Current Prevalence of Myopia and Association of Myopia With Environmental Factors Among Schoolchildren in Japan

Key Points

Question  What are the prevalence of and factors associated with myopia?

Findings  In this cross-sectional study of 1478 schoolchildren at 2 schools in Tokyo, Japan, the prevalence of myopia was 76.5% among those aged 6 to 11 years and 94.9% among those aged 12 to 14 years, and the prevalence of high myopia was 4.0% among those aged 6 to 11 years and 11.3% among those aged 12 to 14 years.

Meaning  These findings suggest that prevalence rates of myopia under noncycloplegic refractive conditions are high among Japanese schoolchildren.

Importance  Given the estimates of increasing prevalence of myopia, especially in Asia, it is important to determine the current prevalence of myopia among populations of schoolchildren in Japan.

Objective  To investigate the current prevalence rate of myopia and the association between environmental factors and myopia in Japanese schoolchildren.

Design, Setting, and Participants  This cross-sectional study assessed 1478 participants, including 726 elementary school students and 752 junior high school students, at 2 schools in Tokyo, Japan, who underwent eye examinations from April 1 to May 31, 2017, that included measurement of the refractive errors by autorefractometry with noncycloplegic refraction and ocular biometric factors. After excluding those who had been treated with atropine or orthokeratology (n = 11), had a history of eye disease (n = 2), had no parental consent (n = 41), and were absent (n = 8), 1416 schoolchildren were analyzed.

Main Outcomes and Measures  The primary outcome was the prevalence of myopia and high myopia. Secondary outcomes were environmental factors that were associated with myopia.

Results  A total of 1416 schoolchildren (mean [SD] age, 10.8 [2.7] years; 792 [55.9%] male) were studied. The prevalence rates of myopia (spherical equivalent ≤−0.5 diopters [D]) were 76.5% (95% CI, 73.4%-79.7%) among the elementary school students and 94.9% (95% CI, 93.3%-96.5%) among the junior high school students. The prevalence rates of high myopia (spherical equivalent ≤−6.0 D) were 4.0% (95% CI, 2.5%-5.4%) among the elementary school students and 11.3% (95% CI, 8.8%-13.7%) among the junior high school students. The prevalence rates of high myopia classified based on axial length of 26.0 mm or longer were 1.2% (95% CI, 0.4%-2.0%) among elementary school students and 15.2% (95% CI, 12.5%-17.8%) among junior high school students. Multiple regression analysis showed that higher-order aberrations and dry eye disease were associated with refractive error in elementary school students (spherical aberration: β = 6.152; 95% CI, 3.161-9.143; P < .001; dry eye disease: β = −0.626; 95% CI, −1.189 to −0.063; P = .03) and with axial length in junior high school students (spherical-like aberration: β = 26.546; 95% CI, 18.708-34.385; P < .001; dry eye disease: β = 0.354; 95% CI, 0.131-0.578; P < .002).

Conclusions and Relevance  Although the use of noncycloplegic autorefraction with a cutoff of −0.50 D could lead to overestimation of results, these findings suggest that the current prevalence rates of myopia among elementary and junior high school students in Asia are high, especially if the results from these 2 schools are generalizable to all schoolchildren in Japan and Asia.

The incidence of myopia has been increasing worldwide, especially in East Asia, during the past 60 years.¹ The incidence of high myopia, which can be sight-threatening in later life, is also increasing worldwide.² A previous study² stated that the population with myopia will be approximately 5 billion in 2050.

One study³ suggested that increased time spent on near-vision tasks is a possible risk factor for myopia. Outdoor activity is also considered to be one of the most prominent environmental factors associated with suppressing myopia progression.^3-6 Although the environment might be associated with the recent increase in myopia,¹ no epidemiologic biometry study of myopia has been conducted among Japanese schoolchildren since 1999 to our knowledge.⁷ Matsumura and Hirai⁷ reported that the prevalence of myopia in Japan in children at 12 years of age was 43.5% from 1989 to 1991. Some other studies^8-11 have reported an association between higher-order aberrations and myopia.

