Update on the Epidemiology and Genetics of Myopic Refractive Error

Justin C Sherwin; David A Mackey


Expert Rev Ophthalmol. 2013;8(1):63-87. 

In This Article

Epidemiology of Myopia

Refraction Across the Lifespan

Refraction is dynamic throughout a lifespan. Newborns are usually hyperopic; over the first few years after birth the prevalence of myopia increases in conjunction with axial elongation, thinning of the lens and flattening of the cornea.[13] There is no simple relationship between lens thickness and lens power. The crystalline lens continues to become thinner and although it starts to thicken early in the second decade of life, it continues to lose power.[27,28] Children with longer AL, vitreous chamber depth and thinner crystalline lenses are more likely to develop myopia.[29] Myopic eye growth associated with an increased AL induces the lens to compensate by becoming much thinner.[28] In adults above 40 years of age, increasing age is often associated with a hyperopic shift, unless nuclear sclerosis/cataract is present, which leads to a late-aged myopic shift.[30,31] There is also evidence that suggests that a hyperopic shift can start in earlier years.[32]

Prevalence of Myopia

The highest prevalence estimates for myopia are for young adults in east Asia, with estimates encroaching 90% in some urbanized and highly educated populations.[33] Using pooled data and a standardized definition of myopia (standard error [SE]: ≤-1.0 D), the crude prevalence of myopia in adults aged ≥40 years in USA, western Europe and Australia was 25.4, 26.6 and 16.4%, respectively.[34] In adults aged ≥40 years in east Asia, the prevalence tends to be higher than in other ethnic populations, but the disparity is less marked than in younger cohorts (Table 1). It is noteworthy that in many studies, participants were often excluded from refraction analyses because of reduced vision and/or if pseudophakic. Adult-onset myopia is not uncommon,[35] although it does not usually progress to the same extent as myopia that arises in childhood. Data from population-based studies in adults do not support a consistently higher prevalence of myopia for either males or females.[36–40] High myopia affects approximately 1–4% of adults aged ≥40 years,[41–46] but is higher in some studies of east Asian adults[47] and in younger east Asian populations.[3,4] Furthermore, the growth rate of high myopia in young adults clearly exceeds that of 'any myopia' in parts of east Asia.[3]

The prevalence of myopia in children and young adults varies according to ethnicity, which could represent inter-ethnic genetic and/or environmental variation. Data from population-based studies performed in children propose that Asian populations, especially of Chinese ethnicity, may possess an inherent susceptibility to myopia compared with western populations.[48] Multiethnic studies of refractive error highlight inter-ethnic differences in the prevalence of refractive error. In Singaporean males aged 15–25 years, the prevalence of myopia was 48.5% in Chinese, 34.7% in Eurasians (mostly white Caucasian), 30.4% in Indians and 24.5% in Malays.[49] In a group of American school children aged 5–17 years, the prevalence of myopia was highest amongst Asians (19.8%) followed by Hispanics (14.5%), African–Americans (8.6%) and whites (5.2%).[50] Inter-ethnic differences persisted even following adjustment for age and sex. The Refractive Error Study in Children (RESC) examined the prevalence of refractive error in children aged 5–15 years in eight locations using a standardized protocol and diagnostic criteria. Across RESC locations in subjects aged 15 years, the prevalence of myopia ranged from less than 1% (0.79%) in rural Nepal[51] to nearly 80% (79.9%) in Guangzhou, China.[7] Such differences in myopia prevalence usually reflect differences in the respective population distributions of AL.[52]

Significant insight into the etiology of myopia can be gained by studying the correlation with migration.[53] Myopia prevalence is higher among second- or later generation Indian immigrants in Singapore than the first-generation immigrants, suggesting that country-specific environmental factors may contribute to the increasing prevalence of myopia in Asia.[54] Rose et al. compared the prevalence and associations of myopia of children of Chinese ethnicity in both Singapore and Sydney.[55] They found a much lower prevalence in Sydney (3.3 vs 29.1%), with the Sydney children spending much more time outdoors per week (mean: 13.75 vs 3.05 h). Reduced time spent outdoors compared with Caucasians was also shown in a clinic-based sample of Chinese–Canadian children with a high prevalence of myopia (64.1% at 12 years of age).[56]

