Worldwide prevalence and risk factors for myopia
Abstract
Citation information: Pan C‐W, Ramamurthy D & Saw S‐M. Worldwide prevalence and risk factors for myopia. Ophthalmic Physiol Opt 2012, 32, 3‐16. doi: 10.1111/j.1475‐1313.2011.00884.x
Abstract
Background: Myopia, the most common type of refractive error, is a complex trait including both genetic and environmental factors. Numerous studies have tried to elucidate the aetiology of myopia. However, the exact aetiology of myopia is still unclear.
Purpose: To summarize the worldwide patterns and trends for the prevalence of myopia and to evaluate the risk factors for myopia in population‐based studies.
Recent findings: The prevalences of myopia vary across populations of different regions and ethnicities. In population‐based studies on children, the prevalence of myopia has been reported to be higher in urban areas and Chinese ethnicity. The regional and racial difference is not so obvious in adult populations aged over 40 years. More time spent on near work, less time outdoors, higher educational level and parental history of myopia have been reported to increase the risk of myopia.
Conclusions: Environmental factors play a crucial role in myopia development. The effect of gene‐environment interaction on the aetiology of myopia is still controversial with inconsistent findings in different studies. A relatively hyperopic periphery can stimulate compensating eye growth in the centre. Longitudinal cohort studies or randomized clinical trials of community‐based health behaviour interventions should be conducted to further clarify the aetiology of myopia.
Myopia is a global public health problem leading to visual impairment and blinding complications.1 The economic costs of myopia are also high. In Singapore, the mean annual direct cost of myopia for each Singaporean school children aged 7–9 years was estimated to be US$148.2 In the United States, the National Health and Nutrition Examination Survey (NHANES) reported the annual direct cost of correcting distance vision impairment due to refractive errors to be between US$3.9 and US$7.2 billion.3 The medical burden of high myopia includes pathologic complications such as myopic macular degeneration, choroidal neovascularisation, cataract and glaucoma.1 Uncorrected refractive error could also impair vision‐related quality of life and increase difficulty in performing vision‐related tasks.4
In the past few decades, numerous epidemiology studies have provided information on the pattern of prevalence and risk factors for myopia. Population‐based studies with sufficient sample sizes, high response rates and few biases provide the strongest evidence for examining the aetiology of myopia. A recent review summarized the data of prevalence and risk factors for myopia published in Ophthalmic and Physiological Optics.5 However, several questions remain unanswered: Are the rates of myopia in Asia higher in East Asians than other ethnic groups in Asia? Is there any gene‐environment interaction? Does outdoor activity play a crucial role in myopia development? In this perspective, we reviewed the major population‐based studies on the epidemiology of myopia, summarized key findings and highlighted future challenges for the research community. The rationale for grouping studies in this review was based on geographic location and ethnicity.
Prevalence of myopia in adults in Asian Countries
In mainland China, the prevalence of myopia for definitions of spherical equivalent (SE) of <−0.50 D, <−1.0 D, <−6.0 D, and <−8.0 D were reported to be 22.9% (95% CI 21.7, 24.2), 16.9% (95% CI 15.8, 18.0), 2.6% (95% CI 2.2, 3.1), and 1.5% (95% CI 1.1, 1.9) respectively, in the Beijing Eye Study (n = 4439, aged 40–90 years).6 The limitation of this study is that refraction was not performed on subjects with an uncorrected visual acuity of 0.0 logMAR (Snellen 6/6) or better. The Shihpai Eye study in Taiwanese adults aged over 65 years reported the prevalence to be 19.4% and 14.5% for myopia of SE < −0.5 D and SE < −1.0 D, respectively. The prevalence of myopia in Taiwan seems to be lower than that of Beijing Eye Study. The difference in prevalence of <3.5% between Taiwan and Beijing is marginal. This difference in prevalence is attributed to the older sample in Taiwan leading to a hyperopic shift in refraction, but this difference in age would also work in the opposite direction with a potential myopic shift due to the onset of nuclear cataract in the older population.7 In Japanese adults, the prevalence was reported to be 41.8% for myopia of SE < −0.5 D.8 The Japanese study may have overestimated the prevalence of myopia due to younger participants and non‐cycloplegic refraction.
In India, the prevalence of myopia for SE < −0.5 D in 40 year and older Indian adults in both urban and rural areas was reported to be 34.6% (n = 3723) in the Indian state of Andhra Pradesh, with a prevalence of 38.0% in rural areas and 31.9% in urban areas. The higher prevalence of myopia in the rural Indian population could be explained by higher rates of nuclear cataract in rural India leading to a myopic shift in refraction.9 In another study of rural Indian adults aged over 39 year in Chennai (n = 2508), the prevalence was reported to be 31% for myopia of SE < −0.5 D.10 The extent of non‐participation bias cannot be elucidated as neither of the studies in India revealed details about the respondents and non‐respondents.
The Tanjong Pagar Survey (TPS), the Singapore Malay Eyes Study (SiMES) and the Singapore Indian Eye Study (SINDI) analyzed the prevalence of myopia of SE < −0.50 D in Singaporean Chinese, Malay and Indian adults aged over 40 years and reported it to be 38.7%,11 26.2%12 and 28.0%,13 respectively. The difference in the prevalences may reflect secular trends over time as well as inter‐ethnic variation since the TPS was conducted a few years prior to SiMES and SINDI.