The current study was the first large-scale study, to our knowledge, performed in 21st-century Japan to investigate the current prevalence rates of myopia and high myopia among students. We also analyzed questionnaire results regarding the association between myopia and environmental factors.

The Keio University School of Medicine Ethics Committee approved this cross-sectional study, which adhered to the tenets of the Declaration of Helsinki.¹² The parents of the participating students provided written informed consent.

A total of 1478 students participated, including 726 elementary school students and 752 junior high school students. Both schools are in Tokyo, Japan; however, the junior high school is a private school affiliated with Keio University and the elementary school is a public school. The examinations were conducted from April 1 to May 31, 2017, in each school when the students had their annual medical checkups. Eighty-eight junior high school students wore contact lenses at the time of the eye examination. Their refractive data were excluded when we analyzed the spherical equivalent and the prevalence of myopia. Their axial lengths were calculated after we subtracted 0.1 mm from the measured value according to a previous study.¹³ When determining the prevalence of myopia, these contact lens wearers were included as having myopia except for 5 students who used hyperopic contact lenses. We also excluded data from students who kept their eyelids open using a measurer (n = 10), who were measured manually (n = 9), or who wore contact lenses (n = 88) when we calculated the higher-order aberrations. Parental consent was obtained from 1437 of the 1478 students (97.2%). Eight students were absent on the examination day, and 1429 students ultimately were examined. After excluding those who had been treated with atropine (n = 4) or who wore orthokeratology contact lenses (n = 7) and had a history of eye disease, such as cataract or morning glory syndrome (n = 2), 1416 (95.8%) had available data for this study (eFigure 1 in the Supplement).

All participants underwent eye examinations that included measurement of the refractive errors by autorefractometry with noncycloplegic refraction and ocular biometric factors, such as central corneal thickness, corneal curvature, anterior chamber depth, lens thickness, axial length, and higher-order aberrations. We also recorded the height, weight, and body mass index of the participants.

Refraction, ocular biometric factors, and higher-order aberrations were measured in a noncycloplegic state using the HOYA iTrace Surgical Workstation (Tracey Technologies). The axial length was measured by noncontact optical biometry using a swept-source optical coherence tomography biometer (IOLMaster 700; Carl Zeiss Meditec AG). We recorded the axial length 10 times and averaged the data. The axial length–corneal curvature radius ratio was calculated because it can be a useful marker of myopia onset and progression.^14,15 Refractions become myopic theoretically when the axial length–corneal curvature radius ratio exceeds approximately 3.0.^14,15 The ocular higher-order aberrations were evaluated with a 3-mm pupillary diameter, and the root mean square values from the third- to sixth-order Zernike coefficients were calculated. From these Zernike coefficients, comalike aberration, spherical-like aberration, fifth-order and sixth-order aberrations, and the total higher-order aberration were calculated.

The students and their parents completed a questionnaire about the students’ lifestyles that included factors such as time spent on outdoor activities, near-work, and sleeping; symptoms or history of dry eye; and parental history of myopia. The time spent outdoors was defined as the sum of outdoor leisure and sports. The mean number of hours spent outdoors daily was calculated using the following formula¹⁶: [(hours spent on a weekday) × 5 + (hours spent on a weekend day) × 2]/7.

We set the maximal time spent on outdoor activities as 180 minutes daily because the association between outdoor activities and suppression of myopia onset and progression were expected to plateau at that point according to a meta-analysis and systematic review.¹⁷ Near-work included doing school homework, reading books, using a computer or smartphone, watching television, and playing video or portable games. We also asked to estimate the time spent doing near-work per day and the reading distance. The symptoms or diagnosis of dry eye disease were recorded using the Schaumberg questionnaire¹⁸ in Japanese, which was translated by our group.^19,20 We also interviewed the students and their parents about the students’ physical activities using the short version of the International Physical Activity Questionnaire in Japanese.²¹

We defined myopia as a spherical equivalent of −0.5 diopters (D) or less because Flitcroft et al²² reported that almost 90% of epidemiologic studies of myopia used less than −0.50 D or −0.50 D or less as the definition of myopia. High myopia was defined as a spherical equivalent of −6.0 D or less or an axial length of 26.0 mm or greater. Only the results from the right eye are presented.