Incidence & Progression of Myopia

Several studies in children and adults have investigated longitudinal changes in refraction. The annual incidence of myopia in Hong Kong was 144.1 cases per 1000 primary school children per annum and was significantly associated with increasing age.[57] The annual mean change in refraction was greater in myopes compared with nonmyopes.[57] An earlier study from Hong Kong showed the incidence of myopia at age 7–8 years and 11–12 years to be 9% and 18–20%, respectively.[58] In Singapore, the 3-year cumulative incidence rates were 47.7 and 32.4% for 7- and 9-year-old children, respectively.[29]

In the Barbados Eye Study (BES), the overall 9-year incidence in adults aged ≥40 years was 12.0% for myopia (2.0% moderate-high myopia) and 29.5% for hypermetropia.[59] The degree of shift in refraction in older adults was age associated, and it can be shown that a myopic shift is more apparent in older adults beginning at approximately 60–70 years of age, which may be due to increasing lenticular nuclear sclerosis.[60,61] At 10 years of follow-up in the Blue Mountains Eye Study (BMES), the gender-adjusted changes in refraction were 0.40, 0.33, -0.02 and -0.65 diopters (D) in persons with baseline ages 49–54, 55–64, 65–74 and ≥75 years, respectively.[31] In the Beaver Dam Eye Study (BDES) there was a mean change of +0.48, +0.03 and -0.19 D for persons aged 43–59, 60–69 and ≥70 years at the baseline examination, respectively.[60]

Cohort Effects on Refraction

In the BDES, participants born in more recent years were more likely to have myopia than those born in earlier years.[60] Here it was shown that the mean refractions in persons 55–59 years at an examination were +0.20, -0.13 and -0.50 D for those born in the years 1928–1932, 1933–1937 and 1938–1942, respectively.[60] A cohort effect has also been suggested in Sweden where a decreased spherical equivalent (more myopic refraction) in 65- to 74-year-olds was observed over time.[62] Further evidence of a cohort effect has been demonstrated in Australian population-based studies.[31,63] There is also evidence of a myopic shift in indigenous Australians over time.[64] Mutti and Zadnik concluded that while most studies have indicated a reduced prevalence of myopia and shortened AL with increasing age in older adults, this is better explained by an intrinsic age-related decrease in an individual's myopia over time rather than a cohort effect.[65]

Results from cross-sectional surveys of younger adults at different time points provide further evidence of cohort effects in refraction. In a study of 421,116 Singaporean military conscripts aged 15–25 years, the estimated prevalence of myopia increased from 26.3% in 1974–1984 to 43.3% in 1987–1991.[66] A further study has been undertaken involving 15,086 new male Singaporean military conscripts aged 16–26 years in 1996 and a repeat survey of 29,170 similarly aged males in 2009–2010.[67] Overall myopia prevalence increased from 79.3 to 81.3% between 1996 and 2009, with a concomitant increase in prevalence of high myopia (13.1–14.8%). In Taiwan, the mean refraction in 4686 freshmen was -4.25 ± 2.74 D in 1988 and -4.93 ± 2.82 D in 2005 among 3709 freshman.[4] Similarly, in a series of cross-sectional surveys of 919,929 Israeli Army conscripts, the overall prevalence of myopia increased from 20.3% in 1990 to 28.3% in 2002.[68]

Notably in east Asia, the population distribution of refractive error has undergone a myopic shift in only a few generations. Even those children without myopic parents will likely become myopic, some severely. The large and rapid increase in myopia prevalence in recent birth cohorts of east Asian origin and elsewhere where large differences in environmental pressures have been observed, coupled with the lower heritability estimates and parent–offspring correlations obtained from parent and sibling–offspring correlations in such populations, supports an important and major role for environmental factors influencing myopia.[33,69–71]

AL: The Major Endophenotype of Refractive Error

AL is composed of the sum of anterior and vitreous chamber depths and lens thickness. In emmetropic, pretreated eyes AL does not correlate with myopia susceptibility; in a study of chicks it was shown that the correlation between genetic variants controlling susceptibility to visually induced myopia and variants controlling normal eye size was negligible.[72] However, when looking at a whole population containing both emmetropes and ametropes, increasing AL is associated with an increasingly myopic refraction, suggesting that factors leading to an increasing AL may share homology with those contributing to myopia and a myopic refraction.[73,74]

AL reaches its fastest rate of growth in the year prior to myopia onset before growing at a much slower pace following onset.[75] In Caucasian children before myopia onset, children with myopic parents display a longer AL than children with no positive family history, even after adjustment for environmental covariates.[76] Notably, parental history of myopia influenced the eye's growth rate rather than size in a study of Chinese children.[77] Peripheral hyperopia may stimulate the axial elongation in myopia.[78] Greatly increased AL is characteristic of individuals with high myopia, and the degree of AL correlates strongly with high myopia severity.[79] In some adults with high myopia, the AL continues to increase, especially in those who develop posterior staphyloma.[80] One benefit of using AL in analyses is that it remains relatively constant following cataract or refractive surgery, which is not true for refractive readings.

