Epidemiology characteristics for non-myopic Chinese children aged 6 to 12 years in Jiangsu Province

Background: We aimed to present epidemiology characteristics including SE value, age, BMI, sex for non-myopic Chinese children after indicating the prevalence of myopia among children aged six to twelve . Based on these we were trying to develop alert values for predicting the future onset of myopia. Methods: All students took part in the ophthalmic examination lled in a questionnaire to provide basic demographic information. We used an autorefractor applying with cycloplegia to obtain spherical equivalent value, and part of their parents lled in a questionnaire of factors related to myopic among children. Results: We nally had 3900 non-myopic observations from 6362 students, and the prevalence of myopia is 38.0% for boys and 39.5% for girls respectively. The average values for SE were 0.5±0.7 for boys and 0.6±0.8 for girls respectively. The mean SE decreased with age in children, and the value of height and BMI took on a stable trend. Alert values for myopia could be set as follows, for children aged six years of age, 0.4-0.6 D for boys and 0.8-1.0 D for girls respectively. For children aged seven years of age, 0.4-0.6 D for boys and 0.3-0.5 D for girls respectively, for children aged eight years, 0.2-0.4 D for boys and 0.3-0.5D for girls, for children aged nine years, 0.2-0.4 D for boys and 0.3-0.4D for girls, 0.1-0.3 D for boys aged ten and 0.3D for girls aged ten years, 0.10-0.3 D for boys aged eleven and -0.3-0.3 D for girls aged eleven years, and for children aged twelve ,-0.1-0.1 D for boys and -0.3-0.1 D for girls aged twelve years. Father’s myopia ( OR:1.22,95%CI:1.01-1.48 ), near work on weekends (OR:2.56,95%CI:1.17-5.61) and outdoor


Background
There is a dramatic increase in the prevalence of myopia in Eastern Asia. In China, a great increase was also seen in the young generation, indicating the importance of prediction of early-onset myopia among juveniles [1][2][3]. When it comes to the intervention of myopia, people often pay more attention to treatment such as optical or pharmaceutical methods to slow down eye growth, and thus retard the progression of myopia [4]. However, from a public health perspective, it is more desirable for non-myopic students to develop an early warning comprehensive system to predict and prevent the onset of myopia.
Holden et al stated the future development trends of myopia from a macroscopic view, predicting that nearly half of the world's population may be myopic by 2050, with as much as 10% having highly myopic [4]. However, the prediction of myopia on an individual level is urgent and relevant research is limited. Karla Zadnik et al noted that spherical equivalent (SE) refractive error is the best single predictor of future myopia, comparing to other factors such as parental myopia, near work, and outdoor activities [6]. In Beijing three-year follow-up eye study, researchers found that children aged 6 to 7 years showed signi cant SE decrease, AL increase, CCT thickening, ACD deepen, LT thinning, and AL/CR increase. These ndings may be an early warning signal of myopia development [7]. BMI is a good indicator of risk, growth, and childhood related disease such as obesity, elevated blood pressure, and so on [8,9]. Obese children aged 7 to 9 years would more likely to have poor visions comparing to those without obesity, and such in uence could last to 12 years of age among boys [10]. A nomogram or nomograph is a form of line chart showing scales for all variables involved in a formula, it is a rule functioned as a simple calculator [11]. Currently, there is a growing application of the nomogram model using rational risk factors in predicting the probability of occurrence, progression or prognosis of an individual's disease [12][13][14]. Therefore, it is essential to predict the myopia onset based on factors such as sex, age, BMI, and SE.
This study aimed to explore the epidemiology characteristics, including age, BMI, sex, and SE value, among non-myopic Chinese children. Alert values were proposed to predict future myopia and the associations between these values and factors associated with myopia were addressed.

Study sites and populations
This study was based on the program "Surveillance for common disease and health risk factors among students, sub-program: ophthalmological investigation" during the 2018-2019 academic year in Jiangsu Province. We enrolled 26,461 students from 12 regions in Jiangsu Province (Supplement gure1) in this program, and A total of 6,363 students were aged from 6 to 12.5, participated in sub-program: ophthalmological physical examination.