In Bangladesh and Pakistani adults aged over 30 years, the prevalence of myopia (SE < −0.5 D) has been reported to be 23.8% (n = 11 624) and 36.5% (n = 14 490) respectively whereas it is about 48.1% in Indonesian young adults aged over 21 years (n = 1043).14-16 The prevalence of myopia in Mongolian adults over 40 years was reported to be 17.2% (n = 1617).17 In the WHO National Blindness and Low Vision Surveys in Bangladesh, non‐cycloplegic refraction and subjective refraction were only performed on those with visual acuity worse than 0.30 logMAR (Snellen 6/12). Thus, the prevalence of myopia may have been overestimated (Table 1).
| Author (year) | Country | N | Age | Definition | Refraction method | Prevalence (%) | 95% CI |
|---|---|---|---|---|---|---|---|
| Cheng (2003) | Taiwan | 1361 | 65+ | SE < −0.5 D | Subjective | 19.4 | 16.7, 22.1 |
| Sawada (2007) | Japan | 3021 | 40+ | SE < −0.5 D | Subjective | 41.8 | 40.0, 43.6 |
| Saw (2002) | Indonesia | 1043 | 21+ | SE < −0.5 D | Objective | 48.1 | 45.0, 51.1 |
| Gupta (2008) | Myanmar | 1863 | 40+ | SE < −1.0 D | Objective | 42.7 | 40.4, 44.9 |
| Xu (2005) | China | 5324 | 40+ | SE < −0.5 D | Subjective | 22.9 | 21.7, 24.2 |
| Krishnaiah (2009) | India | 3642 | 40+ | SE < −0.5 D | Subjective | 34.6 | 33.1, 36.1 |
| Raju (2004) | India | 2508 | 40+ | SE < −0.5 D | Subjective | 31.0 | Not available |
| Shah (2008) | Pakistan | 14 490 | 30+ | SE < −0.5 D | Objective | 36.5 | 35.7, 37.3 |
| Bourne (2004) | Bangladesh | 11 189 | 30+ | SE ≤ −0.5 D | Objective | 23.8 | 23.8, 23.8 |
| Wong (2000) | Singapore | 1232 | 40+ | SE < −0.5 D | Subjective | 38.7 | 35.5, 42.1 |
| Saw (2008) | Singapore | 2974 | 40+ | SE < −0.5 D | Subjective | 26.2 | 26.0, 26.4 |
| Pan (2011) | Singapore | 2805 | 40+ | SE < −0.5 D | Subjective | 28.0 | 25.8, 30.2 |
| Tarczy‐Hornoch (2006) | USA | 5396 | 40+ | SE ≤ −1.0 D | Subjective | 16.8 | Not available |
| Katz (1997) | USA | 5028 | 40+ | SE < −0.5 D | Subjective | 28.1 (white); 19.4 (black) | Not available |
| Vitale (2008) | USA | 12 010 | 20+ | SE < −0.5 D | Objective | 33.1 | 31.5, 34.7 |
| Wu (1999) | USA | 4709 | 40–84 | SE < −0.5 D | Objective | 21.9 | 20.6, 23.2 |
| Wang (1994) | USA | 4926 | 43–84 | SE < −0.5 D | Objective | 26.2 | Not available |
| Wensor (1999) | Australia | 4744 | 40–98 | SE < −0.5 D | Subjective | 17.0 | 15.8, 18.0 |
| Attebo (1999) | Australia | 3654 | 49–97 | SE < −0.5 D | Subjective | 15.0 | Not available |
| Rahi (2011) | UK | 2487 | 44–45 | SE ≤ −0.75 D | Objective | 49.0 | 48.8, 50.8 |
| Midelfart (2002) | Norway | 3137 | 20–25 | SE < −0.5 D | Subjective | 35.0 | Not available |
| 40–45 | 30.3 |
Prevalence of myopia in adults in Western Countries
In the United States, the 1999–2004 NHANES used an autorefractor to measure refractive data on a US non‐institutionalized, civilian population aged 20 years or older. The age‐standardized prevalence of myopia (SE < −1.0 D or less) was 33.1% (95% CI 31.5, 34.7) in 12 010 participants.18 In this study, non‐cycloplegic refraction may have caused an overestimation of myopic persons among younger participants. In the Baltimore Eye Survey (n = 5028), the prevalence of myopia (SE < −0.5 D) was 28.1% among the white and 19.4% among the black.19 The Los Angeles Latino Eye Study reported a myopia prevalence of 16.8% in 40 years or older adults (n = 5927) in the worse eye.20 In the Beaver Dam Eye Study, the age‐gender adjusted prevalence of myopia (SE < −0.5 D) was 26.2% based on the data of the right eye.21 The Barbados Eye Study examined the prevalence of myopia in African‐Americans aged 40–84 years (n = 4709). The age‐gender adjusted prevalence of myopia (SE < −0.5 D) was 21.9% (95% CI 20.6, 23.2) based on objective refraction data.22 The Beaver Dam Eye study of adults aged over 43 years may have overestimated the prevalence of myopia in terms of the younger respondents. On the contrary, the NHANES on people aged over 20 years may have underestimated the prevalence of myopia since the younger working adults were more difficult to recruit than the older ones.