AL does not represent a perfect surrogate for refraction. Male sex is associated with a longer AL than female sex,[81,82] which may reflect differences in stature, but this does not translate into a higher prevalence of myopia.[42,83,84] Furthermore, the inter-ethnic variation in height with resulting longer ALs does not simply translate into a higher prevalence of myopia in taller populations. For example, the prevalence of myopia in the Dutch adult population,[34] widely believed to be among the tallest population in the world, is much less than in the adult Chinese.[47]

Environmental, Socioeconomic & Lifestyle Risk Factors of Myopia

Environmental and lifestyle factors have strongly influenced the rapid increases in the prevalence of myopia over time in some populations.[42,71] Using a lifecourse epidemiological approach, several novel putative risk factors were identified from the 1958 British Birth Cohort Study, which support the overall global increase in myopia prevalence. These factors include increasing maternal age, increasing rates of intrauterine growth retardation, persistence of smoking during pregnancy, changing socioeconomic status and reduced rates of breastfeeding.[85]

Socioeconomic Factors. In ethnically homogenous populations, people growing up in an urban environment have a higher risk of myopia than people from rural regions.[45,86–88] This incongruity may reflect differences in educational level and socioeconomic class rather than the effect of the environment itself. Other explanations may relate to long-distance viewing, differences in light intensity and optical field, and changes in levels of physical and outdoor activity.[33]

One of the strongest environmental determinants of myopia is educational attainment. Prevalence of myopia increases and overall mean spherical error decreases (becomes more myopic) with increasing years of formal education.[89,90] Educational attainment is associated with higher income and decreased unemployment,[91] factors that are also associated with refractive error. In an adult Japanese population, risk of myopia is higher in professional occupations with higher incomes and higher educational attainment.[37] In the 1958 British Birth Cohort study, the association between social class and myopia was significantly attenuated in the later life-stage models, thereby suggesting social class was mediated through other relevant childhood factors, which may include education, growth and reduced time spent outdoors.[85] It is important to acknowledge that educational attainment is also influenced by genetic factors and should not be regarded as merely environmental.[92]

Diet. Studies of the relationship between diet and refraction have been limited by small sample sizes and imprecise methods of dietary assessment. No dietary factor appears relevant in myopia development or its progression. Malnutrition does not appear to be associated with refractive error.[93] Children with vegetarian diets had an increased prevalence of refractive errors compared with omnivorous children,[94] but reasons to account for this are not clear. Using food records, Edwards found statistically significant differences in energy intake, protein, fat, vitamins B1, B2 and C, phosphorus, iron and cholesterol between 24 children who became myopic between 7 and 10 years of age and 68 controls.[95] However, there were no differences in height or weight of cases and controls, and the study was not population-based. No dietary factors for myopia, as assessed by food-frequency questionnaire, were identified in a cross-sectional sample of 851 Chinese school students aged 12.81 ± 0.83 years and, although some dietary components (cholesterol and fat) were associated with AL, multiple testing was not accounted for in analyses.[96] Several studies have shown that being breastfed as an infant is associated with increased hyperopic refraction,[97,98] but this finding is not consistently replicated.[99]

