mean AL growth trajectories: ‘non-rapid growth’ (n=2429, 92.6%) and ‘rapid growth’ (n=193, 7.4%) (figure 1C). Similarly, among 1813 participants aged 7–9 years, classes were designated as ‘non-rapid growth’ (n=1565, 86.3%) and ‘rapid growth’ (n=248, 13.7%) (figure 1D).
Figure 1. Trends and trajectories in AL growth. (A–B) Splines depicted the growth trends of AL and AL/CR ratio during the follow-up years for all participants and between 3–6 years and 7–9 years age groups, respectively. P values were calculated by repeated-measure analysis of variance (ANOVA) to capture the developmental changes within each group across the 3 years, with significance denoted by ***, p<0.001. (C–D) Two distinct AL growth trajectories were identified as ‘non-rapid growth’ and ‘rapid growth’ in 3–6 years and 7–9 years age groups. AL/CR ratio, axial length to corneal radius of curvature ratio.
Three-year AL growth and mean growth per year
During the 3 years, the non-rapid AL growth group of 3–6-year-old children had AL measurements of 22.30±0.74 mm, 22.59±0.78 mm, 22.84±0.84 mm and 23.26±0.84 mm, with a mean AL growth per year of 0.28±0.13 mm. In contrast, the rapid AL growth group had AL measurements of 22.69±0.94 mm, 23.55±1.00 mm, 24.19±0.97 mm and 24.65±0.76 mm, with a mean AL growth per year of 0.69±0.10 mm. Among children aged 7–9 years, the non-rapid AL growth group had AL measurements of 23.11±0.79 mm, 23.43±0.83 mm, 23.75±0.86 mm and 23.90±0.84 mm, with a mean AL growth per year of 0.32±0.16 mm. In contrast, the rapid AL growth group had AL measurements of 24.42±1.07 mm, 25.11±0.87 mm, 25.54±0.80 mm and 25.38±0.81 mm, with a mean AL growth per year of 0.50±0.19 mm.
Multivariate analysis of associations between AL growth and risk factors
To further explore the independent factors influencing AL growth trajectories, we used multivariate logistic regression models while considering confounding variables. The findings are shown in tables2 3. For children aged 3–6 years, having a myopic father (2.33 (1.54 to 3.85), p<0.001), especially with both parents affected (3.22 (1.96 to 5.45), p<0.001), was associated with a higher risk of rapid AL growth. Additionally, persistent excessive homework (2.86 (1.31 to 6.95), p=0.013) and prolonged near-work activities (2.13 (1.12 to 4.10), p=0.022) were also linked to increased risk in this age group.
Table 2. Multivariate analysis of risk factors of 3-year AL growth trajectories among 3–6-year-old children.
| Variables* | Rapid growth (%)† | Non- rapid growth (%) |
OR (95% CI) | P value‡ |
|---|---|---|---|---|
| Age at baseline | ||||
| 3–4 | 51 (3.8) | 1296 (96.2) | 1 (ref.) | |
| 5 | 56 (10.2) | 492 (89.8) | 1.48 (0.92 to 2.36) | 0.104 |
| 6 | 86 (11.8) | 641 (88.2) | 1.99 (1.25 to 3.18) | 0.004 |
| Gender | ||||
| Male | 92 (7.3) | 1173 (92.7) | 1 (ref.) | |
| Female | 101 (7.4) | 1256 (92.6) | 0.99 (0.73 to 1.35) | 0.949 |
| Pencil age | ||||
| 2–3 | 37 (6.3) | 552 (93.7) | 1 (ref.) | |
| 4–5 | 134 (7.6) | 1621 (92.4) | 1.00 (0.66 to 1.53) | 0.997 |
| 6 | 22 (7.9) | 256 (92.1) | 0.68 (0.36 to 1.27) | 0.234 |
| Birth weight | ||||
| Normal | 187 (7.4) | 2332 (92.6) | 1 (ref.) | |
| Low | 6 (5.8) | 97 (94.2) | 0.73 (0.26 to 1.72) | 0.504 |
| Preterm birth | ||||