In the UK, among a total of 2487 randomly selected 44‐year‐old members of the 1958 British birth cohort, 1214 individuals (49%; 95% CI 48.8, 50.8) were myopic. Refraction was measured by autorefraction using the Nikon Retinomax 2 (Nikon Corp., http://www.nikon.com/), under non‐cycloplegic conditions. Thus, myopia prevalence may have been overestimated.23 In Norway, non‐cycloplegic refraction was measured in a population‐based sample of young (20–25 years) and middle‐aged (40–45 years) adults. A total of 3137 persons (1248 young and 1889 middle‐aged adults) with corrected visual acuity worse than 0.3 logMAR (Snellen 6/12) in either eye were included in the study. The prevalence of myopia (SE < −0.5 D) was 35.0% in the young adult group and 30.3% in the middle‐aged group. Prevalence of myopia was overestimated especially for the young adult group due to the non‐cycloplegic refraction.24
In Australia, the Blue Mountains Study reported a prevalence of myopia in adults aged 40–97 years of 15.0% (n = 3654).25 The Visual Impairment Project reported a myopia (SE < −0.5 D) prevalence of 17.0% (95% CI 15.8, 18.0).26 A meta‐analysis by the Eye Diseases Prevalence Research Group estimated the crude prevalence rates for myopia of −1.0 D or less as 25.4%, 26.6%, and 16.4% in the United States, Western Europe and Australia, respectively.27 (Table 1)
Based on the published data of myopia prevalence on adults, it is still unclear whether the myopia prevalence is higher in East Asian Countries than in Western Countries. The prevalence of myopia is 38.7% in Singaporean Chinese (SE < −0.5 D).11 However, the meta‐analysis by Kempen et al.27 showed that the prevalence of myopia is 25.4% and 26.6% for White subjects in the United States and Western Europe using a more conservative definition of myopia (SE < −1.0 D), respectively. The cut off used to define myopia is arbitrary but the prevalence might change significantly by a small shift in this cut‐off value.26 In Singapore, the Chinese have a higher prevalence of myopia compared with Malays and Indians living in the same country and the myopia prevalence in South Asia in the Indian population is only marginally lower than the Singaporean Chinese. The myopia prevalence reported in the Singaporean Malays12 and Indians28 are also lower than those from North America.19, 21
Worldwide prevalence of myopia in children
Tables 2 and 3 summarize the overall and age‐specific prevalence of myopia in children. The Refractive Error Study in Children (RESC) was conducted in different countries using the same sampling strategies, procedures to measure refraction and definitions of myopia, in order to compare the prevalence of myopia across different study populations. In Nepal, the prevalence of myopia ranged from 10.9% in 10‐year‐old children, 16.5% in 12‐year‐olds, to 27.3% in 15‐year‐old children living in the urban region, whereas it was <3% in 5–15 year old children in rural Nepal.29, 30 In urban India, the prevalence of myopia was 4.7%, 7.0% and 10.8% in 5, 10 and 15 year‐olds, respectively. On the other hand, the prevalence of myopia was 2.8%, 4.1% and 6.7% in 7, 10 and 15‐year‐olds, respectively in the rural region.31, 32 Among urban Chinese children the prevalence of myopia ranged from 5.7% in 5‐year‐olds, 30.1% in 10‐year‐olds and increased to 78.4% in the 15‐year‐olds.33 In rural parts of northern China, the prevalence of myopia was almost nil in 5‐year‐olds and steadily increased to 36.7% and 55.0% in 15‐year‐old males and females respectively.34 In the rural region of Southern China, 36.8% of 13‐year‐olds, 43.0% of 15‐year‐olds and 53.9% of 17‐year‐olds were found to be myopic.35 In brief, the prevalence of myopia was highest (78.4%) in 15‐year‐old urban Chinese children33 and lowest (1.2%) in 5–15 year old rural Nepalese children.30
| Author (Year) | Location | N | Age range | Myopia definition | Prevalence (%) | 95% CI |
|---|---|---|---|---|---|---|
| Pokharel (2000) | Mechi Zone, Nepal | 5067 | 5–15 years | ≤−0.5 D | 1.2 | Not available |
| Sapkota (2008) | Kathmandu, Nepal | 4282 | 10–15 years | ≤−0.5 D | 19.0 | 17.8, 20.2 |
| Murthy (2002) | New Delhi, India | 6447 | 5–15 years | ≤−0.5 D | 7.4 | 5.0, 9.7 |
| Dandona (2002) | Andhra Pradesh, India | 4074 | 7–15 years | ≤−0.5 D | 4.1 | 3.3, 4.9 |
| Goh (2005) | Gombak district, Malaysia | 4634 | 7–15 years | ≤−0.5 D | 20.7 | 17.3, 24.1 |
| Zhao (2000) | Shunyi District, Beijing, China | 5884 | 5–15 years | ≤−0.5 D | 21.6 | Not available |
| He (2004) | Guangzhou, China | 4364 | 5–15 years | ≤−0.5 D | 38.1 | 36.3, 39.8 |
| He (2007) | Yangxi,Guangdong province,China | 2454 | 13–17 years | ≤−0.5 D | 42.4 | 35.8, 49.0 |
| Naidoo (2003) | South Africa | 4890 | 5–15 years | ≤−0.5 D | 4.0 | 3.3, 4.8 |
| Maul (2000) | La Florida, Chile | 5303 | 5–15 years | ≤−0.5 D | 7.3 | Not available |
| Saw (2005) | Singapore | 1453 | 7–9 years | ≤−0.5 D | 36.7 | 34.2, 39.2 |
| Dirani (2009) | Singapore | 2369 | 6–72 months | ≤−0.5 D | 11.0 | 10.9, 11.2 |
| Zadnik (1997) | USA | 716 | 6–14.9 years | ≤−0.75 D | 6 years: 2, 12 years: 20 | Not available |
| Ip (2008) | Australia | 2353 | 12 years | ≤−0.5 D | 11.9 | 6.6, 17.2 |
| Rudnicka (2010) | UK | 1053 | 10–11 years | ≤−0.5 D | 3.4 | Not available |
| O’Donoghue (2010) | Northern Ireland | 1053 | 6–7 years | ≤−0.5 D | 2.8 | 1.3, 4.3 |
| 12–13 years | 17.7 | 13.2, 22.2 | ||||
| Logan (2011) | England | 327 | 6–7 years | ≤−0.5 D | 9.4 | Not available |
| 12–13 years | 29.4 |
| Author (Year) | Study design/Population (N) | Response rate (%) | Cycloplegic refraction | Myopia definition | Prevalence (95% CI) |