We examined the associations between outcomes (axial length, spherical equivalent, and axial length–corneal curvature radius) and other factors, including age, sex, body mass index, physical activity (evaluated using the International Physical Activity Questionnaire), dry eye disease, time spent outdoors, time spent on near-work, reading distance, time spent sleeping, parental myopia, and higher-order aberrations (spherical aberration, comalike aberration, spherical-like aberration, and total higher-order aberration) using univariate regression analysis. We then performed multiple regression analysis. Age and sex were evaluated by forced entry, and the other factors were evaluated by stepwise analysis because age and sex have already been reported as factors associated with myopia.²³

We performed the χ² test with Bonferroni correction to compare the prevalence rates of myopia between 6-year-old students and students in other grades in elementary school and between 12-year-old students and students in other grades in junior high school. All statistical analyses were performed using SPSS software, version 24.0 (IBM Inc). All P values were 2-sided. P < .05 was considered to be statistically significant. Multicollinearity was not a factor.

A total of 1416 schoolchildren (mean [SD] age, 10.8 [2.7] years; 792 [55.9%] male) were studied. The mean (SD) spherical equivalent was −2.40 (2.23) D, and the mean (SD) axial length was 24.09 (1.30) mm for all participants. Among elementary school students, the mean (SD) spherical equivalent was −1.73 (1.98) D and the mean (SD) axial length was 23.41 (1.03) mm. Among junior high school students, the mean (SD) spherical equivalent was −3.09 (2.26) D and the mean (SD) axial length was 24.73 (1.19) mm (Table 1).

Table 2 gives the distributions of the ages of the participants and the prevalence rates of myopia in each age range. The prevalence of myopia (spherical equivalent ≤−0.5 D) was 76.5% (95% CI, 73.4%-79.7%) among all elementary school students. The prevalence rates of myopia between 6-year-old students and 8- to 11- year-old students in elementary school differed (prevalence of 63.1% in 6-year-olds, 78.6% in 8-year-olds, 79.5% in 9-year-olds, 85.3% in 10-year-olds, and 83.0% in 11-year-olds; P < .001). The prevalence of high myopia (spherical equivalent ≤−6.0 D) was 4.0% (95% CI, 2.5%-5.4%) among all students and 9.8% among those aged 10 years (eFigure 2 in the Supplement). On the basis of the axial length, the prevalence of high myopia (axial length ≥26.0 mm) was 1.2% (95% CI, 0.4%-2.0%) among all students (eFigure 3 in the Supplement).

Among the junior high school students, the overall prevalence of myopia was 94.9% (95% CI, 93.3%-96.5%), whereas the prevalence rates in each junior high school grade did not differ from each other (eFigure 4 in the Supplement). The prevalence of high myopia (spherical equivalent ≤−6.0 D) was 11.3% (95% CI, 8.8%-13.7%) for students who attended the examination without contact lenses. The prevalence of high myopia based on axial length was 15.2% (95% CI, 12.5%-17.8%) among all students (eFigure 5 in the Supplement). Table 3 gives the distributions of the spherical equivalents of the participants.

Multiple regression analysis performed to estimate the association between myopia and environmental factors found that among elementary school students, the axial length was associated with age (β = 0.303; 95% CI, 0.261-0.344; P < .001), sex (β = 0.427; 95% CI, 0.287-0.567; P < .001), and parental myopia (1 parent: β = 0.218; 95% CI, 0.001-0.435; P = .049; both parents: β = 0.570; 95% CI, 0.358-0.781; P < .001). The spherical equivalent was associated with age (β = −0.245; 95% CI, −0.336 to −0.154; P < .001), sex (β = 0.309; 95% CI, 0.005-0.612; P = .046), and dry eye disease (β = −0.626; 95% CI, −1.189 to −0.063; P = .03). The axial length–corneal curvature radius was associated with age (β = 0.033; 95% CI, 0.028-0.038; P < .001), dry eye disease (β = 0.033; 95% CI, 0.001-0.064; P = .04), and parental myopia (both parents: β = 0.051; 95% CI, 0.024-0.077; P < .001) (Table 4).