| Term delivery | 184 (7.4) | 2302 (92.6) | 1 (ref.) | |
| Early delivery | 9 (6.6) | 127 (93.4) | 0.97 (0.42 to 1.98) | 0.927 |
| Parental myopia | ||||
| Neither | 20 (3.8) | 501 (96.2) | 1 (ref.) | |
| Mother only | 26 (5.2) | 336 (91.8) | 1.18 (0.76 to 1.85) | 0.477 |
| Father only | 30 (8.2) | 473 (94.8) | 2.33 (1.54 to 3.85) | <0.001 |
| Both | 117 (9.5) | 1119 (90.5) | 3.22 (1.96 to 5.45) | <0.001 |
| BMI | ||||
| Persistently normal | 76 (5.2) | 1395 (94.8) | 1 (ref.) | |
| Intermittently overweight | 55 (7.4) | 688 (92.6) | 1.37 (0.94 to 1.98) | 0.097 |
| Persistently overweight | 59 (14.5) | 349 (85.5) | 1.43 (0.95 to 2.15) | 0.085 |
| Outdoor time | ||||
| Inactive | 131 (8.3) | 1446 (91.7) | 1 (ref.) | |
| Intermittently active | 53 (5.6) | 889 (94.4) | 0.73 (0.51 to 1.02) | 0.071 |
| Persistently active | 9 (8.7) | 94 (91.3) | 1.27 (0.56 to 2.59) | 0.531 |
| Homework time | ||||
| Persistently normal | 8 (3.2) | 245 (96.8) | 1 (ref.) | |
| Intermittently excessive | 52 (3.7) | 1371 (96.3) | 0.99 (0.47 to 2.32) | 0.971 |
| Persistently excessive | 133 (14.1) | 813 (85.9) | 2.86 (1.31 to 6.95) | 0.013 |
| Sleeping time | ||||
| Persistently sufficient | 63 (5.6) | 1062 (94.4) | 1 (ref.) | |
| Intermittently insufficient | 64 (6.2) | 969 (93.8) | 0.89 (0.61 to 1.30) | 0.552 |
| Persistently insufficient | 66 (14.2) | 398 (85.8) | 0.93 (0.60 to 1.45) | 0.752 |
| Academic time | ||||
| Persistently normal time | 50 (5.3) | 887 (94.7) | 1 (ref.) | |
| Intermittent excessive time | 49 (7.1) | 644 (92.9) | 1.19 (0.77 to 1.83) | 0.429 |
| Persistent excessive time | 94 (9.5) | 898 (90.5) | 1.35 (0.91 to 2.01) | 0.138 |
| Exercise time | ||||
| Persistently inactive | 45 (5.5) | 779 (94.5) | 1 (ref.) | |
| Intermittently active | 50 (6.9) | 676 (93.1) | 1.02 (0.65 to 1.59) | 0.937 |
| Persistently active | 98 (9.1) | 974 (90.9) | 0.81 (0.53 to 1.24) | 0.330 |
| Near-work activities | ||||
| Persistently normal | 28 (3.5) | 774 (96.5) | 1 (ref.) | |
| Intermittent near-work | 121 (7.6) | 1464 (92.4) | 1.40 (0.88 to 2.30) | 0.167 |
| Persistent near-work | 44 (18.7) | 191 (81.3) | 2.13 (1.12 to 4.10) | 0.022 |
| Bad posture | ||||
| Persistently normal | 49 (4.2) | 1125 (95.8) | 1 (ref.) | |
| Intermittent bad posture | 111 (8.8) | 1145 (91.2) | 1.30 (0.87 to 1.95) | 0.207 |
| Persistent bad posture | 33 (17.2) | 159 (82.8) | 1.42 (0.77 to 2.60) | 0.254 |
| Screen use | ||||
| Persistently normal | 44 (6.3) | 658 (93.7) | 1 (ref.) | |
| Intermittently frequent | 82 (6.8) | 1129 (93.2) | 0.92 (0.61 to 1.38) | 0.675 |
| Persistently frequent | 67 (9.4) | 642 (90.6) | 0.92 (0.59 to 1.44) | 0.704 |
The variable ‘Myopia correction need at baseline’ was excluded from the multivariate analysis for the 3–6 years group due to its low proportion in this age range. The model failed to converge when including this variable. Therefore, it was excluded to maintain the integrity of the analysis.
Rapid growth (%) or non-rapid growth (%) indicated the number and percentage of children classified as having rapid or non-rapid AL growth in the corresponding category, respectively.
P values were calculated by multivariable logistic regressions, and p < 0.05 was considered significant (marked in bold).
AL, axial length; BMI, body mass index.