|---|---|---|---|---|---|
| Dirani (2009) | Population‐based cross‐sectional study, N = 2369 Chinese children | 72.3 | Cycloplegic autorefraction | ≤−0.5 D | 6–11.9 months: 15.8% (10.6, 22.2) |
| 12–23.9 months: 14.9% (11.7, 18.5) | |||||
| 24–35.9 months: 20.2% (16.5, 24.2) | |||||
| 36–47.9 months: 8.6% (6.3, 11.3) | |||||
| 48–59.9 months: 7.6% (5.5, 10.1) | |||||
| 60–72 months: 6.4% (4.5, 8.8) | |||||
| Saw (2005) | School‐based cross‐sectional study, N = 1453 Chinese children | 66.3 | Cycloplegic autorefraction | ≤−0.5 D | 7 years: 29.0% (25.5, 32.6) |
| 8 years: 34.7% (30.4, 39.0) | |||||
| 9 years: 53.1% (47.9, 58.4) | |||||
| Sapkota (2008) | Population‐based N = 4282 children from Kathmandu, Nepal | 95.1 | Cycloplegic autorefraction | ≤−0.5 D | 10 years: 10.9% (7.00, 14.7) |
| 11 years: 13.8% (10.5, 17.2) | |||||
| 12 years: 16.5% (13.2, 19.8) | |||||
| 13 years: 19.4% (16.7, 22.1) | |||||
| 14 years: 23.3% (20.0, 26.7) | |||||
| 15 years: 27.3% (22.6, 32.0) | |||||
| Murthy (2002) | Population‐based N = 6447 children from New Delhi, India | 92.0 | Cycloplegic retinoscopy | ≤−0.5 D | 5 years: 4.68% (2.54, 6.83) |
| 6 years: 5.87% (2.59, 9.15) | |||||
| 7 years: 3.13% (1.17, 5.08) | |||||
| 8 years: 5.67% (2.50, 8.84) | |||||
| 9 years: 5.33% (2.61, 8.05) | |||||
| 10 years: 6.95% (3.44, 10.5) | |||||
| 11 years: 9.85% (5.91, 13.8) | |||||
| 12 years: 9.66% (5.64, 13.7) | |||||
| 13 years: 10.6% (6.02, 15.2) | |||||
| 14 years: 10.2% (6.85, 13.5) | |||||
| 15 years: 10.8% (6.71, 14.8) | |||||
| Dandona R (2002) | Population‐based N = 4074 children from Andhra Pradesh, India | 92.3 | Cycloplegic retinoscopy | ≤−0.5 D | 7 years: 2.80% (1.28, 4.33) |
| 8 years: 2.83% (1.50, 4.16) | |||||
| 9 years: 3.90% (2.05, 5.74) | |||||
| 10 years: 4.06% (2.09, 6.03) | |||||
| 11 years: 2.73% (1.38, 4.09) | |||||
| 12 years: 4.79% (2.91, 6.97) | |||||
| 13 years: 5.43% (3.25, 7.60) | |||||
| 14 years: 6.74% (3.31, 10.2) | |||||
| 15 years: 6.72% (4.31, 9.12) | |||||
| Goh (2005) | Population‐based N = 4634 children from Gombak district, Malaysia | 32.8 | Cycloplegic autorefraction | ≤−0.5 D | 7 years: 10.0% (6.8, 13.1) |
| 8 years: 14.0% (10.3, 17.6) | |||||
| 9 years: 16.3% (11.7, 20.9) | |||||
| 10 years: 16.2% (11.6, 20.7) | |||||
| 11 years: 22.6% (17.0, 28.2) | |||||
| 12 years: 24.8% (19.1, 30.6) | |||||
| 13 years: 25.3% (19.5, 31.1) | |||||
| 14 years: 32.5% (25.5, 39.6) | |||||
| 15 years: 32.5% (25.5, 39.6) | |||||
| Zhao (2000) | Population‐based N = 5884 children from Shunyi District, Beijing, China | 95.9 | Cycloplegic autorefraction | ≤−0.5 D | Males: |
| 5 years: 0 | |||||
| 15 years: 36.7% (29.9, 43.4) | |||||
| Females: | |||||
| 5 years: 0 | |||||
| 15 years: 55.0% (49.4, 60.6) | |||||
| He (2004) | Population‐based cluster sampling, N = 4364 children from Guangzhou, China | 86.4 | Cycloplegic autorefraction | ≤−0.5 D | 5 years: 5.7% (2.3, 9.0) |
| 6 years: 5.9% (2.6, 9.2) | |||||
| 7 years: 7.7% (4.7, 10.8) | |||||
| 8 years: 14.0% (10.4, 17.6) | |||||
| 9 years: 25.9% (22.0, 29.8) | |||||
| 10 years: 30.1% (24.4, 35.8) | |||||
| 1 years: 41.7% (37.3, 46.1) | |||||
| 12 years: 49.7% (44.7, 54.6) | |||||
| 13 years: 57.4% (52.1, 62.6) | |||||
| 14 years: 65.5% (62.4, 68.5) | |||||
| 15 years: 78.4% (74.5, 82.2) | |||||
| He (2007) | Population‐based N = 2454 children from Yangxi, Guangdong province, China | 97.6 | Cycloplegic autorefraction | ≤−0.5 D | 13 years: 36.8% (29.2, 44.3) |
| 14 years: 38.8% (30.8, 46.7) | |||||
| 15 years: 43.0% (34.5, 51.4) | |||||
| 16 years: 46.8% (37.7, 55.9) | |||||
| 17 years: 53.9% (39.6, 68.1) | |||||
| Giordano (2009) | Population‐based cross‐sectional study, N = 1268 African‐American and N = 1030 White children | Not stated | Cycloplegic autorefraction | ≤−1.0 D | African‐American: |
| 6–11 months: 7.5% | |||||
| 12–23 months: 10.5% | |||||
| 24–35 months: 5.9% | |||||
| 36–47 months: 6.2% | |||||
| 48–59 months: 6.6% | |||||
| 60–72 months: 7.4% | |||||
| Whites: | |||||
| 6–11 months: 0% | |||||
| 12–23 months: 2.3% | |||||
| 24–35 months: 1.1% | |||||
| 36–47 months: 0% | |||||
| 48–59 months: 1.5% | |||||
| 60–72 months: 1.1% | |||||
| Naidoo (2003) | Population‐based N = 4890 children from South Africa | 87.3 | Cycloplegic autorefraction | ≤−0.5 D | 5 years: 3.2% (0.6, 5.7) |
| 6 years: 4.6% (2.4, 6.7) | |||||
| 7 years: 2.5% (0.8, 4.2) | |||||
| 8 years: 2.9% (1.2, 4.6) | |||||
| 9 years: 3.1% (1.4, 4.8) | |||||
| 10 years: 1.9% (0.6, 3.2) | |||||
| 11 years: 4.4% (2.8, 6.1) | |||||
| 12 years: 4.4% (2.2, 6.6) | |||||
| 13 years: 3.4% (1.7, 5.2) | |||||
| 14 years: 6.3% (3.6, 8.9) | |||||
| 15 years: 9.6% (6.4, 12.7) | |||||
| Maul (2000) | Population‐based N = 5303 children from La Florida, Chile | 75.8 | Cycloplegic autorefraction | ≤−0.5 D | Males: |
| 5 years: 3.4% (1.87, 5.00) | |||||
| 15 years: 19.4% (13.6, 25.2) | |||||
| Females: | |||||
| 5 years: 3.4% (1.72, 5.05) | |||||
| 15 years: 14.7% (10.1, 19.2) | |||||
| Solang (2008) | Population‐based N = 2441 children from Brazil | 86.4 | Cycloplegic autorefraction | ≤−0.5 D | 11 years: 5.4% (3.72, 7.08) |
| 12 years: 4.52% (2.53, 6.65) | |||||
| 13 years: 5.83% (4.57, 7.08) | |||||
| 14 years: 6.05% (4.2, 7.89) |
