Differences in Prevalence of Refractive Errors in Young Asian Males in Singapore between 1996-1997 and 2009-2010

Purpose: To determine the differences in prevalence rates of refractive errors in young Singaporean males over a 13-year period between 1996–1997 and 2009–2010 Methods: Non-cycloplegic autorefraction data were obtained in a survey of 15,085 young male subjects (mean age 19.5 ± 1.4 years, range 16–25 years) in 1996–1997 using a Nikon NRK-8000 autorefractor. A second survey of 28,908 male subjects (mean age 19.8 ± 1.2 years, range 17–29 years) was conducted in 2009–2010 using a Huvitz MRK-3100 autorefractor. Results: The overall prevalence of myopia (spherical equivalent, SEq, worse than −0.5 diopters, D) was similar in 2009–2010 (81.6%) and 1996–1997 (79.2%; odds ratio 1.38, 95% confidence interval 1.31–1.46; p < 0.001). The prevalence of high myopia (SEq worse than −6.0D) increased from 13.1% in 1996–1997 to 14.7% in 2009–2010 (p < 0.001). The prevalence of refractive astigmatism increased significantly from 41.4% in 1996–1997 to 50.9% in 2009–2010 (p < 0.001). Conclusions: Our results show that the prevalence of myopia and high myopia remained high and the prevalence of astigmatism increased in young male adults in Singapore over a 13-year period after controlling for age, education and ethnicity.

Keywords: Astigmatism; epidemiology; myopia; refractive error; Singapore

Refractive errors, especially myopia, are among the most common causes of visual impairment in the world.[[1]] Unlike many ocular diseases, refractive error can be easily corrected if there are no myopia-related complications which affect the macula permanently, such as lacquer cracks, choroidal neovascularization or retinal detachment. Such complications tend to affect individuals with pathological myopia due to an excessively elongated globe. The prevalence of myopia among East Asians is one of the highest in the world.[7],[8]

Although there are multiple studies on the prevalence of myopia in white and Asian populations, there is a lack of longitudinal data on the prevalence of refractive error over time. To date, only two studies, one from Israel among a military cohort (comprised of young military conscripts aged between 16 and 22 years) and the other from the United States using data from the National Health and Nutrition Examination Survey (NHANES; comprised of subjects aged between 12 and 54 years), have reported increasing trends of refractive error. Results from Israel showed that myopia prevalence was 20.3% in 1990 and 28.3% in 2002,[9] and a similar pattern emerged from the NHANES data with 25.0% in 1971–1972 and 41.6% in 1999–2004.[10]

Singapore is a small, urbanized country with a total population of approximately 5.3 million and a population density of 7,422/km2. The major ethnic groups include Chinese (74.2%), Malay (13.3%) and Indian (9.2%). The median age is 37.4 years and 73.7% of the population is between 15 and 64 years old. The literacy rate is 96% and the gross domestic product is 345.6 billion Singapore dollars.[11] The risk of developing myopia at a young age in Singapore is very high. Since the late 1970s, the prevalence of myopia in Singapore has been on an upward trend. Starting from as low as 26.3% in 1978–1984, the prevalence of myopia increased to 44.2% in 1987–1992.[12],[13] Myopia was recognized as a potential public health problem and the National Myopia Prevention Programme (NMPP) was introduced in 2001. The aim of the NMPP was to prevent and reduce myopia progression through public education and to increase awareness of myopia and vision screening in school age children. To our knowledge, there is no recent longitudinal data on the differences in prevalence rates of refractive errors in Asia.

Therefore, our study aimed to examine differences in the prevalence of refractive errors including myopia, hypermetropia, astigmatism and anisometropia in two groups of young Asian males in Singapore in 1996–1997 and 2009–2010.

This is an observational cohort study comparing the prevalences of refractive errors in young Asian males in Singapore in 1996–1997 and 2009–2010. The exclusion criteria for both cohorts included history of ocular pathology that affected the vision of either eye, such as cataract, retinal detachment, glaucoma and history of refractive surgery. Subject demographics collected included age, ethnicity and highest education level attained. Education level was categorized as primary school education (completed up to 6 years of education) and above primary school education (completed more than 6 years of education). This study was approved by the institutional review boards of Singhealth and DSO National Laboratories, and followed the tenets of the Declaration of Helsinki.