Table 3. Multivariate analysis of risk factors of 3-year AL growth trajectories among 7–9-year-old children.
| Variable | Rapid growth (%)* | Non-rapid growth (%) | OR (95% CI) | P value† |
|---|---|---|---|---|
| Age at baseline | ||||
| 7 | 149 (13.6) | 945 (86.4) | 1 (ref.) | |
| 8 | 72 (15.0) | 407 (85.0) | 0.85 (0.58 to 1.24) | 0.395 |
| 9 | 27 (11.3) | 213 (88.8) | 0.47 (0.27 to 0.77) | 0.003 |
| Gender | ||||
| Male | 83 (9.4) | 800 (90.6) | 1 (ref.) | |
| Female | 165 (17.7) | 765 (82.3) | 2.05 (1.45 to 2.90) | <0.001 |
| Myopia correction need at baseline | ||||
| No need | 204 (12.0) | 1495 (88.0) | 1 (ref.) | |
| Spectacles wearing | 44 (38.6) | 70 (61.4) | 3.19 (2.02 to 5.02) | <0.001 |
| Pencil age | ||||
| 2–3 | 20 (14.6) | 117 (85.4) | 1 (ref.) | |
| 4–5 | 113 (13.6) | 719 (86.4) | 0.72 (0.37 to 1.46) | 0.339 |
| 6 | 115 (13.6) | 729 (86.4) | 0.52 (0.27 to 1.06) | 0.060 |
| Birth weight | ||||
| Normal | 240 (13.7) | 1508 (86.3) | 1 (ref.) | |
| Low | 8 (12.3) | 57 (87.7) | 1.16 (0.44 to 2.73) | 0.749 |
| Preterm birth | ||||
| Term delivery | 238 (13.7) | 1498 (86.3) | 1 (ref.) | |
| Early delivery | 10 (13.0) | 67 (87.0) | 0.64 (0.26 to 1.40) | 0.293 |
| Parental myopia | ||||
| Neither | 36 (8.6) | 381 (91.4) | 1 (ref.) | |
| Mother only | 39 (11.3) | 249 (88.0) | 1.31 (0.78 to 2.15) | 0.350 |
| Father only | 34 (12.0) | 306 (88.7) | 1.34 (0.80 to 2.24) | 0.315 |
| Both | 139 (18.1) | 629 (81.9) | 2.22 (1.41 to 3.70) | <0.001 |
| BMI | ||||
| Persistently normal | 150 (11.5) | 1151 (88.5) | 1 (ref.) | |
| Intermittently overweight | 55 (18.6) | 240 (81.4) | 1.50 (0.98 to 2.28) | 0.055 |
| Persistently overweight | 43 (19.8) | 174 (80.2) | 1.32 (0.84 to 2.06) | 0.224 |
| Outdoor time | ||||
| Inactive | 196 (14.8) | 1125 (85.2) | 1 (ref.) | |
| Intermittently active | 50 (10.7) | 417 (89.3) | 0.65 (0.43 to 0.97) | 0.039 |
| Persistently active | 2 (8.0) | 23 (92.0) | 0.28 (0.02 to 1.53) | 0.237 |
| Homework time | ||||
| Persistently normal | 30 (12.0) | 221 (88.0) | 1 (ref.) | |
| Intermittently excessive | 81 (9.3) | 792 (90.7) | 1.03 (0.52 to 2.11) | 0.933 |
| Persistently excessive | 137 (19.9) | 552 (80.1) | 1.30 (0.69 to 2.59) | 0.431 |
| Sleeping time | ||||
| Persistently sufficient | 50 (12.0) | 367 (88.0) | 1 (ref.) | |
| Intermittently insufficient | 93 (11.3) | 728 (88.7) | 0.97 (0.57 to 1.67) | 0.915 |
| Persistently insufficient | 105 (18.3) | 470 (81.7) | 0.73 (0.44 to 1.25) | 0.242 |
| Academic time | ||||
| Persistently normal time | 84 (13.2) | 551 (86.8) | 1 (ref.) | |
| Intermittent excessive time | 45 (10.2) | 398 (89.8) | 0.78 (0.48 to 1.27) | 0.329 |
| Persistent excessive time | 119 (16.2) | 616 (83.8) | 1.11 (0.74 to 1.66) | 0.613 |
| Exercise time | ||||
| Persistently inactive | 91 (12.7) | 624 (87.3) | 1 (ref.) | |
| Intermittently active | 64 (13.4) | 414 (86.6) | 1.17 (0.75 to 1.81) | 0.489 |
| Persistently active | 93 (15.0) | 527 (85.0) | 1.25 (0.84 to 1.87) | 0.268 |
| Near-work activities | ||||
| Persistently normal | 28 (8.8) | 292 (91.3) | 1 (ref.) | |
| Intermittent near-work | 153 (12.9) | 1033 (87.1) | 1.02 (0.56 to 1.92) | 0.941 |
| Persistent near-work | 67 (21.8) | 240 (78.2) | 1.63 (0.79 to 3.43) | 0.188 |
| Bad posture | ||||
| Persistently normal | 56 (11.5) | 429 (88.5) | 1 (ref.) | |
| Intermittent bad posture | 137 (13.2) | 904 (86.8) | 1.07 (0.67 to 1.73) | 0.782 |
| Persistent bad posture | 55 (19.2) | 232 (80.8) | 0.84 (0.46 to 1.56) | 0.585 |
| Screen use | ||||
| Persistently normal | 40 (10.4) | 344 (89.6) | 1 (ref.) | |
| Intermittently frequent | 114 (12.8) | 778 (87.2) | 1.16 (0.71 to 1.93) | 0.555 |
| Persistently frequent | 94 (17.5) | 443 (82.5) | 1.29 (0.77 to 2.19) | 0.345 |
