Prevalence and associated factors of myopia among rural school students in Chia-Yi, Taiwan

We recruited students of 33 elementary schools and 3 junior high schools from July 2014 to June 2015 in Chia-Yi, Taiwan, for this cross-sectional study. In total, 5417 participants underwent surveys of refraction and eye health as well as associated life style investigations. We included two cohorts, namely the age groups of 7–12 years and 13–15 years, that comprised students from elementary and junior high schools, respectively. We contacted the local administration of the Education and School Board to request their cooperation. The Institutional Review Board of Chang Gung Foundation (102-4827B) approved our study, and we followed the tenets of the Declaration of Helsinki. Written consent was obtained from the parents or guardians of all children before the study.

We used an examination environment setting of a classroom with blackout curtains for all the participants. The same investigating team administered all survey questionnaires and examinations to reduce individual errors. The body weight and height were measured. Gender, sleep duration and eye related symptoms was obtained through questionnaire. Noncycloplegic refractive errors were assessed using an auto-refractor (Autorefractometer, ARK-1, Nidek Co.,LTD., Aichi, Japan) [9]. Children presenting with a history of ocular and physical pathology, strabismus, and amblyopia were excluded. We obtained this medical data through questionnaire from children’s parents. (Supplementary doc. 1) The definition of refractive errors was based on spherical equivalent (SE) refraction calculated as the spherical dioptre plus one half of the cylindrical dioptre. An SE of ≤ − 1.0 D in one or both eyes was defined as myopia and categorised into low myopia (from − 1.0 D to − 6.0 D) and high myopia (≤ − 6.0 D). We evaluated ocular alignment using corneal reflections observations (Hirschberg test), followed by the monocular cover–uncover test. Briefly, one eye of the participant was covered, and the examiner looked for any movement in the opposite uncovered eye. Tropia was indicated by such movement. While heterophoria was indicated in case of movement of the covered eye when the cover was applied and movement in the opposite direction (a fusional movement) as the cover was removed. If the patient had heterophoria, the eye would be straight before and after the cover–uncover test. The deviation appeared during the test because of the interruption of binocular vision. We included both exophoria and esophoria in the horizontal heterophoria group for further analysis. We analysed the risk factors and prevalence of myopia and high myopia for both children with and without myopia and presented the adjusted odds ratio values for all risk factors.

Despite the implementation of the myopia prevention programme, the prevalence of myopia remained high in the rural prefecture of Taiwan. Horizontal heterophoria was an associated factor for myopia in children in both elementary and junior high school groups, with ORs of 2.37 and 2.20, respectively. It was also associated with high myopia, with ORs of 9.79 and 6.36 for children in elementary and junior high school groups, respectively. Rose et al. and Dirani et al. suggested that outdoor activity is an independent factor negatively associated with myopia [10, 11]. Thus, a strategy based on these findings can be developed to prevent myopia and high myopia [12]. Although the Ministry of Education in Taiwan launched a project using these very strategies in 2009, our study demonstrated that the prevalence of myopia remained high. This may be possible because of children being increasingly exposed to computers, communication and electronic consumer products, and after-school tutoring classes in Taiwan. Meanwhile, controversy exists whether near work is associated with myopia [13‐15].

Our results demonstrated an increase in the prevalence of myopia with increasing age, which is compatible with the results in the literature [16]. The literature also shows that the annual incidence of myopia increased with age from 19.1% at aged 7 years to 30.2% at aged 11 years [16]. The increase in the prevalence with age would not be solely accumulation of myopic children, but also the increasing new myopia with age [16]. Therefore, more effort is needed in preventing new myopia as well as delaying the onset of myopia.

The association between myopia prevalence and horizontal heterophoria was further investigated in our study. Heterophoria was previously suggested to be associated with myopia in pre-school and elementary school children [17‐20]. We demonstrated a close link between horizontal heterophoria and myopia. Children with heterophoria may progress to myopia though convergence accommodation [21‐24]. Reduction of concomitant accommodation has been found to delay the progression of moderate myopia [25]. However, because we were unable to determine whether heterophoria is a risk factor for myopia or vice versa, further study is needed to clarify the link between myopia and heterophoria [18].

This study has several limitations. First, we used noncycloplegic refraction because this was a population-based screening study, and cycloplegia was not well accepted by children and their parents. Choong et al. reported that autorefractors had a tendency towards minus overcorrection under noncycloplegic conditions [26]. Therefore, we used − 1.0 D to define myopia. The noncycloplegic autorefraction test was more convenient for vision-screening than the best-corrected visual acuity test and cycloplegic refraction test [9]. Second, we only evaluated a rural region in Taiwan. The prevalence of myopia is usually higher in urban areas than in rural areas. However, our results showed a high prevalence of myopia in the rural region, which warrants further evaluation. The study still retained sufficient statistical power to evaluate the presence of various risk factors for myopia given the rather large sample size. Finally, because our measurements were performed at a single point in time, they may not reflect long-term exposure to important demographic or biochemical factors.