We analyzed data from 14,482 individuals from the TMM BirThree Cohort Study aged 4 to 19 years to determine the distribution of their AL. Of note, the data from 8,889 individuals aged 4 to 6 years was particularly valuable, as such a large data set has not been previously reported on a global scale. A segmented regression analysis identified that the break point for mean AL was approximately at the age of 12 years. Before this age, AL tended to elongate (β = 0.27), while after it, the elongation rate slowed (β = 0.12). Furthermore, when analyzed separately by sex, the break point was identified as occurring at approximately 10 years of age in boys and 16 years in girls. AL elongation was steeper in boys (β = 0.29) than in girls (β = 0.25). We determined the proportion by age of individuals with an AL ≥ 24.5 mm and ≥ 26 mm, which began to increase from ages 8 and 10 respectively. The proportion of individuals with LAL was higher among boys than girls between the ages of 6 and 11. After age 12, no significant sex differences were observed in the proportion of individuals with LAL.

The AL of the 4- to 6-year-old children was 22.14 ± 0.73 mm (among 8,899 children) in our study, significantly lower (p < 0.001) than the 22.35 mm ± 0.67 mm (among 457 children) reported by Matsumura et al. in Japan [16]. Chen et al. report that among East Asians and non-East Asians aged 4.0 to 6.9 years, AL was 22.40 (95% CI: 22.32-22.48) mm and 22.31 (95% CI: 22.16-22.47) mm, respectively [22]. Compared to these reports, we found that the AL of 4- to 6-year-old children was shorter than non-East Asians or East Asians. AL at ages 4 to 6 is thought to be influenced by genetic factors such as ethnicity and body size. However, even within the same Japanese population, significant differences in AL have been observed. One possible explanation is that the study by Matsumura et al. was conducted among preschool children in urban Kanagawa Prefecture, whereas the present study was conducted in Miyagi and Iwate prefectures, which are relatively rural. This environmental difference may have contributed to the variation in AL [23, 24]. Moreover, we compared mean, median, and quartile AL at 6, 9, and 15 years old with values for European children reported by Tideman et al. [11], as well as with values for Chinese children reported by Sans Diez et al. [12]. As shown in the Supplemental Table, we found significant differences in mean AL among the 3 groups at ages 6 and 9. However, at age 15, no significant difference was found between the Japanese and Chinese groups. Among the 3 groups, the Chinese participants had the longest AL, followed by the Japanese and the European children at ages 6 and 9. At age 15, AL in the Japanese and Chinese groups was similar, and both were longer than the European group. The median AL of the European children aligned with the 25th percentile of AL in the Japanese and Chinese distributions. Differences in AL among ethnic groups are observed at age 6, suggesting the influence of genetic factors [25]. By age 15, the AL of Chinese and Japanese participants had become comparable, indicating that lifestyle and environmental factors may also play a role [25].

The segmented regression analysis of the slope representing the speed of elongation showed that the break point for AL by age occurred at approximately 12 years. Before this age, the AL tended to elongate (β = 0.27), while after this age, the elongation rate slowed (β = 0.12). A review by Chen Y. et al. found that children’s AL grew at a rate of 0.3 to 0.4 mm per year between the ages of 3 and 5 [22], and Chen D.Z. et al. report that AL elongated at an average length of 0.80 mm at the ages of 3 to 6 years [26]. Similar to our results, a report by Fabian et al. targeting emmetropic eyes, found that the rate of AL elongation after the age of 12 was less than the reproducibility value (± 0.04 mm) of the AL measuring device [27].

In our data, the average AL of boys was longer than the average AL of girls from 4 to 19 years of age, except at the age of 18 years. Boys had significantly longer AL than girls between the ages of 4 and 12. Previous reports show that among adults, men have a longer AL than women [19, 28]. One study reports that there was a significant difference in mean AL between men and women among both those with myopia (p < 0.001) and those with hyperopia (p = 0.025) [29]. This is also supported by another study [30]. In a previous report on sex differences in AL in children under 7 years of age, both in children aged 0.0 to 1.9 years and in 2.0 to 3.9 years, AL was not significantly different between boys and girls (p = 0.85 and 0.37, respectively). In older children (4.0-6.9 years), boys were reported to have a longer AL than girls (boys: 22.49 mm, 95% CI 22.34-22.64, girls: 21.99 mm, 95% CI 21.82-22.15; p < 0.001) [22]. Recent studies report that it is possible to determine sex from fundus photographs in adults [31, 32], and Yamashita et al. note that boys with more masculine features in the fundus tend to have significantly longer ALs [33]. It is reported that the accuracy of sex identification from fundus images was low in younger elementary-school-aged children but significantly increased with age [34]. Specifically, the accuracy was recorded at 56.3% at 8.5 years, 46.1% at 9.5 years, 65.5% at 10.5 years, and 73.1% at 11.5 years [35]. It is not clear why boys have a longer AL than girls. One possible explanation could be that boys generally have a larger body size, greater height, and larger ocular volume [15, 28]. Segmented regression analysis revealed a difference in the break point of AL elongation speed between boys and girls. In boys, AL elongation was steep and tended to slow by the age of 10, while in girls, AL continued to elongate gradually until the age of 16. The reason for this sex difference in AL elongation patterns remains unclear, but differences in choroidal blood flow and hormones are considered possible factors. The choroid plays an important role in the regulation of eye growth and the development of myopia [36, 37]. Choroidal vascularity and choriocapillaris blood perfusion are lower in the more myopic eyes [38]. Choroidal circulation changes follow the menstrual cycle [39]. This may underlie the difference in break points for mean AL between the sexes and the tendency for AL to continue to extend in girls. To understand the reasons for sex differences in AL, further investigation is needed into ocular morphology as well as ocular volume, height, blood perfusion, and hormone levels.