Rapid growth (%) indicated the number and percentage of children classified as having rapid AL growth in the corresponding category
P values were calculated by multivariable logistic regressions, and P < 0.05 was considered significant (marked in bold)
AL, axial length; BMI, body mass index.
Conversely, for children aged 7–9 years, having both myopic parents increased the risk of rapid AL growth (2.22 (1.41 to 3.70), p<0.001), while myopia in either parent did not significantly impact AL growth. Furthermore, intermittent engagement in outdoor activities over 3 years (0.65 (0.43 to 0.97), p=0.039) reduced the risk. Age-specific comparisons indicated that 6-year olds faced higher risks compared with 3–4-year olds (1.99 (1.25 to 3.18), p=0.004), while 9-year olds had reduced risk compared with 7-year olds (0.47 (0.27 to 0.77), p=0.003). Among children aged 7–9 years, females (2.05 (1.45–2.90), p<0.001) and those needing myopia correction at baseline (3.19 (2.02 to 5.02), p<0.001) exhibited an elevated risk of rapid AL growth. As myopic eyes elongate faster than normal eyes, to adjust for the effects caused by the initial myopia status, we separately assessed risk factors for non-myopic eyes in the older age group (online supplemental table S8, multiple linear regression models could not converge due to few myopic children in this group at baseline). The results showed that baseline age (0.35 (0.18 to 0.63), p<0.001), female (1.85 (1.27 to 2.72), p=0.002), both myopic parents (2.15 (1.40 to 3.36), p<0.001) and intermittently actively engaged in outdoor activity (0.59 (0.37 to 0.92), p=0.022) were associated with different patterns of AL growth, indicating the importance of preventive strategies focusing on particular group of children and behaviours to mitigate the risk of myopia onset.
AL growth trajectories and myopia progression
It is important to elucidate the impact of diverse AL growth trajectories on the characterisation of refractive development in children. Based on the outcomes of repeated-measures ANOVA, the growth trajectories of UCVA, SE, corneal curvature and AL/CR ratio among children with rapid AL growth were significantly different from their non-rapid AL growth counterparts (all p<0.001) (figure 2). Specifically, within both age groups, children with a rapid AL growth trajectory experienced a more rapid decline in UCVA, a notable intensification in the degree of SE, a tendency towards flatter corneal curvature, and an increased AL/CR ratio.
Figure 2. Refractive development patterns linked to varying AL growth trajectories over time. (A–B) Curves represented variations in UCVA, spherical equivalent, corneal curvature and AL/CR ratio in 3–6 years and 7–9 years age groups. Error bars denote the SD of observed values. P values were calculated by repeated-measure analysis of variance (ANOVA) to capture the developmental changes within each group across the 3 years, with significance denoted by ***, p<0.001. AL/CR ratio, axial length to corneal radius of curvature ratio D, diopters; UCVA, uncorrected visual acuity.
Discussion
To better understand the lifestyle changes caused by environmental exposures and how they relate to the progression of myopia, we conducted a school-based cohort study of 4435 individuals over 3 years in Shanghai. Herein, we employed LCGMs to discern AL growth trajectories over 3 years, which is a semiparametric analysis that delineates groups of individuals, to examine AL growth patterns from a group rather than an individual perspective. We converted the complex lifestyle variables into patterns of change in behavioural styles according to the aforementioned criteria, further increasing the interpretability and clinical relevance of the analysis.