A total of 43,993 consecutive young males who underwent pre-employment screening for the public service for the period of 1 year from either 1996–1997 or from 2009–2010 were recruited for participation. All subjects were asked to complete a questionnaire which included subject demographics, and medical and refraction history.[8]

This first cohort consisted of 15,085 males with a mean age of 19.5 ± 1.4 years (range 16–25 years). Refractive data was obtained with a Nikon NRK-8000 autorefractor (Nikon, Tokyo, Japan), which provides measures of spherical power ranging from −18.0 diopters (D) to +23.0 D, cylindrical power from −12 D to +12 D and cylindrical axis from 1–180°. The minimum pupil diameter compatible with the Nikon NRK-8000 autorefractor is 2.9 mm. The mean of three measurements of spherical power, cylindrical power and cylindrical axis was taken as the final reading.

The second cohort consisted of 28,908 males with a mean age of 19.8 ± 1.2 years (range 17–29 years). Non-cycloplegic refractive data was obtained using a Huvitz MRK-3100P (Huvitz, Geumjeong-dong, Korea) autorefractor which measures spherical power ranging from −25.0 D to +25.0 D, cylindrical power from −10 D to +10 D and cylindrical axis from 1–180°. The minimum pupil diameter compatible with the Huvitz MRK-3100P autorefractor is 2.0 mm. The mean of three measurements of spherical power, cylindrical power and cylindrical axis was taken as the final reading.

Spherical equivalent (SEq) was defined as sphere plus half negative cylinder. Myopia was defined as SEq <−0.5 D and high myopia as <−6.0 D. Hypermetropia was defined as >+0.5 D, astigmatism as >−0.5 D cylinder and anisometropia as >1.0 D SEq absolute difference between both eyes. Astigmatism was stratified into hypermetropic, simple and myopic. This was further stratified by the axes of astigmatism (with-the-rule, WTR, against-the-rule, ATR, and oblique) using the negative cylinder convention. Astigmatism within 15° of 180° was classified as WTR, 15° of 90° was classified as ATR and other axes were categorized as oblique. Simple astigmatism was defined as the presence of astigmatism and spherical refractive error between +0.5 D and −0.5 D. Myopic astigmatism was defined as the presence of astigmatism with spherical refractive error <−0.5 D. Hypermetropic astigmatism was defined as the presence of astigmatism with spherical refractive error >+0.5 D.

The prevalence and distribution of myopia, hypermetropia and astigmatism in the two cohorts were analyzed using only the right eye, as the correlation between right and left eyes was high for refractive error (Pearson correlation coefficient 0.94). Prevalence and magnitude of anisometropia were analyzed using data from both eyes. Chi-square tests and z-tests were used to evaluate statistical differences between the two cohorts. Differences in prevalence of myopia, high myopia, astigmatism, and anisometropia between the two cohorts were analyzed using a logistic regression model with adjustment for age, education and ethnicity. Odds ratios (ORs) with 95% confidence intervals (CIs) are reported. Refractive error of each cohort was further stratified by ethnicity and 95% CIs for the prevalence of each refractive error by ethnicity were calculated. Statistical analysis was conducted using the SPSS version 19.0 (IBM Corp, Armonk, NY, USA).

Of 15,094 subjects in the 1996–1997 cohort, 9 were excluded (3 had congenital cataract, 4 had corneal refractive surgery and 2 had ocular trauma) and the remaining 15,085 subjects were included in the analyses. Of 28,916 subjects in the 2009–2010 cohort, 8 were excluded (1 had congenital cataract and 7 underwent corneal refractive surgery) and the remaining 28,908 subjects were included in the analyses.

