Prevalence and risk factors of myopia in Han and Yugur older adults in Gansu, China: a cross-sectional study

Few studies have investigated the prevalence of myopia in Northwest China. This cross-sectional study aimed to investigate the prevalence and associated factors of myopia and high myopia in adults aged 40–80 years in the Han and Yugur populations living in Gansu Province, Northwest China. A total of 3,845 participants were included. The overall age- and sex-adjusted prevalence of myopia (spherical equivalent (SE) < −0.5 D), high myopia (SE < −6.0 D) and hyperopia (SE > + 0.5 D) were 16.4%, 0.7% and 26.2% in Yugur participants, respectively, and 34.3%, 5.0% and 19.2% in Han participants, respectively. The prevalence of myopia and high myopia in Han participants was significantly higher than that in Yugur participants (both P < 0.001). Yugur population, birth in rural areas, smoking history and outdoor work were found to be negatively associated with myopia. Higher education level and a family history of myopia were found to be positively associated with myopia in the study population. High myopia was negatively associated with Yugur population, aging, birth in rural areas and was positively associated with a family history of myopia. This study provided valuable information regarding the environmental risk factors of myopia and revealed an ethnic disparity in the prevalence of myopia in Gansu Province, Northwest China.


Results characteristics of the Han and Yugur populations.
Of the entire study sample, 3,845 participants aged 40-80 years completed face-to-face questionnaire interview, physical examination and refractive error assessment. There were 2,788 Han participants (1190 men; 42.7%) and 1,057 Yugur participants (507 men; 48.0%). Compared with the Han participants, the Yugur participants spent more time in rural areas (P < 0.001), had lower education levels (primary school or below: men 52.1% vs. 24.5%, P < 0.001; women: 72.2% vs. 38.8%, P < 0.001), were more engaged in outdoor work (P < 0.001) and had higher activity levels (P < 0.001). More participants in the Yugur group had ever been smokers (P < 0.001) and had ever been drinkers (men P = 0.008; women P < 0.001) than those in the Han group. There were also differences between the Han and Yugur participants regarding the prevalence of hypertension (men: 35.3% vs. 44.4% P < 0.001, respectively; women: 28.5% vs. 34.9%, P = 0.005, respectively) and diabetes in the male group (11.9% vs. 7.3% P = 0.004, respectively) ( Table 1).

Discussion
Our study explored the prevalence of myopia in Han and Yugur adults aged 40-80 years in Gansu Province, Northwest China. In this study, the age-and sex-adjusted prevalence of low and moderate myopia decreased with age but mildly increased in the 70-80-year age group (Fig. 1). The U-shaped curve of myopia prevalence was consistent with the findings of previous studies in China 26,27 . A birth cohort study in Europe showed that the prevalence of myopia was low for cohorts born before 1940, but there was a considerable increase in the prevalence of myopia in a more recent cohort when measured at the same age 28 . Data in this birth cohort suggested that increases in the prevalence of myopia may be due to extended education to meet the needs of modern industry 29 . Recent studies have indicated that the rapidly growing incidence of myopia in youths was mainly attributed to environmental factors, among which education was believed to be strongly associated with the increased incidence of myopia; in other words, education is a key causal factor of myopia 29 . A study conducted among German adults revealed that higher levels of school and postschool professional education are associated with more myopic refraction 30 . In our study, most elderly people who were born in prerevolutionary China received unsystematic education in a tumultuous period of Chinese history. The youngest subjects were educated in a society characterized by rapid economic development and the expansion of the education system. Therefore, the younger participants had higher educational levels than the elderly participants had and completed many more years of school. Therefore, it is reasonable to consider that younger individuals may have high levels of reading with a large amount of near work activity. In this scenario, the difference in the exposure level of environmental risk factors among generations may explain the pattern of decreasing myopia prevalence with increasing age. The increased prevalence of myopia among subjects aged 70-80 years might be related to the increased density of the crystalline lens in this group 31 . The prevalence of hyperopia increased with age in both ethnic groups. The age-related change in prevalence was also concluded in previous studies [32][33][34][35] .