Notably, the prevalence of rapid AL growth doubled from 7.4% in the 3–6-year group to 13.7% in the 7–9-year group, likely reflecting environmental influences. The transition from preschool to elementary school increases academic demands, such as prolonged near-work and reduced outdoor activity, both established contributors to rapid AL elongation and myopia development.23 Understanding how risk factors affect AL growth is crucial for developing targeted interventions to mitigate rapid eye growth and myopia progression.
Educational factors are significant environmental factors associated with myopia, and they can have a considerable impact on other lifestyle aspects.24 To minimise the effects of educational factors and provide a school-based perspective for myopia prevention and control strategies, we categorised the patients into two groups: preschool children (aged 3–6 years) and low-grade pupils (aged 7–9 years). Interestingly, we found girls aged 7–9 were more prone to rapid AL growth than same-aged boys, aligning with East Asian studies.15 Nevertheless, whether sex truly influences myopia risk is uncertain due to factors like varying outdoor activities, near-work exposure and simultaneous adolescent and growth stages.25 Notably, the influence of age on rapid AL growth was found to differ between the two age groups. In younger children, being 6 years old emerged as a specific risk factor for rapid AL growth, potentially attributable to the transition from preschool to elementary school, which is associated with increased near work and reduced outdoor activity. As children grow older, the natural deceleration of AL growth becomes more pronounced, and advancing age within the 7–9 years group may reflect a shift towards more stable ocular development, serving as a protective factor.9 Our study links parental myopia to increased rapid AL growth risk in children aged 3–9 years, consistent with prior research.26 This study found that father-only myopia significantly increased the risk of rapid AL growth in children aged 3–6 years, but not in those aged 7–9 years. Prior research has established that having one myopic parent raises the child’s risk of myopia,27 but this study uniquely examined whether paternal or maternal myopia contributes differently. Larger studies are needed to clarify the extent of paternal or maternal myopia’s influence on childhood myopia development. For children aged 7–9 years, the need for myopia correction at baseline became an independent risk for accelerated AL growth, underlining the importance of interventions to manage myopia progression in early-onset cases.
In the present study, the lifestyle risk factors of rapid AL growth were homework time and near-work activities for children aged 3–6 years and outdoor time for those aged 7–9 years. These findings are consistent with those of a study by Wang et al, where children engaging in homework for over an hour on an average day exhibited higher AL compared with those with less than an hour of homework time.28 Prolonged near-work activities cause lagging eye regulation such that objects are imaged behind the retina, leading to hyperopic defocused growth and compensatory AL elongation.29 Our findings on the impact of near-work activities mirror a previous study that demonstrated a link between increased near-work time and the progression of AL (≥ 0.28 mm) and SE (≤ −0.5 D) over a 3-year follow-up period.30 A longitudinal cohort study in Australia also found that the effects of near work on myopia onset were greater for 6-year olds but not significant for 12-year olds, and the findings may be consistent across countries for younger children, which may suggest that accommodative lag may play an important role in 3-year to 6-year olds, and that early mitigation of near work is a viable strategy to help reduce early-onset myopia.31 Nonetheless, several studies have reported no significant association between myopia progression and near-work activities. This discrepancy might stem from variations in the age composition of the studied populations and potential measurement biases.32
Time spent outdoors is widely recognised as a crucial protective factor against myopia onset, but there are still controversies about its effectiveness in slowing progression. In this study, we found that intermittent active outdoor activity had a protective effect against rapid AL growth in children aged 7–9 years, but not in those aged 3–6 years. The protective effect was more pronounced in the older age group, likely due to increased exposure to myopia risk factors on entering primary school. In contrast, the younger group had a lower incidence of myopia, which could explain why the protective effect of outdoor activity against myopia onset, though present, was not statistically significant. Additionally, the lack of significance for persistently active outdoor activity could be due to the low proportion of persistently actively engaged in outdoor activities. Accounting for reduced outdoor activity during the COVID-19 pandemic, further research with a larger cohort is needed to better understand this association.33
According to our findings, children with rapid AL growth showed distinct refractive changes in both age groups, including poorer visual acuity, more severe myopia, flatter corneas and larger AL/CR ratios during follow-ups. Interestingly, in children aged 3–6 years, ocular biometric measurements at baseline were not significantly different between the groups with rapid and non-rapid AL growth, but significant differences were seen at follow-up, whereas children aged 7–9 years with rapid AL growth had significantly different ocular biometric measures at baseline. These findings suggest that children who experience rapid AL growth and myopia onset at the preschool age develop significant differences in their ocular biometric measures by the time they begin schooling. Therefore, our results suggest that the time window for interventions for myopia must be shifted forward to preschool age, with effective identification of those at risk, close follow-up and behavioural interventions to delay the onset and slow the progression of myopia.