The proportions of Chinese, Malay and Indian subjects in the 1996–1997 cohort were 82.0%, 12.8% and 4.3%, respectively. In the 2009–2010 cohort, the proportions were 69.2%, 21.3% and 7.6%, respectively. The proportion of primary school and above primary school education in the 1996–1997 cohort was 14.2% and 85.8%, respectively. In the 2009–2010 cohort, 2.6% and 97.4% had primary school and above primary school education, respectively.

The distribution of SEq refractive errors for each cohort is shown in Figure 1. The median SEq refraction was the same in both cohorts at −2.25 D (p = 0.127). The overall prevalence of myopia remained similar; 79.2% in 1996–1997 and 81.6% in 2009–2010 (p < 0.001; Table 1). The prevalence of high myopia increased from 13.1% in 1996–1997 to 14.7% in 2009–2010 (p < 0.001). After adjusting for age, education and ethnicity, the prevalence for myopia and high myopia remained significantly higher in the later cohort. The increase in prevalence of myopia between the two cohorts was more marked in Malay men (from 64.9 to 70.7%; p < 0.001) and Indian men (from 68.6 to 74.6%; p = 0.030). There was only a slight increase in prevalence of myopia in Chinese subjects from 82.2 to 85.9% (p < 0.001; Table 2). The prevalence of high myopia increased slightly in Chinese subjects from 14.8 to 18.2% (p < 0.001) and Malay subjects from 5.0 to 6.6% (p = 0.011).

Graph: FIGURE 1. Distribution of spherical equivalent refraction (a) and astigmatism (b) in right eyes of subjects in cohorts from 1996–1997 and 2009–2010, Singapore.

TABLE 1. Prevalence and odds ratios of refractive errors in the two cohorts, Singapore.

5 aAdjusted for age, education and ethnicity. CI, confidence interval; D, diopter; OR, odds ratio

TABLE 2. Comparison of prevalence of refractive errors between two cohorts by ethnicity, Singapore.

6 aChi-square test comparing the proportions of refractive error between the 1996–1997 and 2009–2010 cohorts CI, confidence interval; D, diopter

Median astigmatism was −0.50 D in the 1996–1997 survey and −0.75D in the 2009–2010 survey (p < 0.001). This difference in astigmatism is shown in Figure 1 with the distribution being slightly more myopic in the 2009–2010 compared to the 1996–1997 survey. This increase in prevalence of astigmatism was statistically significant: 41.4% with astigmatism in 1996–1997 compared to 50.9% in 2009–2010 (p < 0.001; Tables 1 and 3). Subjects from the 2009–2010 cohort were more likely to have astigmatism compared to the 1996–1997 cohort (OR 1.62, 95% CI 1.55–1.69; Table 1). Results presented in Table 3 show that myopic astigmatism was the more prevalent form of astigmatism and remained similar in both cohorts. In contrast, the prevalence of simple astigmatism decreased slightly from 9.9 to 8.8% over the same period (p < 0.001). The most common form of astigmatism was WTR for both studies at 51.4% in 1996–1997 and 60.5% in 2009–2010.

TABLE 3. Prevalence of different axes of astigmatism stratified by spherical equivalent refraction over the two cohorts, Singapore.

7 ap value compares proportions between the 1996–1997 and 2009–2010 cohorts. CI, confidence interval; D, diopter; ATR, against-the-rule (within 15° of 90°); WTR, with-the-rule (within 15° of 180°);oblique, all other axes

The prevalence of hypermetropia (SEq >+0.5 D) remained similar: 0.8% in 1996–1997 and 0.9% in 2009–2010 (Table 1).

Table 1 shows that the prevalence of anisometropia remained similar between the two cohorts after adjusting for age, education and ethnicity: 17.9% in 1996–1997 compared to 18.3% in 2009–2010 (p < 0.001). Table 1 shows that persons from the 2009–2010 cohort were more likely to have anisometropia compared to the 1996–1997 cohort (p = 0.023). Differences across the three ethnicities are presented in Table 2.