In Singapore, the prevalence of myopia was 29.0% in 7‐year‐olds, 34.7% in 8‐year‐olds and 53.1% in 9‐year‐olds in the school‐based population of the Singapore Cohort Study of Risk factors for Myopia (SCORM)36 while the Strabismus, Amblyopia and Refractive error Study in Singapore Preschool Children (STARS) reported that the prevalence of myopia was 11.0% in Chinese children aged 6–72 months.37 In Hong Kong, a large cross‐sectional survey reported that the prevalence was 17.0% in children aged <7 years and which increased to 37.5% among those aged 8 years and 53.1% in children aged more than 11 years.38 The prevalence of myopia among Taiwanese Chinese primary school children aged 7 years was 5.8% in 1983, 3.0% in 1986, 6.6% in 1990, 12.0% in 1995 and 20.0% in 2000. Among Taiwanese children aged 12 years, the myopic rates were 36.7%, 27.5%, 35.2%, 55.5% and 61.0% correspondingly. At the junior high school level, the prevalence was 64.2%, 61.6%, 74.0%, 76.0% and 81.0% respectively. Among children aged 16–18 years, the myopia prevalence was almost constant at around 74–75% in studies conducted in 1983, 1986 and 1990. However, the prevalence rate increased to 84% in studies in 1995 and 2000.39
The prevalence of myopia has also been reported in non‐Asian populations. Among South African children, the prevalence of myopia was about 3% or 4% increasing to 6.3% in 14‐year‐olds and 9.6% in 15‐year‐olds.40 In Chile, 3.4% of the 5‐year‐olds were myopic and the prevalence rate increased to 19.4% and 14.7% in the 15‐year‐old males and females respectively.41 In Australia, the Sydney Myopia Study (SMS) reported the myopia prevalence to be 1.4% among 6‐year‐olds (n = 1765) with 0.8% in the White children and 2.7% among other ethnic groups.42 Among 12‐year‐old children (n = 2353), the overall myopia prevalence was 11.9%, which was lower among European Caucasian children (4.6%) and Middle Eastern children (6.1%) and higher among East Asian (39.5%) and South Asian (31.5%) children,43 although the sample size of non‐White groups in SMS was very small. In the Orinda Longitudinal Study of Myopia (OLSM), the prevalence of myopia increased from 4.5% in 6–7‐year‐old children to 28% in 12‐year‐old children in a predominantly white population in the United States.44 In the USA Collaborative Longitudinal Evaluation of Ethnicity and Refractive Error (CLEERE), Asians had the highest prevalence (18.5%), followed by Hispanics (13.2%). Whites had the lowest prevalence of myopia (4.4%), which was not significantly different from African Americans (6.6%). In the CLEERE study, however, children with different ethnicities were from different geographical areas so that the comparison of prevalence was affected by both genetic and environmental factors.45
In a Swedish school‐based sample of 1045 children aged from 12 to 13 years, refraction was performed using 1 drop of 0.5% tropicamide and measured by retinoscopy. The prevalence of myopia (SE ≤ −0.5 D) was reported to be 49.7% and the prevalence of bilateral myopia was reported to be 39.0%.46 In another study in the UK, non‐cycloplegic autorefraction data were available for 7554 children at the age of 7 from a birth cohort study. Using a definition of ‘likely to be myopic’ as SE ≤ −1.50 D, this study reported a prevalence of myopia of 1.5% in 7‐year‐old white children.47 The Northern Ireland Childhood Errors of Refraction study, a population‐based cross‐sectional study, examined 661 white 12–13‐year‐olds and 392 white 6–7‐year‐old children between 2006 and 2008. The prevalence of myopia was reported to be 2.8% (95% CI 1.3, 4.3) in the 6–7‐year‐old age group and 17.7% (95% CI 13.2, 22.2) in the 12–13‐year‐old age group.48 The Aston Eye Study, an ongoing multi‐racial sample of school children from the metropolitan area of Birmingham, England, reported preliminary cross‐sectional data on 213 South Asian, 44 black African Caribbean and 70 white European children aged 6–7 years and 114 South Asian, 40 black African Caribbean and 115 white European children aged 12–13 years and found that myopia prevalence was 9.4% and 29.4% for the two age groups, respectively. Ethnic differences in myopia prevalence were found with South Asian children having higher levels than white European children (36.8% vs 18.6%) for the children aged 12–13 years.49 The Child Heart and Health Study in England used population‐based sampling stratified by socioeconomic status and reported the prevalence of myopia to be 3.4% in White children aged 10–11 years. However, non‐cycloplegic refraction in this study might have led to an overestimation of the myopia prevalence.50 In Greece and Bulgaria, four schools from the centre of a Greek city were chosen and two schools from the centre of a Bulgarian city. Non‐cycloplegic auto‐refraction was performed on children aged 10–15 years. The prevalence of myopia (SE ≤ −0.75 D) was 37.2% in Greek children and 13.5% in Bulgarian children.51
In summary, the prevalence of myopia in Chinese children is higher than other ethnic groups. Moreover, the prevalence of myopia in European children seems to be lower than that in Asian children generally. Data from most studies have also documented a clear urban–rural difference in the prevalence of myopia. Studies on populations with very similar genetic backgrounds growing up in different environments in India, Nepal and China have shown that those growing up in rural environments have a lower prevalence of myopia. For the Chinese ethnicity, the prevalence of myopia in cities such as Guangzhou and Hong Kong is comparable to those reported for Singapore and urban areas of Taiwan. However, recent evidence showed that the prevalence in rural southern China is also very high. Whether this high prevalence of myopia in rural China is due to rapid economic development and high educational achievement is unclear.