This study has some limitations. First, the sample of 20 Shanghai kindergartens and primary schools may not represent the broader Chinese demographic, especially given Shanghai’s high myopia rates.23 Although our study reports on 4435 children selected from an initial cohort of 18 961, we recognise that there may be limitations in generalisability and that future validation with a large external cohort is needed. Second, not using cycloplegic refraction, the gold standard for myopia measurement might overestimate refraction development differences tied to distinct AL growth trajectories.34 Third, data on lifestyle transitions were obtained from annual parental questionnaires, introducing potential reporting and recall biases that could amplify noise and impede precise assessment. Quantitative measures such as wearable smart devices offer promise for refining findings in further studies.35 Fourth, the present study’s approach to dichotomising AL growth is data driven and clinically annotated, but other patterns of AL growth may still exist; however, this study’s first analysis from a trajectory perspective of AL growth can provide new insights. Fifth, it is noteworthy that the initial AL in the rapid growth group was longer than in the non-rapid growth group in both age groups. This difference may be due to an uneven age distribution in the cohort. Future studies could provide more refined results by including more younger age participants to optimise the cohort structure. Finally, we recognise that some participants’ school environments changed over time, which could potentially mitigate the original grouping based on baseline educational factors. However, this grouping still reflects the early exposure to distinct educational intensities, which is critical in understanding the rapid AL growth. Future studies with more appropriate stratifications may help address the school transitional issues.
This study provides novel insights into AL growth trajectories and the escalating prevalence of myopia in China. The findings establish the AL growth trajectories for preschool children and low-grade primary school students over a 3-year longitudinal study period. To address these trends, our study highlights persistent excessive homework and near-work activities as risk factors for rapid AL growth in preschool children. Additionally, among low-grade primary school students, baseline myopia correction requirements and female gender are recognised as risk factors, while engagement in outdoor activities acts as a protective factor against rapid AL growth. Further research with improved methodologies and broader participant representation is essential to confirm these associations.
Supplementary material
Acknowledgements
We would like to thank Editage (www.editage.cn) for English language editing.
Footnotes
Funding: This work was supported by National Key Research and Development Program of China (Grant Number 2024YFC2510805); National Natural Science Foundation of China (Grant Number 82271119, 82304260); Shanghai Rising-Star Program (Grant Number 23QA1401000); China Postdoctoral Science Foundation (Grant Number 2023M730626). Healthy Young Talents Project of Shanghai Municipal Health Commission (Grant Number 2022YQ015); Project of Shanghai Science and Technology (Grant Number 20410710100). The funding agencies had no role in study design, data collection and analysis, interpretation of data or writing the manuscript.
Provenance and peer review: Not commissioned; externally peer-reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study was approved by the Ethics Committee of the Eye & ENT Hospital of Fudan University (approval codes 2019-022 and 2020-10-29) and adhered to the principles of the Declaration of Helsinki. Participants gave informed consent to participate in the study before taking part.
Data availability statement
Data are available upon reasonable request.