Data collection and ethics statement
Myopia was de ned as -0.50 diopters(D) in the worse eye and the worse eye was de ned as the eye with the greater absolute value of refractive error (spherical equivalent). All students took part in this subprogram were required to provide basic demographic information including name, sex, regional, and some of their parents lled in questionnaires concerning myopic related questions including family, near work, sleep duration and outdoor activities. An autorefractor (Topcon RM-8900 or KR-800; Topcon Co., Tokyo, Japan) was applied with cycloplegia, and the spherical equivalent of the refractive error was calculated as the spherical value of refractive error plus one half of the cylindrical value.
The study protocol was approved by the Institutional Review Board of Ethics committee of Jiangsu Province, and detailed information can be found in the previous article [10]. We used an autorefractor with cycloplegia under parents' informed consent.

Statistical analysis
Statistical analyses were performed with R software (www.R-project. Org, version 3.5.3) with additional rms package [15] and Microsoft o ce ware. The age-speci c spherical equivalent values were calculated for the percentiles of 5 th , 10 th , 25 th ,50 th ,75 th ,90 th and 95 th , and age-speci c BMI were calculated for the percentiles of 25 th , 50 th ,75 th for boys and girls. We then performed a logistic model to select variables t for nomogram model. Nomograph was drawn by R software with rms packages. Continuous variables were presented as the mean with standard deviation (SD).

Baseline characteristics of this study
The prevalence of myopia for children aged 6 to 12 was 38.0% for boys and 39.5% for girls. We eventually obtained 3,900 non-myopic students, and the ratio of male to female was 1.16. The mean SE decreased with age in children, and the value of height and BMI took on a stable trend. (Table1) Myopic boys or girls had higher values of BMI or height than these of non-myopic boys or girls. (Supplement table1) Alert values of spherical equivalent value for non-myopic students according to percentiles of BMI and SE Table2 showed the 25 th , 50 th , 75 th percentiles for BMI value by age (6-12.5) for both sexes, and presented two different 5 th , 10 th , 25 th ,50 th ,75 th ,90 th and 95 th percentiles for SE value. Figure1 showed a comparison of two alert values for non-myopic students, and both of them indicated the same trend. We set cut-off points as: for children aged

Nomograph for predicting childhood myopia onset
We selected 50 th percentiles for both SE and BMI to create a nomogram model. Total points ranged from 64 to 93 corresponding to risk probability of future myopia onset in half-year ranged from 0% to 100%. When SE value ranged from -0.6 to 2, the corresponding points ranged from 100 to 0. Observation's age increased (from 6 to 12.5) with the corresponding points decreased (25 to 0). Sex seemed to have little impact on future myopia onset: male and female are corresponding to 0 and4 points respectively.
Relationship between BMI and points can be presented as 8-0 points, 12-1 points, 16-2 points, 20-3 points, 24-4 points,and 30-5 points. (Figure2) Relationship between myopic alerting for non-myopic students and myopia associated factors Non-myopic students with alerting had a higher proportion of their myopic father. (P 0.05) A family which had more than one kid is a protective factor for non-myopic students to prevent the onset of myopia. After school homework especially on weekends had higher OR values for students with no myopia. (OR: OR:2.56,95%CI:1.17-5.61). Sleep and outdoor activities also had more impact on non-myopic students with alerting. (Table3, Figure3) Discussion This is the rst study using age, BMI, and sex to propose alert values and predict myopia onset among children aged 6 to 12 years in the world. It is also the rst time using a nomogram model to predict risk of myopia onset among children. Moreover, we rstly addressed the epidemiology characteristics for nonmyopic Chinese children aged 6 to 12 years in Jiangsu Province, China.
The prevalence of myopia among children aged 6 to 12 years was 38.0% for boys and 39.5% for girls, which was higher than Chinese adults. The prevalence of myopia for de nitions of SE of <−0. Heredity, outdoor activities, and near work had a great in uence on the onset and progression of myopia [19]. In this study, these factors also had a signi cant impact on non-myopic children.
Nomograms may be valuable tools to estimate the likelihood of diseases in the future [18]. Based on alert values, we built up a nomogram model to give warnings to children with alerting.
There were some limitations in this study. First, cut-off points were built based on a cross-sectional study.
Therefore, a long-term cohort study is needed to improve the accuracy of this study. Second, higher sensitive factors associated with myopia required further analysis for better forecasting.

Conclusion
This study presented the epidemiology description among non-myopic students in Jiangsu Province, China. A series of alert values were proposed to provide early warning reference for Chinese children aged 6 to 12 years. Heredity, near work, and outdoor activities had an impact on non-myopic students with myopic alerting.
14. Mazouni C, Spyratos F, Romain S, et al: A nomogram to predict individual prognosis in node-negative   Nomogram for predicting childhood myopia onset Figure 3 Forest graph of relationship between alert values for non-myopic students and myopic factors

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. supplement.pdf