Diabetes & Glycemia. The relationship between diabetes, acute and chronic changes in glycemia and refractive error is complex. Acutely transient changes of refraction due to hyperglycemia could lead to either a myopic or hyperopic shift, depending on whether changes occur in the refractive indices or in the lens. Hyperglycemia-associated changes in the lens nucleus and cortex often lead to transient myopia and hyperopia, respectively; corneal edema often also ensues.[100] In the cross-sectional Handan Eye Study and Los Angeles Latino Eye Study (LALES), diabetes was significantly associated with myopia.[41,46] These results have not been well supported by longitudinal studies. Diabetes was not associated with incident myopia in a 5-year follow-up study of the BMES.[101] The 9-year follow-up of the BES showed that diabetes was not significantly associated with any myopia, but was predictive of moderate-to-severe myopia (spherical equivalent ≤-3 D).[59] Furthermore, the BDES found that people with diabetes even underwent a hyperopic shift over time.[60] In a cross-sectional study of subjects with Type 2 diabetes in an urban Indian population aged ≥40 years, poor glycemic control was associated with myopia.[102] In a Danish population, increasing HbA1c was associated with an increased odds of myopia, and risk of myopic shift was increased with HbA1c ≥8.8 compared with HbA1c <8.8, suggesting that myopia may be associated with chronic hyperglycemia.[103] By contrast, other studies have found that hyperglycemia and elevated HbA1c have been associated with a hyperopic refraction.[104,105]

Smoking & Alcohol. Parental smoking is independently associated with reduced myopia and increased hyperopic refraction.[106] For adults, higher frequency of personal smoking is independently associated with hyperopic refraction.[107] In a population-based cross-sectional study of 6491 adults aged 30–99 years in China, smoking was protectively associated with myopia, following adjustment for age, diabetes, reading and family history of myopia.[46] In the Beijing Eye Study, subjects who consumed alcohol had less myopic refraction than those who did not, but this association was attenuated following adjustment for possible confounders.[108]

Anthropometric Measures. The finding that body stature is associated with myopia is unconvincing. Although in some studies myopes were significantly taller than nonmyopes,[109,110] many studies found only a weak or negligible association.[111,112] In the Genes in Myopia study, following adjustment for age, gender, educational attainment and ocular biometric characteristics, height was not associated with myopic refraction.[113] In a study of 106,926 consecutive Israeli male military recruits aged 17–19 years, myopia was not associated with taller height, and indeed an association in the opposite direction was found.[114] Because genetic coregulation exists between stature and AL, taller people have longer eyes,[115] yet because the same genes also appear to control other ocular component dimensions[116] the eyes still end up emmetropic in general.

Myopia has also been associated with low birthweight for gestational age.[85] The relationship between myopia and BMI is absent[117] or negligible.[114] Elsewhere in Asian populations, a higher BMI has been associated with hyperopic refraction in children[110] and adults.[107]

Physical & Sporting Activity. In a 2-year prospective study of Caucasian Danish medical students, duration of physical activity was associated with a myopic refraction (0.175 D per hour of physical activity per day; p = 0.015) after multivariate adjustment.[118] Adult physical activity levels do not appear to show the same association.[119] Using objective measurements of physical activity, 12-year-old children with myopia spent less time in moderate–vigorous physical activity than other children, and also increased time in sedentary behavior, even following adjustment for social and behavioral confounders.[120] Playing sports is inversely associated with myopia,[121,122] but not indoor sports,[123] suggesting exposure to an outdoor environment may be more important than the exercise component.

Time Spent Outdoors. The current available evidence favors time spent outdoors, rather than time spent in physical activity, as being protective against myopia. A systematic review and meta-analysis has demonstrated a small but significant association between reduced time outdoors and myopia in children and young adults in observational studies.[23] Time spent outdoors is protective against incident myopia independently of physical activity level,[124] and before becoming myopic, children spend less time outdoors than those who do not become myopic.[125] Accordingly, findings from a small RCT[126] have been encouraging, and several other RCTs investigating the effect of time spent outdoors, of which several are currently underway, will shed further important insights. Why being outside is protective is incompletely understood, although exposure to natural light outdoors represents a possible mechanism. This has been supported by a protective association of myopia with increasing conjunctival ultraviolet autofluorescence,[127] a novel biomarker of time spent outdoors.[128,129]

Near Work. Traditionally, time spent doing activities involving near work has been regarded as an important risk factor for myopia, although animal data and epidemiological data are inconsistent.[130] Some studies have demonstrated that increased time spent on activities involving near vision is associated with myopia in primary school children.[131] Further, myopia is associated with proxy measures of near work in adults,[85,132] but reverse causality is likely. In any case, there appears to be little difference in near work activity between and after development of myopia compared with emmetropia, suggesting other factors may be more important.[133] A number of studies have found that television watching is not associated with myopia in children.[122,134,135] Television viewing habits also do not appear to differ before and after the onset of myopia.[133]