In East Asia, the prevalence of myopia in the native population is very high [1, 2]. However, governmental intervention in both Taiwan and Singapore to control the progression of myopia in school-aged children is proving successful [3,4,5]. In Japan, data regarding uncorrected visual acuity (VA) (UCVA) in the annual school healthcare checkup has been the only statistical data on the prevalence of myopia in school children. We have been conducting the ‘Kyoto Childhood Refractive Error Study’ (KRES), a study on the RE of elementary and junior high school children in the Prefecture of Kyoto, Japan [7,8,9] since 2013, and in 2021, a national survey on refractive errors (RE) was initiated in Japan [6].

There are several reports on gender-based differences of myopic prevalence in the published literature. In the 1990s, it was reported that myopia was more common in adult males than females [10], and recently it has been reported that it is more common in girls [11,12,13,14] and, conversely that there is no difference in prevalence between boys and girls [15], so it is likely that not only genetic but also environmental factors are involved. To date, various environmental factors, such as education, near-vision work time, and outdoor activity time, have been pointed out as risk factors for the progression of myopia [16,17,18,19,20].

Since there are gender-related differences in infant growth curves, it is likely that there are also gender-related differences in eye development. Physical growth in school-aged children is closely related to growth hormones, yet after puberty, growth becomes more closely related to sex hormones. It is reported that girls, whose sex hormones are more likely to be secreted early, grow taller first [21]. Another report shows that school-aged children with earlier peak height velocities experienced peak axial length (AL) and spherical equivalent velocities earlier than children in whom puberty occured later [22], thus clearly illustrating the biological differences between males and females. Moreover, it is demonstrated in the United States, Europe, Australia, and China that AL is longer and corneal keratometry is flatter in males than in females [15, 23,24,25,26], and in many studies, the growth curves of AL have been evaluated in a gender-specific manner.

The purpose of this present study was to investigate and analyze gender-based differences in myopia prevalence in native Japanese school-aged children via the examination of RE, AL, corneal keratomery, and higher-order aberrations (HOAs), as well as via a questionnaire on environmental factors that may possibly impact the myopia progression.

The protocols of this observational study were approved by the Ethics Review Committee of Kyoto Prefectural University of Medicine, Kyoto, Japan (Approval No. RBMR-E-467-5). This study was conducted in accordance with the tenets set forth in the Declaration of Helsinki, and followed the guidelines of the Ethics Review Committee, an explanatory document including an option to not participate was presented to all children prior to their involvement in the study.

The participants in this study were students attending two elementary/junior high schools; Kyoto-City Ryofu Gakuen (School A) and Fukuchiyama-City Yakuno Gakuen (School B), these two institutions of learning were selected as being typical public schools by the Boards of Education of Kyoto City and Kyoto Prefecture.

In School A, the study population in the first year (2013) consisted of 6- to 9-year-old students enrolled in elementary school Grades 1 through 3. Participants in the second year of the study were 6- to 10-year-old students enrolled in elementary school Grades 1 through 4. Participants in the third year of the study were 6- to 11-year-old students enrolled in elementary school Grades 1 through 5. Participants in the fourth year of the study were 6- to 12-year-old students enrolled in elementary school Grades 1 through 6. The number of grades was then gradually increased. From the seventh through ninth year of the study (i.e., from 2020 through 2022), the study population consisted of 6- to 15-year-old students enrolled in elementary school Grades 1 through 6 and junior high school Grades 7 through 9, respectively. In School B, the first year of the examinations was 2014, and all participants were enrolled in the same manner as described above for School A (Online Resource 1).

Since it was conducted as a school health check, students in the support class were also examined as much as possible.

For the annual ocular examinations at both schools, the instruments used for the examination were brought to the school, and all examinations were performed by accredited ophthalmologists or nationally certified orthoptists.

In all participants, monocular distance UCVA and best-corrected VA (BCVA) examinations were performed, first in the right eye and then in the left eye with a Landolt ring at 5-meter distance. The findings were converted to Logarithm of the Minimum Angle of resolution (LogMAR) VA.

Objective RE and pupil distance were measured using the WR-5100K open-field binocular auto refractometer/keratometer (Grand Seiko) set to a 5-meter distance index, and subjective RE was examined by orthoptists. If distance UCVA was better than LogMAR 0, subjective RE was set to 0D. When there was a difference between objective and subjective RE in a single participant, objective RE was measured again for confirmation.

AL was measured with an IOL Master® 500 Optical Biometer (Carl Zeiss Meditec AG). HOAs, corneal keratometry, and pupil diameter were measured using a KR-1W Corneal Wavefront Analyzer (Topcon Corporation). The ocular and corneal HOAs were analyzed at a pupil diameter of 4.0mm and 6.0mm, and the root-mean-square (RMS) value from the third- to sixth-order of Zernike coefficients were calculated. From those coefficients, corneal and ocular spherical aberrations, coma-like aberrations, spherical-like aberrations, and total aberrations were calculated.

Each year, the parents of the children were instructed to complete a questionnaire form on the following lifestyle factors; (1) time spent on outdoor activities, (2) time spent reading, (3) time spent playing mobile-phone-app games, and (4) time spent watching television, as well as a question regarding whether or not the parents were myopic (spectacles or contact lens wearers). The time for each activity was divided into four stages (< 1 hour, < 2 hours, < 3 hours, and ≥ 3 hours). Moreover, beginning in 2016 a question regarding height and weight was added to the questionnaire.

Myopia was defined as a subjective spherical equivalent (SE) of less than or equal to -0.50 diopters (D), in addition to a UCVA of more than LogMAR 0 [27, 28]. High myopia was defined as a subjective SE of less than or equal to -6.00D.

For the statistical analysis, UCVA, BCVA, subjective and objective SE, keratometry (D), AL, AL (mm)/corneal radius (mm) ratio (AL/CR ratio), pupil diameter, pupil distance, HOAs, and questionnaire were investigated. The following data were excluded from the statistics analysis as outliers; 0 mm of AL, more than 1.0 μ RMS value of HOAs analyzed at 4 mm, and 2.0 μ RMS value of HOAs analyzed at 6 mm. Data of both eyes were used for the statistical analysis. A linear mixed model or generalized linear mixed model with participant as a random effect and gender as a fixed effect was used to compare the differences in mean or difference in proportion, respectively, of each factor. To account for intra-subject correlation arising from the use of both eyes, robust standard errors (Huber-White sandwich estimator) were used to provide valid inference despite the non-independence of observations [29]. In addition to the usual estimates, standardized estimates were calculated that adjusted for differences in the number of students in each grade as a sensitivity analysis.

Regarding HOAs it might be affected by the corneal refractive power. Thus, the association between HOAs and keratometry were investigated via linear mixed models. Data with missing values were not included in the analysis (complete case analysis). In all analyses, a P-value of < 0.05 (two-sided) was considered statistically significant. All data analysis was performed using the SAS version 9.4 (SAS Institute Inc.).