Among the junior high school students, the axial length was associated with age (β = 0.128; 95% CI, 0.028-0.229; P = .01), sex (β = 0.819; 95% CI, 0.650-0.988; P < .001), dry eye disease (β = 0.354; 95% CI, 0.131-0.578; P = .002), and parental myopia (1 parent: β = 0.714; 95% CI, 0.449-0.978; P < .001; both parents: β = 1.157; 95% CI, 0.893-1.422; P < .001). The spherical equivalent was associated with sex (β = −0.529; 95% CI, −0.892 to −0.166; P = .004), reading distance (β = 0.025; 95% CI, 0.002-0.047; P = .03), and parental myopia (1 parent: β = −1.097; 95% CI, −1.629 to −0.565; P < .001; both parents: β = −1.880; 95% CI, −2.416 to −1.345; P < .001). The axial length–corneal curvature radius was associated with sex (β = 0.068; 95% CI, 0.047-0.089; P < .001), physical activity (β = −0.000026; 95% CI, −0.000048 to −0.000004; P = .02), reading distance (β = −0.001; 95% CI, −0.003 to −0.0002; P = .02), and parental myopia (1 parent: β = 0.068; 95% CI, 0.036-0.100; P < .001; both parents: β = 0.112; 95% CI, 0.080-0.144; P < .001) (Table 4).

Multiple regression analysis was performed to estimate the association between myopia and higher-order aberrations among the elementary school and junior high school students (Table 5). We also performed the same multiple regression analysis among students who had high myopia (spherical equivalent ≤−6.0 D or axial length ≥26.0 mm); however, no significant factor was identified.

Our data found high prevalence rates of myopia (76.5% in the elementary school and 94.9% in the junior high school students) in Tokyo. Furthermore, the prevalence rates of high myopia were also high. In particular, the 15.2% (axial length–based) prevalence rate in the junior high school was high. Although the use of noncycloplegic autorefraction with a cutoff of −0.50 D could overestimate results, these findings suggest that the current prevalence rates of myopia among elementary and junior high school students in Asia are high.

The current myopia prevalence rate in Tokyo is one of the highest compared with data obtained previously from other Japanese cities, Singapore, and Taiwan.²⁴ Only 2 previous reports^7,25 on Japanese schoolchildren are available to date. A study²⁵ published in 1977 reported that the prevalence of myopia was 12.5% among elementary school students and 37.6% among junior high school students. In the other study,⁷ which included only 12-year-old individuals during the period 1989 to 1991, the prevalence of myopia (spherical equivalent <−0.5 D) was 43.5%. A systematic review²⁴ reported that the prevalence of myopia (spherical equivalent <−0.5 D) in 15-year-old individuals was 86.2% in Singapore and 80.0% in Taiwan. In the current study, the prevalence of myopia in 14-year-old individuals was 94.2%. The current result might be higher than the actual prevalence rate because we did not use cycloplegics and there were more boys than girls in the junior high school. In any case, this rate is unprecedentedly high.

The scenario was similar for high myopia. Among the third-year junior high school students in Beijing, the prevalence rates were 3.96% in 2005 and 6.69% in 2015.²⁶ In Singapore, the prevalence was 7.3% in 12- to 16-year-old individuals.²⁷ Although the age range is not exactly the same, the high myopia data from the current junior high school students in Tokyo were higher than those results.

A possible reason for such high prevalence rates for myopia is the decrease in the amount of time spent outdoors. It is well known that the myopia prevalence tends to decrease when individuals spend more than 2 hours daily engaged in outdoor activities, according to the studies of Rose et al⁴ and Jones et al.⁵ In the current study, the mean daily time spent outdoors was not associated with myopia in multiple regression analysis. A previous study²⁸ also reported that the time spent outdoors was not associated with the axial length among third-grade students in an elementary school in Kagoshima in rural Japan, and 87.7% of these students reported that their daily time spent outdoors was less than 1 hour. The reason that the time spent outdoors was not associated with myopia in the current study may have been the lack of variance in refractive error and time spent outdoors. All these results suggest that less than 2 hours of time spent outdoors may not have a sufficient association with suppression of myopia and the axial length.