In this study, we found that the right-eye AL was longer than the left-eye AL in participants aged 4, 5, and 8 years, a similar finding as previous studies. Several reports find that the right eye is more myopic [19, 40, 41]. The reason why the right eye is more predisposed to myopia is unknown, but it may be related to differences in the blood supply to the eye or hand dominance. Mansour et al. report that 93% of subjects in their study were right-handed and that SE was significantly less in their right eyes than their left eyes, but that there was no correlation between hand dominance and SE [41]. The right common carotid artery may have lower blood flow because of anatomical differences [42], which may be related to lower choroidal blood flow and a longer AL in the right eye [36, 38].

In this study, we calculated the age-specific proportions of participants with AL ≥ 24.5 mm and ≥ 26 mm, defined as LAL, which were considered indicative of suspected myopia. Among children aged 4 to 7 years, the prevalence of participants with AL ≥ 24.5 mm was quite low; however, a few children with longer AL were still observed, and these individuals were highly suggestive of having already reached the stage of high myopia. The proportions AL ≥ 24.5mm began to increase from age 8, continued to rise with age, and reached approximately half of the population by age 13 and older. The proportion of children with AL ≥ 26 mm, associated with significantly increased risk of serious complications later in life, increased particularly between ages 10 and 12, and about 10% were observed among those aged 13 years and older. When comparing boys and girls, we found that the proportion with LAL was significantly higher among boys aged 7 to 11 years. After the age of 12, no gender differences were observed in the prevalence of LAL. These results suggest that 8 years may be the age at which the onset of LAL occurs, and that many children are at risk of developing high myopia by the time they reach age 10. The higher prevalence of LAL in boys compared to girls before age 11, and the absence of sex differences after age 12, may be attributed to differences in the rate of axial elongation, as suggested by the results of the segmented regression analysis. According to a report by the International Myopia Institute [5], myopia is less common in children younger than 6 years, and even in East Asia and Singapore, where the prevalence of adult myopia is high, most studies show that myopia is less common in children younger than 6 years; the prevalence has been shown to be less than 5% [43,44,45,46]. According to a survey of myopia by Matsumura et al. based on an auto reflex meter, myopia increased from the ages of 7 to 8 in Japan [16]. Our results are similar to previous reports. Chua et al. observed schoolchildren aged 7 to 9 until age 11 and show that the age at onset of myopia was the strongest predictor of high myopia [47]. With high myopia, 85% of children were likely to have developed myopia before age 7. Olavi et al. report that 32% of children who received their first prescription for eyeglasses between the ages of 8.8 and 12.8 years developed high myopia in adulthood [48]. These reports and our current data indicate that AL before age 8 plays an important role in predicting future myopia.

As this study encompassed data collected before, during, and after the COVID-19 lockdowns in Japan—periods during which excessive AL elongation among preschool children has been widely reported [49,50,51,52]. The potential influence of the COVID-19 lockdown on AL should also be considered. When comparing age-specific AL before and after the nationwide lockdown in Japan, which took place around April 2020, a statistically significant difference was observed only in the right eyes of 13-year-old participants, whereas no significant differences were found in other age groups. In 13-year-olds, the mean AL before the lockdown was 24.13 ± 1.05 mm (n = 81), and after the lockdown it was 24.81 ± 1.12mm (n = 35), with a p-value of 0.002.

There are concerns that many children are at risk of visual impairment in the future. Since the progression of myopia is irreversible, it is important to detect high-risk cases early and link them to treatment. The results of this study can be used as a reference for determining standard values for AL at each age and sex and predicting future myopia. It may be useful for designing appropriate myopia-prevention programs [53, 54]. Early onset and prolonged axial elongation may increase the risk of developing high myopia in the future. Therefore, a longitudinal approach that includes follow-up over time will be necessary. We plan to continue to observe this cohort over the long term, and we expect to obtain further valuable findings.

The present study has several limitations. Refractive power directly indicates the degree of myopia, but in this study, refractive test results were not included in the analysis. In addition to AL, various factors, such as the corneal radius of curvature, anterior chamber depth, and crystalline lens thickness determine the refractive power of the eye, but these test results were not included in the results of this study. Due to the characteristics of the cohort enrollment, measurements were more frequently obtained at ages 4 and 8. The data for children aged > 8 years are derived only from sibling category of the TMM BirThree Cohort Study. A total of 2,801 participants provided two measurements. Therefore, the age distribution is skewed, which may have influenced the analysis. Potential biases due to the unbalanced age distribution include reduced precision of estimates in the older age group because of the small sample size, distortion of age-related trends caused by the over-representation of younger age group, and limited generalizability of the findings to older children and adolescents. In addition, subtle changes in AL could remain undetected owing to the insufficient number of cases in that age range. Follow-up AL measurements have not yet been performed with the participants and therefore changes associated with growth cannot be determined. Furthermore, the number of participants in older age groups was small; therefore, the data from these groups should be interpreted carefully.

This study was the first large-scale AL survey in Japan and reveals the age and sex related distribution of AL and the proportion with myopia among Japanese children and adolescents aged 4 to 19 years.