Environmental risk factors of myopia and axial length
Outdoor activities
In Australia, students who performed high levels of near work but low levels of outdoor activity had the least hyperopic mean refraction. On the other hand, those who carried out low levels of near work but high levels of outdoor activity had the most hyperopic mean refraction. Furthermore, in an analysis combining the amount of outdoor activity and near work activity spent, children with low outdoor time and high near work were two to three times more likely to be myopic compared to those performing low near work and high outdoor activities.52
In Singapore, a cross‐sectional study was conducted to analyze the effect of outdoor activities on 1249 teenagers aged 11–20 years (71.1%, Chinese, 20.7% Malays and 0.8% other ethnicities). After adjusting for confounders, there was a significant negative association between myopia and outdoor activity. Adjusting for the same confounders, for each hour increase in outdoor activity per day, SE increased by 0.17 D (i.e. a hyperopic shift) and the AL decreased by 0.06 mm.53
The OLSM found that children who became myopic (SE < −0.75 D) by the 8th grade spent less time in sports and outdoor activity (hours per week) at the 3rd grade compared to those who did not become myopic (7.98 ± 6.54 h vs 11.65 ± 6.97 h). In predictive models for future myopia, the combined amount of sports and outdoor hours per week was predictive of future myopia.54
Additional recent studies have found that outdoor activity is an independent factor negatively associated with myopia. The Sydney Myopia Study measured both near work and outdoor activities simultaneously and found that near work activities had little impact on refraction.52 This study also found no effect of indoor sport on myopia, which implicates that more time spent outdoors, rather than sport itself, as the essential protective factor. A recent animal study on chicks found that light intensity modulates the process of emmetropization and that a low intensity of ambient light is a risk factor for developing myopia.55 To answer questions related to cause and effect, randomized clinical trials (RCT) of community‐based health behaviour interventions may be conducted. In Singapore, a RCT on children aged 7–10 years using a novel incentive‐based family intervention to increase time spent outdoors is ongoing. This study aims to examine the hypothesis that children in the intervention group will show smaller shifts of refraction toward myopia as a result of increased outdoor time.
The biological mechanism behind this association is not yet clearly understood. It is postulated that higher light intensity outdoors could make the depth of field greater and reduce image blur. In addition, the release of dopamine from the retina is stimulated by light, and dopamine can inhibit eye growth.52 However, the hypothesis that it is the high light intensity outdoors that is crucial has been contradicted by a study suggesting that it is the spectral composition of the light, rather than the intensity, which is the primary cause of the tendency for myopia to be associated with more time indoors.56 In a recent animal study, chicks exposed to high illuminances (15 000 lux) for 5 h per day significantly slowed compensation for negative lenses compared with those under 500 lux. Compensation for positive lenses was accelerated by exposure to high illuminances but the end point refraction was unchanged, compared with that of the 500‐lux group. High illuminance also reduced deprivation myopia by roughly 60%, compared with that seen under 500 lux. This protective effect was abolished by the daily injection of spiperone, a dopamine receptor antagonist. This study showed that the retardation of myopia development by light is partially mediated by dopamine.57 A very recent animal study (Smith et al., 2011 ARVO e‐abstract 3922) showed that high‐light‐reared monkeys exhibited significantly lower average degrees of myopic anisometropia (+0.14 ± 4.12 vs −3.56 ± 3.33 D, p = 0.04) and average treated‐eye refractive errors that were significantly more hyperopic than those observed in monocularly form‐deprived monkeys reared under normal light levels (+4.44 ± 5.24 vs −0.65 ± 3.84 D, p = 0.03). Thus, high ambient light levels can dramatically retard the development of form‐deprivation myopia. This study indicated that absolute light levels are a fundamental variable impacting the vision‐dependent regulation of ocular growth in primates and suggested that the seemingly protective effects of outdoor activities against myopia in children are due to exposure to the higher light levels normally encountered in outdoor environments. In a recent publication, Charman hypothesized that a consistent relationship between the astigmatic image fields and the retina are likely to be favourable to peripherally‐based emmetropization. This condition is satisfied by outdoor environments, since dioptric stimuli may not vary widely across the visual field.58
Near work
In the SMS, near work was quantified by the continuous time and close reading distance in 12‐year‐old children.59 Children who read continuously for more than 30 min were more likely to develop myopia compared to those who read for <30 min continuously. Meanwhile, children who performed near‐work at a distance of <30 cm were 2.5 times more likely to have myopia than those who worked at a longer distance. Similarly, children who spent a longer time reading for pleasure and those who read at a distance closer than 30 cm were more likely have higher myopic refractions.