In the current study, the axial length but not the spherical equivalent was associated with parental myopia in elementary school students. This finding is consistent with several previous studies^28,29 that reported that the axial length was associated with parental myopia. We assumed that the spherical equivalent was not associated with parental myopia only in elementary school students because emmetropization still continues at that age range.³⁰ We believe that this is the reason why there was no association with spherical equivalent and parental myopia only in elementary school students.

Among elementary school students, boys had longer axial lengths and a lower degree of myopia. The reason for this discrepancy is unknown. Some previous studies^23,31 have reported that boys had longer axial lengths than girls, as in the current study. The spherical equivalent may not match the axial length completely because the refraction is composed of other factors, such as the cornea and lens.

In the current study, we evaluated only 2 schools in Tokyo because the background, including the height, weight, and visual acuity, of the students in the current study was almost equivalent to the general Japanese schoolchildren’s data provided by the government. These results suggest that the prevalence of myopia among the Japanese schoolchildren may be high. However, the junior high school students probably had a more intensive educational environment, with more near-work and less time outdoors, because the school is a private school affiliated with a university. Furthermore, we assumed that children attending this school were from a higher social class and had higher educational attainment and potentially more parental myopia.

The association between higher-order aberrations and myopia attracted attention^8-11 because of the assumption that in patients with high degrees of higher-order aberrations there might be a blurred retinal image that eventually can lead to further development of myopia.¹⁰ A previous study¹¹ reviewed that it was unclear whether the optical aberrations occurred before or after myopia onset. Furthermore, increased higher-order aberrations are also associated with dry eye disease.³² Tear film instability in the central corneal regions caused by dry eye disease is associated with decreased stability and resultant increases of the higher-order aberrations. We are interested to know whether there is an association between dry eye disease and myopia. The current results were noteworthy (ie, the greater the symptoms of dry eye, the higher the myopic refraction was in the elementary school students and the longer the axial length was in the junior high school students). Our data suggest the possibility that there could be an association between dry eye disease and myopia. Thus, we would like to evaluate the association between these conditions in future studies.

The current study has some limitations. First, the current study was not completely reflective of the population in Japan because students in only 2 schools in Tokyo were evaluated. Second, we did not use cycloplegics because we conducted the current study during the annual school medical checkups. However, the cycloplegic autorefraction is the criterion standard based on some studies and the white paper presented by the International Myopia Institute.^22,33-36 The panel consensus was that cycloplegic autorefraction was the preferred method to determine the amount of myopia.²² Considering that, there is a possibility that the prevalence of myopia reported in the current study was overestimated. Meanwhile, we measured the axial length in the current study, and it was unaffected regardless of whether cycloplegia was used, which is the strength of the current study. Third, 88 junior high school students wore contact lenses. Although the number of contact lens wearers did not affect the prevalence of myopia, we excluded them from the evaluation of the refraction. Considering that, the mean refraction in junior high school would be underestimated. Furthermore, another 11 elementary and junior high school students were excluded because of previous myopia treatment. This exclusion may have led to underestimation of the refraction, axial length, and myopia prevalence. Fourth, we speculated that there is an association between myopia and dry eye disease, although there was substantial ascertainment bias because children with myopia were more likely to see eye care practitioners frequently and thus were more likely to be diagnosed with dry eye. Fifth, there was recall bias because the lifestyle data were collected from a questionnaire. Sixth, multiplicity was a possibility in the current study because we analyzed the axial length and refraction data in both elementary school and junior high school students.

The current study of Japanese schoolchildren suggests that there may be a high prevalence of myopia in Tokyo. Although the use of noncycloplegic autorefraction with a cutoff of −0.50 D could lead to overestimation of the results, these findings suggest that the current prevalence rates of myopia among elementary and junior high school students in Asia are high, especially if the results from these 2 schools are generalizable to all schoolchildren in Japan and Asia.

Accepted for Publication: May 26, 2019.