As Han is the predominant ethnicity in China, surveys focusing on the prevalence of myopia in the Han population have been carried out worldwide. To make the comparison more specific, we compared the prevalence of myopia (30.8%) and high myopia (5.0%) in the 50-59 age group in the Han population with the prevalence estimates from other studies in Asia or worldwide (Table 5). Previous studies conducted in Han Chinese adults have reported that the prevalence of myopia and high myopia were higher among populations in well-developed countries (35.9-41.1%) 11 20 . This socioeconomic disparity in the prevalence of myopia suggests that the population in more developed countries or areas may experience more intensive education and limited outdoor time, which could promote the development of myopia 37,38 .
Recent knowledge on the etiology of myopia tends to suggest that environmental risk factors play the predominant role in the development of myopia 29 . In our study, the prevalence of myopia in the Yugur population was significantly lower than that in Han population. In the present study, compared with Han participants, the Yugur population, most of whom were born in rural areas, have a much lower educational level and a higher proportion of outdoor work. As stated previously, a longer school year and a lower outdoor activity level are strongly associated with the risk of myopia [39][40][41] . Therefore, environmental factors, such as educational level and outdoor activity level, could, to some extent, explain the ethnic difference in the prevalence of myopia between the Yugur and Han populations in this study. Although no genomic study has explored differences in the susceptibility to myopia between the Yugur and Han groups, there has been study indicated that the Yugur population is closely related to the Han population and to the Mongolian population and is clearly discrete from the Tibetan and the Uygur populations genetically 42 . Thus, we suppose that the difference in the myopia prevalence between the Yugur and Han populations was more related to environmental risk factors than to genetic background. Evidence from other studies also suggests that differences in environmental exposures, rather than genetic background, are the primary cause of ethnic variation in myopia. For example, Singapore provides a multiethnic laboratory in which the role of environmental factors in generating ethnic differences can be assessed 29 . The Singapore study 11 , which included Malays, Indians and Chinese individuals aged 40-80, found that the Chinese group had the highest myopia prevalence. However, it is notable that all ethnic groups in Singapore are more myopic than the same ethnic groups in other parts of the world. For example, the prevalence of myopia in children of Indian origin is much higher in Singapore than in India 43 . Likewise, the prevalence of myopia is higher in Singapore Malays than in Malays in Malaysia 44 . A study conducted in Yunnan Province (Southwest China) among the Yi and Han populations found that the Han ethnicity was no longer associated with myopia after adjustments were made for socioeconomic characteristics in the regression model 35 .
In addition, since 2010, the prevalence of myopia among Europeans and Americans over 40 years of age has increased from that in previous decades. In the Multi-Ethnic Study of Atherosclerosis (2013) in the United States, the prevalence of myopia was 31.0% in white people 12 45,47 . The prevalence of myopia reported in the 2010s in the United States and Europe was close to the prevalence in East Asia in the same age group. Thus, we believe that the ethnic disparity in myopia in this study mainly resulted from different levels of environmental exposure rather than genetic backgrounds.
In the present study, we also found that smoking history was negatively associated with myopia. The nicotinic cholinergic receptor is one of the main acetylcholine receptors distributed in the retina 48 . The results of studies on the impact of cigarette smoking on myopia are inconsistent. A study including 1,334 Chinese children from three schools in Singapore found no significant association between parental smoking and refractive error 49 . The CNHS study in Yunnan Province revealed that smoking history was not associated with myopia (OR, 1.27; 95% CI: 0.76-2.13) 35 . However, several studies indicated that current smoking was negatively associated with myopia (OR, 0.7; 95% CI: 0.5-0.9) 50 . We believe that the impact of smoking on myopia development has duration and dose  accumulation effects. It is possible that neither the duration nor dosage of smoking by the subjects themselves were enough to affect the development of myopia in those studies. The impact of nicotine on the development of myopia requires further laboratory and prospective studies. Furthermore, the association between smoking and myopia could also be modified by socioeconomic status (SES) and education level. SES inequalities may be linked with smoking behavior: people with lower SES and lower education level are more likely to have higher smoking prevalence [51][52][53][54] . We should consider the possibility of the impact of SES and education level on smoking behavior when discussing the association between smoking and myopia. In the present study, we obtained a statistically significant result, but a causal inference cannot be obtained because of the cross-sectional design. The conclusion of the relationship between myopia and smoking requires more exploration.