The SCORM found that children who read more than two books per week were about three times more likely to have higher myopia (SE < −3.0 D) compared with those who read <2 books per week. Children who read for more than 2 h a day were 1.5 times more likely to have higher myopia compared to those who read <2 h, but this was not significant. Every book read per week, was associated with an AL elongation of 0.04 mm. Children who read more than two books per week had 0.17 mm longer axial lengths compared to children who read two or fewer books per week.36
The OLSM examined 366 eighth‐grade predominantly Caucasian children and found that the OR of myopia (SE < −0.75 D) was 1.02 (95% CI 1.008, 1.032) for every dioptre‐hours of near work spent per week, after controlling for parental myopia and achievement scores.60
Near work was also shown not to be associated with myopia in several other studies.61, 62 In a 5‐year follow‐up longitudinal study on 1318 children aged 6–14 years, hours per week spent reading or using a computer did not differ between the groups before myopia onset. Studying and TV watching were also not significantly different before myopia onset. This study failed to show evidence of a relationship between near visual activities and the development of myopia.63 Most studies on myopia and near work are cross‐sectional which cannot examine the temporal relationship between outcomes and predictors. It is also likely that myopes engage in more near work as as it is more difficult to take part in some sporting tasks due to spectacle wear. A prospective study reported that myopic children may be more at risk of having lower levels of physical activity than their non‐myopic peers.64 This argument should be resolved by more prospective studies with longitudinal evidence. In addition, most information on near work and time outdoors in previous studies were reported by parents. Thus, recall bias or reporting bias may have occurred. In the future more accurate and more tightly standardised methodology for quantifying near work needs to be used, which should facilitate precise comparison between different studies. Some modifiable kinds of near work, such as reading posture, breaks during reading, and proper lighting should also be studied so that children could benefit through health promotion efforts of modifiable behaviour.65
Education
Numerous studies that have examined the effect of education on myopia have found a consistent correlation between higher educational level and higher prevalence of myopia.19, 21, 26, 66 There appears to be an association between myopia and higher academic achievements as well.60, 67, 68 In a study on the Chinese children in Singapore and Sydney, early schooling in Singapore has also been found to be associated with the high levels of myopia compared with schooling in Sydney.69 This study indicated that exposure to a more intensive schooling system at an early age may be an independent risk factor for myopia. Higher educational level was also positively associated with longer AL. In Singapore Malay adults, increasing AL was associated with higher educational levels (standardized regression coefficient = 0.118, p < 0.001).70 In Singapore Chinese adults, an AL increase of 0.60 mm is associated with every 10 years of education.71
In epidemiological studies, educational level is usually measured either as years of formal education or level of academic achievement. Both the duration and level of education are highly correlated with time spent on reading and writing. Hence, educational level may be a surrogate for near work.72 Meanwhile, the association between education and myopia may also reflect common genetics of intelligence and refraction.
Parental myopia
In the SMS, children with one and two myopic parents had two times and eight times higher risks, respectively, of developing myopia (SE ≤ −0.5 D) compared to those with no myopic parents. In addition, an increasing severity of parental myopia led to a greater risk of myopia. The odds ratio for mild myopia (SE −0.5 to −3 D), moderate myopia (SE −3 to −6 D) and high myopia (SE at least −6 D) were 6.4 (95% CI 1.5, 27.8), 10.2 (95% CI 2.6, 40.1) and 21.8 (95% CI 5.3, 89.4) respectively.73
It was also reported that children with myopic parents have longer AL than those without myopic parents. Zadnik et al. investigated 716 Caucasian children aged 6–14 years and demonstrated that the pre‐myopic eyes in children with myopic parents had a longer AL than those without myopic parents. This suggests that the size of the pre‐myopic eyes might be already influenced by parental myopia. Moreover, it was found that children with two myopic parents developed myopia more often (11%) than children with one myopic parent (5%) or children without myopic parents (2%). (SE ≤ −0.75 D).74
The SCORM cohort showed that having one and two myopic parents was associated with an increase in AL of 0.14 and 0.32 mm, respectively, compared with no myopic parents. The study also showed that having one myopic parent and two myopic parents increased the degree of myopia by 0.39 and 0.74 D, respectively.36
Most studies have shown a consistently higher prevalence of myopia among those with myopic parents as compared with those without. Parental myopia is considered as a marker for both genes and a shared family environmental exposure. Myopic parents are more likely to create myopigenic environments such as more intensive education or less time spent outdoors.60, 73, 75
The gene‐environment interaction for myopia is still inconclusive. The SCORM study found an interaction between parental myopia and near‐work. However, both the OLSM and the SMS found all children are protected by outdoor activities but the risk declined in parallel for children with and without myopic parents, indicating there might be no interaction between outdoor activities and parental myopia. Since myopic parents may create myopigenic environments for their children, interaction observed between parental myopia and near‐work may not represent gene‐environment interaction.
Peripheral refraction
Central refractive error is determined by foveal vision on the visual axis. However, the foveal area is only a small part of the overall visual field and more peripheral retinal areas might also be important in refractive status. Animal studies have shown that the peripheral retina plays an important role in determining eye growth. A study in monkeys in which the central region of the retina was ablated demonstrated that treated eyes recovered as quickly from visual deprivation or lens‐induced myopia as did untreated eyes.76 This suggests that the peripheral retina has an effect on AL growth, and may participate in the process of emmetropization.