Corresponding Author: Toshihide Kurihara, MD, PhD, Laboratory of Photobiology, Department of Ophthalmology (kurihara@z8.keio.jp), and Kazuo Tsubota, MD, PhD, Department of Ophthalmology (tsubota@z3.keio.jp), Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan.

Published Online: August 15, 2019. doi:10.1001/jamaophthalmol.2019.3103

Author Contributions: Drs Yotsukura and Torii contributed equally to this work and share dual first authorship. Drs Yotsukura and Torii had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Yotsukura, Torii, Tokumura, Uchino, Kurihara, Tsubota.

Acquisition, analysis, or interpretation of data: Yotsukura, Torii, Inokuchi, Uchino, Nakamura, Hyodo, Mori, Jiang, Ikeda, Kondo, Negishi.

Drafting of the manuscript: Yotsukura, Torii, Inokuchi, Tokumura, Uchino.

Critical revision of the manuscript for important intellectual content: Yotsukura, Torii, Nakamura, Hyodo, Mori, Jiang, Ikeda, Kondo, Negishi, Kurihara, Tsubota.

Statistical analysis: Yotsukura.

Obtained funding: Torii.

Administrative, technical, or material support: Yotsukura, Torii, Inokuchi, Uchino, Nakamura, Hyodo, Mori, Jiang, Ikeda, Kondo, Negishi, Kurihara.

Supervision: Torii, Tokumura, Uchino, Negishi, Kurihara, Tsubota.

Conflict of Interest Disclosures: Dr Torii reported receiving grants from the Japan Society for the Promotion of Science and grants from Takeda Science Foundation during the conduct of the study. Dr Kondo reported receiving personal fees from Tsubota Laboratory Inc. outside the submitted work. Dr Kurihara reported receiving personal fees from Tsubota Laboratory Inc and Bayer Yakuhin Ltd. outside the submitted work and receiving grants from Fuji Xerox Co Ltd, Kowa Company Ltd, Novartis Pharmaceuticals Japan, Santen Pharmaceutical Co Ltd and Rohto Pharmaceutical Co Ltd. Dr Tsubota reported receiving a grant from Japan Society for the Promotion of Science during the conduct of the study and serving as chief executive officer for Tsubota Laboratory; having a patent issued for Myopia Prevention Article; having a patent issued for Irradiation Device; and having a patent pending for Optical Member (for violet light transmitting eyeglass lens), outside the submitted work. No other disclosures were reported.

Funding/Support: This work was supported by grant 18K16934 from the Japan Society for the Promotion of Science and Takeda Science Foundation (Osaka, Japan).

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.

Additional Contributions: We acknowledge the contributions of the following individuals to this study: Mitsuyo Tsubota, the principal of the elementary school; Kazue Sasaki, the school nurse of the elementary school; Toshihiro Abe, the Parent-Teacher Association officer; and Yasuko Kunegi, the school nurse of the junior high school. Yusaku Katada, MD, Yukihiro Miwa, MD, Hiromitsu Kunimi, MD, Mari Ibuki, MD, Nobuhiro Ozawa, MD, Yangsong Wang, MD, Yumi Hagiwara, MD, Chiho Shoda, MD, Ayako Ishida, MD, Keiko Takahashi, MD, Yuki Hidaka, MD, Tomoki Kurihara, MD, Saho Murakami, MD, Natsumi Takizawa, MD, Taiichirou Katayama, MD, Maho Sato, MD, Sho Okuyama, MD, Yuta Shigeno, MD, Kaoru Kobayashi, MD, Takumi Ariga, MD, Yuka Aonuma, MD, and Honami Nakayama, MD, assisted with the measurements; Yoko Arita assisted with sorting out the data; Yasunori Sato, PhD, Ryota Ishii, PhD, and Ryo Takemura, PhD, assisted with statistics; and Lynda Charters edited the manuscript for English. Lynda Charters was paid for her editing services. Toshihiro Abe, Yuta Shigeno, Kaoru Kobayashi, Takumi Ariga, Yuka Aonuma, and Honami Nakayama were paid for their work as orthoptists.