High myopia is an important cause of eye problems that need to be considered in the clinic 55 . In parallel with the epidemic of myopia, an epidemic of high myopia has appeared 29,56 . The etiology of high myopia includes genetic and environmental aspects 56 . High myopia of genetic etiology tends to be early in onset and severe. Although an increasing number of genes associated with high myopia have been found [57][58][59][60] , previous studies have indicated that the epidemic of high myopia is caused by environmental factors 56 . "Acquired" high myopia is associated with the early onset of myopia because of the early imposition of severe study pressure, which gives myopia more time to progress before it stabilizes 56,61 . In line with other studies 35,46 , the present study revealed that the prevalence of high myopia in young adults is higher than that in older groups. In this case, younger patients with  www.nature.com/scientificreports www.nature.com/scientificreports/ high myopia who receive more education may be more likely to have an environmentally induced early onset of myopia that then progresses to high myopia. Although myopia and high myopia may share some epidemiological risk factors, the association between education and high myopia was inconsistent 62,63 . Education was not found to be associated with high myopia in some older cohorts 62 . The absence of a positive association between education and high myopia in our study may indicate that older people with relatively low educational levels are more likely to develop high myopia that is genetic in origin. The positive association between a family history of myopia and high myopia could also support this idea. Birth in rural areas was also negatively associated with high myopia. The difference in birthplace may lead to variation in environmental risk factors for myopia, such as education attainment, outdoor time, and childhood nutrition.
Strengths of our study include a large multiethnic population-based sampling strategy, a detailed questionnaire and a high response rate. The limitations of the present study should also be acknowledged. First, the nature of the cross-sectional design limited the ability of the study to conclude a causal effect regarding risk factors for myopia. Second, the study population was only older adults in Gansu Province, and thus, the external validity was limited. Third, cycloplegia refraction was not performed in our study, and ocular axial length and other biometric data were not measured; thus, we were unable to study the association between these factors and other biometric measurements and myopia. Fourth, we did not collect information on current cataract history but only cataract surgery history. Because myopia caused by cataract is quite different in etiology, prevention or treatment from that caused by axial myopia, the mixed sample may influence the prevalence as well as the estimation of the association in the present study, especially in the older group.
In summary, for the first time, we described the prevalence of myopia among adults aged 40-80 in Gansu Province, China. The prevalence of myopia in the Han population was significantly higher than that in the Yugur population. Several environmental and lifestyle factors were found to be associated with myopia, while high myopia was only associated with Han ethnicity, birthplace and a family history of myopia. These findings present a rough impression of the prevalence of myopia in Gansu Province. Our study has valuable implications for myopia prevention and control in Northwest China.