Human studies on peripheral refraction have been largely conducted on Caucasians. The OLSM assessed peripheral refractive error in 822 children aged 5–14 years. This study indicated that myopic children had greater relative hyperopia in the periphery, compared to emmetropes and hyperopes.77 Another study included 116 subjects in the age range 18–35 years and reported that myopia had more effect on peripheral refraction along the horizontal rather than vertical visual field.78
A longitudinal study on 605 children aged 6–14 years explored ethnic differences and found that Asian‐Americans (n = 579) had the largest degree of relative peripheral hyperopia, whereas African‐Americans who were myopic (n = 724) had no significant peripheral hyperopia.79
One study determined relative peripheral refractive error in eyes of a group of Chinese. Central and peripheral refractive errors were obtained from cyclopleged eyes of 40 children and 42 adults. In this study, subjects with moderate myopia had relatively greater hyperopic shifts in the periphery than those with low hyperopia who showed a myopic shift (p < 0.05).80
In a recent study in Singapore, 250 Chinese children with a mean age of 83 months were included in analysis. This study found that children with high and moderate myopia had relative hyperopia at all peripheral eccentricities (p < 0.001), whereas children with low myopia had relative hyperopia only at the temporal and nasal 30° (p < 0.001), but not at the nasal and temporal 15°. Children with emmetropia and hyperopia had peripheral relative myopia at all eccentricities (p < 0.001).81
Animal and human evidence indicates that relative peripheral hyperopia occurs in tandem with a prolate shape of the eyeball in myopic individuals. Longitudinal studies are needed to provide evidence for peripheral refraction determining the onset of myopia. However, longitudinal data with this research aim are few. In a longitudinal study on 187 children (mean age: 7.2 years) in Singapore, cycloplegic refraction was performed at five eccentricities: central axis and 15° and 30° eccentricities in the nasal and temporal visual fields. At follow‐up, children who remained non‐myopic (n = 24) retained relative peripheral myopia at all eccentricities, whereas those who became myopic (n = 67) developed relative peripheral hyperopia at the nasal (+0.44 ± 0.72 D) and temporal 30° (+0.13 ± 0.74 D). This study showed that baseline peripheral refraction did not predict the subsequent onset of myopia or influence myopia progression.82
Animal models of myopia
In animal models, macaque monkeys with surgically fused eyelids, i.e. form deprivation, experienced excessive AL elongation and eventually developed myopia.83 Another early study on chicks found that monocular deprivation of form vision also produced myopia and eye enlargement.84 These landmark studies ushered a new era in experimental myopia study and in the years since, models of form deprivation of myopia have been developed in a wide variety of animal species, including chicks,85, 86 tree shrews,87, 88 guinea pigs89, 90 and adult monkeys.91 Other experimental methods using positive or negative lens as modulators of refractive error in chicks showed that the eye grows more slowly (developed hyperopia) or more rapidly (developed myopia), respectively.85 Recent experiments also indicated that the low levels of lighting in laboratories played a major part in the development of myopia in these animal models of myopia, as they appear to be directly countered by high light levels.57 (Smith et al., 2011 ARVO e‐abstract 3922) The experimental models of myopia suggest that both retinal image degradation (hyperopic and myopic defocus) and accommodation play important roles in AL elongation and myopia formation in animals.92 Experimental models of myopia appear to suggest an important role of environmental factors in degradation of image quality, which could lead to myopia development.83, 84, 87 The latest animal study on chicks also found that genetic factors are the major determinant of susceptibility to myopia induced by retinal image degradation. Selective breeding for susceptibility to myopia reveals a gene‐environment interaction on refractive development.93 However, questions remain on the applicability of animal models of myopia to physiological human myopia.94
Conclusions
Population‐based data in children indicate that Asian populations, especially those of Chinese ethnicity, may be more susceptible to myopia compared with Western populations. However, as for adults, the situation is more complex. The prevalence of myopia in Singapore Chinese adults was only slightly higher than similarly aged white populations. In addition, studies in Taiwan, Singapore and China indicated that the rates of myopia in other Asian adult populations are not much higher than rates in White adult populations. This is possibly due to the expansion of mass intensive education in some areas such as China, Taiwan and Singapore. While the overall rates of myopia may vary between Asian populations, most studies demonstrate a clear trend of declining prevalence of myopia and mean AL with age, with younger participants generally having higher myopia rates and longer AL than older participants, which may be attributable to differences in birth cohorts and age.
The precise biological mechanisms through which the environment influences ocular refraction in humans are, however, still a matter of debate. It is still controversial that exogenous variables interact with heritable factors to modulate eye growth during ocular development. To date, time spent outdoors is an important modifiable environmental factor that may play an important role in our efforts to prevent myopic refractive shifts in children.
Appendices
For details of the authors of this review please see the next page.#5;
Chen‐Wei Pan
Chen‐Wei Panis a PhD candidate under Professor Seang‐Mei Saw’s supervision at the National University of Singapore. His research area covers the prevalence of myopia, genetic and environmental risk factors for myopia, and the association between ocular biometry and myopia. He has published several papers on the epidemiology of myopia in Investigative Ophthalmology & Visual Science.
Dr Dharani Ramamurthy
Dr Dharani Ramamurthy is a research fellow working with Professor Seang‐Mei Saw at the National University of Singapore. She obtained her Bachelor in Optometry degree from the Elite School of Optometry, India and her PhD degree from Anglia Ruskin University, UK with the thesis titled ‘Effect of family history of myopia and other known risk factors on myopia progression’. Her research interests are myopia research and relevant risk factors, the influence of accommodation on myopia and the control of myopia progression.
Seang‐Mei Saw
Prof Seang‐Mei Saw is currently a Professor at the Saw Swee Hock School of Public Health, and Vice‐Dean (Research), Yong Loo Lin School of Medicine, National University of Singapore. She received her MBBS degree from the National University of Singapore and both her MPH and PhD from the Johns Hopkins Bloomberg School of Public Health, Baltimore, USA. She has published more than 200 peer‐reviewed international journals, including the Lancet and Journal of the American Medical Association (JAMA).