References
- 1.Dolgin E. The myopia boom. Nature New Biol. 2015;519:276–8. doi: 10.1038/519276a. [DOI] [PubMed] [Google Scholar]
- 2.Chua SYL, Sabanayagam C, Cheung Y-B, et al. Age of onset of myopia predicts risk of high myopia in later childhood in myopic Singapore children. Ophthalmic Physiol Opt. 2016;36:388–94. doi: 10.1111/opo.12305. [DOI] [PubMed] [Google Scholar]
- 3.Pärssinen O, Kauppinen M. Risk factors for high myopia: a 22-year follow-up study from childhood to adulthood. Acta Ophthalmol. 2019;97:510–8. doi: 10.1111/aos.13964. [DOI] [PubMed] [Google Scholar]
- 4.Hu Y, Ding X, Guo X, et al. Association of Age at Myopia Onset With Risk of High Myopia in Adulthood in a 12-Year Follow-up of a Chinese Cohort. JAMA Ophthalmol. 2020;138:1129–34. doi: 10.1001/jamaophthalmol.2020.3451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Fricke TR, Jong M, Naidoo KS, et al. Global prevalence of visual impairment associated with myopic macular degeneration and temporal trends from 2000 through 2050: systematic review, meta-analysis and modelling. Br J Ophthalmol. 2018;102:855–62. doi: 10.1136/bjophthalmol-2017-311266. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Haarman AEG, Enthoven CA, Tideman JWL, et al. The Complications of Myopia: A Review and Meta-Analysis. Invest Ophthalmol Vis Sci. 2020;61:49. doi: 10.1167/iovs.61.4.49. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Jonas JB, Jonas RA, Bikbov MM, et al. Myopia: Histology, clinical features, and potential implications for the etiology of axial elongation. Prog Retin Eye Res. 2023;96:101156. doi: 10.1016/j.preteyeres.2022.101156. [DOI] [PubMed] [Google Scholar]
- 8.Xu L, Li Y, Wang S, et al. Characteristics of highly myopic eyes: the Beijing Eye Study. Ophthalmology. 2007;114:121–6. doi: 10.1016/j.ophtha.2006.05.071. [DOI] [PubMed] [Google Scholar]
- 9.Tideman JWL, Polling JR, Vingerling JR, et al. Axial length growth and the risk of developing myopia in European children. Acta Ophthalmol. 2018;96:301–9. doi: 10.1111/aos.13603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Pan CW, Ramamurthy D, Saw SM. Worldwide prevalence and risk factors for myopia. Ophthalmic Physiol Opt. 2012;32:3–16. doi: 10.1111/j.1475-1313.2011.00884.x. [DOI] [PubMed] [Google Scholar]
- 11.Lyu Y, Zhang H, Gong Y, et al. Prevalence of and factors associated with myopia in primary school students in the Chaoyang District of Beijing, China. Jpn J Ophthalmol. 2015;59:421–9. doi: 10.1007/s10384-015-0409-x. [DOI] [PubMed] [Google Scholar]
- 12.You QS, Wu LJ, Duan JL, et al. Factors associated with myopia in school children in China: the Beijing childhood eye study. PLoS One. 2012;7:e52668. doi: 10.1371/journal.pone.0052668. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Foreman J, Salim AT, Praveen A, et al. Association between digital smart device use and myopia: a systematic review and meta-analysis. Lancet Digit Health. 2021;3:e806–18. doi: 10.1016/S2589-7500(21)00135-7. [DOI] [PubMed] [Google Scholar]
- 14.Suhr Thykjaer A, Lundberg K, Grauslund J. Physical activity in relation to development and progression of myopia - a systematic review. Acta Ophthalmol. 2017;95:651–9. doi: 10.1111/aos.13316. [DOI] [PubMed] [Google Scholar]
- 15.Wang SK, Guo Y, Liao C, et al. Incidence of and Factors Associated With Myopia and High Myopia in Chinese Children, Based on Refraction Without Cycloplegia. JAMA Ophthalmol. 2018;136:1017–24. doi: 10.1001/jamaophthalmol.2018.2658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Li S-M, Wei S, Atchison DA, et al. Annual Incidences and Progressions of Myopia and High Myopia in Chinese Schoolchildren Based on a 5-Year Cohort Study. Invest Ophthalmol Vis Sci. 2022;63:8. doi: 10.1167/iovs.63.1.8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Liu S, He X, Wang J, et al. Association between axial length elongation and spherical equivalent progression in Chinese children and adolescents. Ophthalmic Physiol Opt. 2022;42:1133–40. doi: 10.1111/opo.13023. [DOI] [PubMed] [Google Scholar]
- 18.Ministry of Education of the People’s Republic of China The ministry of education and other three departments deploy prevention and control work for myopia at key stages in kindergartens and primary schools. [1-Dec-2024]. www.moe.gov.cn/jyb_xwfb/gzdt_gzdt/s5987/202411/t20241112_1162408.html Available. Accessed.