Methods
Study population. Our study is part of the CNHS in Gansu Province, Northwest China. The CNHS is an ongoing cross-sectional study to evaluate the national health status using a multistage cluster sampling method, conducted by the Chinese Academy of Medical Sciences 25 . Survey in Gansu Province was from June to August in 2016. The criteria for participant recruitment were: 1) aged 40-80; 2) Yugur or Han population; 3) local resident for at least one year. The exclusion criteria were: 1) women who were currently pregnant; 2) soldiers in service; 3) disabled individuals (who maybe not able to complete the whole physical examination) and 4) individuals with severe mental disorders. Considering the ethnic distribution, the Han and Yugur subjects were recruited from 6 centres: the Gansu Disease Control Centre; 3 county centres for Disease Control (Sunan Yugur autonomous county; Zhangye county; Gaotai county) and 2 village-level health centres (Lianhua village, Minghua township,  view by experienced interviewers and a routine physical examination. The questionnaire contained demographic, lifestyle and health-related information, such as birthplace, current address, ethnicity, education level, occupation, smoking status, alcohol consumption, physical exercise level, medical history (day of diagnosis and treatment for diabetes and hypertension, any diagnosis or surgery history of eye diseases) and myopia family history. The physical examination included anthropometric and blood pressure measurements. Body mass index (BMI) was defined as weight (kg)/height (m) 2 . Educational level was divided into three groups (primary school or below, Middle/high school, undergraduate and above). The occupational group included outdoor and indoor work. Occupational physical labour was divided into light, medium and heavy. Exercise status was divided into 5-7 days a week (at least 20 minutes per day), 3-4 days a week, 1-2 days a week, less than 3 days a month, and never exercise. According to the degree of physical labour and exercise status, the total activity level was divided into three groups: low, moderate and high 64 . The time spent in rural areas was calculated according to the place of residence and the time at which the subjects became residents of the area. Smoking or alcohol consumption was divided into never-smoker/drinker and ever-smoker/drinker. Ever smoker/drinker included past and current smoker/drinker 25 .
Assessment of refractive error. We performed noncycloplegic autorefraction and corneal curvature radius measurements using an auto-refractor (ARK-510A, Nidek Co., Ltd., Tokyo, Japan). An examination of the anterior segment of the eye was performed with a hand-held slit lamp (KJ5S2, Suzhou Kangjie Medical Co. Ltd., Jiangsu, China). Uncorrected visual acuity and best corrected visual acuity were measured by a logarithmic E chart (Wehen Co., Ltd., Guangzhou, China). We use spherical equivalent (SE) to evaluate the refractive error data, which was defined as a sphere plus half cylinder. In our study, emmetropia, low myopia, moderate myopia and high myopia were defined as −0.5≤ SE ≤ 0.5, −3.0≤SE < −0.5D, −3.0D < SE ≤ −6.0D, and SE < −6.0D, respectively. Hyperopia was defined as SE > + 0.5D.

Statistical analyses.
After excluding those who had eye diseases, including glaucoma, pterygium, retina disease, and eye surgical histories (including cataract surgry) or eye injury histories, the refractive error portion of the CNHS in Gansu Province included 4,599 participants aged 40 to 80 (July 2016 to August 2016). Finally, a total of 3,845 participants provided examination data and questionnaire information with no missing value on key risk factors. The response rate was 83.6%. As the correlation coefficients for SE in the left and right eye were high (Spearman correlation test, r s = 0.89), only data of right eye was reported. The age-and sex-specific prevalence of low myopia, moderate myopia, high myopia and hyperopia in each ethnic group was calculated. Chi-square tests were used to compare the demographic, lifestyle-related information and physical examination data between the Han and Yugur subjects. Risk factors for myopia and high myopia were identified by multivariable logistic regression models. A P-value<0.05 (two-tailed) was considered statistically significant. Age-and sex-standardization was performed by direct method using the 6 th national census (2010) data of Chinese population as the standard population. Analyses were performed using Stata version 13.1 (StataCorp, USA) and SAS version 9.4 (SAS Institute Inc, Cary, NC, USA).
ethics approval and consent to participate. Our study was conducted according to the tenets of the Declaration of Helsinki. Ethics approval was received from the bioethics committee of the Institute of Basic Medical Sciences, the Chinese Academy of Medical Sciences. Written informed consent was obtained from every Han or Yugur participant.

Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. This study investigated the prevalence and risk factors of myopia and high myopia in adults aged 40-80 years in the Han and Yugur populations living in Gansu Province, Northwest China.