- 19.Kaup AR, Byers AL, Falvey C, et al. Trajectories of Depressive Symptoms in Older Adults and Risk of Dementia. JAMA Psychiatry. 2016;73:525–31. doi: 10.1001/jamapsychiatry.2016.0004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Paksarian D, Cui L, Angst J, et al. Latent Trajectories of Common Mental Health Disorder Risk Across 3 Decades of Adulthood in a Population-Based Cohort. JAMA Psychiatry. 2016;73:1023–31. doi: 10.1001/jamapsychiatry.2016.1921. [DOI] [PubMed] [Google Scholar]
- 21.Pierce M, McManus S, Hope H, et al. Mental health responses to the COVID-19 pandemic: a latent class trajectory analysis using longitudinal UK data. Lancet Psychiatry. 2021;8:610–9. doi: 10.1016/S2215-0366(21)00151-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Andruff H, Carraro N, Thompson A, et al. Latent Class Growth Modelling: A Tutorial. TQMP. 2009;5:11–24. doi: 10.20982/tqmp.05.1.p011. [DOI] [Google Scholar]
- 23.He X, Sankaridurg P, Xiong S, et al. Prevalence of myopia and high myopia, and the association with education: Shanghai Child and Adolescent Large-scale Eye Study (SCALE): a cross-sectional study. BMJ Open. 2021;11:e048450. doi: 10.1136/bmjopen-2020-048450. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Mountjoy E, Davies NM, Plotnikov D, et al. Education and myopia: assessing the direction of causality by mendelian randomisation. BMJ. 2018;361:k2022. doi: 10.1136/bmj.k2022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lyu IJ, Kim MH, Baek SY, et al. The Association Between Menarche and Myopia: Findings From the Korean National Health and Nutrition Examination, 2008-2012. Invest Ophthalmol Vis Sci. 2015;56:4712–8. doi: 10.1167/iovs.14-16262. [DOI] [PubMed] [Google Scholar]
- 26.Tideman JWL, Polling JR, Jaddoe VWV, et al. Environmental Risk Factors Can Reduce Axial Length Elongation and Myopia Incidence in 6- to 9-Year-Old Children. Ophthalmology. 2019;126:127–36. doi: 10.1016/j.ophtha.2018.06.029. [DOI] [PubMed] [Google Scholar]
- 27.Jiang X, Tarczy-Hornoch K, Cotter SA, et al. Association of Parental Myopia With Higher Risk of Myopia Among Multiethnic Children Before School Age. JAMA Ophthalmol. 2020;138:501–9. doi: 10.1001/jamaophthalmol.2020.0412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wang Y, Lin Y, Jiang D, et al. Differences in close-work activities and optical axis length between only children and non-only children: a cross-sectional study. BMC Pediatr. 2022;22:538. doi: 10.1186/s12887-022-03586-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Huang HM, Chang DST, Wu PC. The Association between Near Work Activities and Myopia in Children-A Systematic Review and Meta-Analysis. PLoS ONE. 2015;10:e0140419. doi: 10.1371/journal.pone.0140419. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Lanca C, Emamian MH, Wong YL, et al. Three-year change in refractive error and its risk factors: results from the Shahroud School Children Eye Cohort Study. Eye (Lond) 2023;37:1625–32. doi: 10.1038/s41433-022-02219-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.French AN, Morgan IG, Mitchell P, et al. Risk factors for incident myopia in Australian schoolchildren: the Sydney adolescent vascular and eye study. Ophthalmology. 2013;120:2100–8. doi: 10.1016/j.ophtha.2013.02.035. [DOI] [PubMed] [Google Scholar]
- 32.Li S-M, Li H, Li S-Y, et al. Time Outdoors and Myopia Progression Over 2 Years in Chinese Children: The Anyang Childhood Eye Study. Invest Ophthalmol Vis Sci. 2015;56:4734–40. doi: 10.1167/iovs.14-15474. [DOI] [PubMed] [Google Scholar]
- 33.Wang J, Li Y, Musch DC, et al. Progression of Myopia in School-Aged Children After COVID-19 Home Confinement. JAMA Ophthalmol. 2021;139:293–300. doi: 10.1001/jamaophthalmol.2020.6239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Hashemi H, Khabazkhoob M, Asharlous A, et al. Cycloplegic autorefraction versus subjective refraction: the Tehran Eye Study. Br J Ophthalmol. 2016;100:1122–7. doi: 10.1136/bjophthalmol-2015-307871. [DOI] [PubMed] [Google Scholar]
- 35.Ramamurthy D, Samuel Paulraj AK, Lakshmi T, et al. Development of smart spectacles to monitor and modify myopia-related health behaviour in children. Ophthalmic Physiol Opt. 2023;43:517–24. doi: 10.1111/opo.13119. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available upon reasonable request.