Our study showed that the prevalence of myopia and high myopia remained high in young male adults in Singapore over a 13-year period after stratifying for age, education and ethnicity. In addition, our results showed that there was a significant increase in the prevalence of astigmatism in this young age group over the same period. Our data suggested that the prevalence of myopia and high myopia in this young age group is not decreasing in Singapore. Table 4 gives a summary of studies examining the prevalence of myopia in young Asian and White men.[[14]] In particular, there were two studies on Korean male conscripts who were living in urban and rural environment.[15],[16] The higher prevalence of myopia and high myopia in the urban population suggested that living environment may play a significant role in the development of myopia. Compared to young Danish men, the prevalence of myopia and high myopia were considerably higher in Asian men.[18] Interestingly, the Jacobsen et al also compared their results (year 2007) comprised of Danish conscripts with previous similar studies in year 1882 and 1964 and concluded that there is decreasing prevalence of myopia and high myopia in Denmark.

TABLE 4. Summary of population-based studies on the prevalence of myopia in young Asian and White males.

8 aMyopia defined as spherical equivalent < −0.50 to −0.75 diopters. High myopia defined as spherical equivalent <−6.00 to −6.50 diopters.

In Singapore, the NMPP helps to detect myopia early and allows affected children to be referred appropriately to refraction clinics or ophthalmologists for further assessment and intervention if necessary. In contrast to the 1996–1997 cohort, the subjects included in the 2009–2010 survey comprised of young male subjects who were exposed to NMPP after its introduction. The results of our study between year 1996 and 2009 suggested that the prevalence of myopia and high myopia did not decrease despite the measures of NMPP for the past 10 years. This highlighted the difficulty of myopia prevention and treatment due to the multi-factorial nature of myopia development including influence from genetic and environmental factors. Interestingly, in several multi-ethnic population-based studies in Singapore comprised of older adults aged above 40 years, the prevalence of myopia and high myopia was consistently lower compared to our study of a much younger age group.[7],[19],[20] Apart from genetic and environmental factors, this could be attributed to increasing popularity of cornea refractive surgery and earlier cataract surgery combined with correction of refractive errors in Singapore. Previous studies have also shown a hypermetropic shift after the age of 40 years old which was postulated to be due to reduction in vitreous depth with age.[[21]]

Compared to previous longitudinal studies conducted in Israel and the USA,[10] the rise in prevalence of myopia in the present study was consistent but the magnitude was much smaller. The Israeli military performed repeated surveys on 919,929 subjects (382,139 females and 537,790 males, age 16 to 22 years old, year 1990 to year 2002).[9] Refraction was done using non-cycloplegic autorefraction, validated by subjective visual acuity. Compared to our study, the study in Israel showed a significant increase in the prevalence of myopia from 20.3% in 1990 to 28.3% in 2002. Compared to our study, the larger upward trend in the prevalence of myopia in the Israeli military and NHANES could be attributable to differences in the study population demographics, methodology, definitions and national myopia programs. Alternatively, there could be a ceiling effect in our study resulting in a smaller increase in the prevalence of myopia considering the high baseline prevalence of myopia in Singapore. The results of our study may also be partly attributed to the differences in the proportion of ethnic groups and education levels between the two cohorts of young men.

Our results showed a high prevalence of high myopia in this young age group. In the NHANES, the prevalence of high myopia (SEq worse than or equal to −7.9 D) was also on the upward trend for the age group 18–24 years (p < 0.02) which was similar to our study population (age range 17–29 years).[10] However, the prevalence of high myopia in the NHANES was only 1.4%. This group of young high myopes needs to be identified early and strategies and interventions are required to help prevent sight-threatening complications of pathological myopia from developing. Currently, a larger prospective study is being conducted in Singapore involving young Asian males with high myopia which will better quantify and stratify the risk of retinal complications.

Our results showed that the prevalence of astigmatism has increased significantly over a 13-year period. The prevalence of astigmatism in Singapore was considerably higher than seen in NHANES which reported that the prevalence of astigmatism in the USA was 36.2% (95% CI, 34.9%, 37.5%).[24] Our results were consistent with another population-based study of older Singapore adults (age more than 40 years) which also showed a high prevalence of astigmatism (58.5%).[25] Our results also showed that the prevalence of myopic astigmatism remained the highest. The relationship between astigmatism and myopia is still controversial and more studies are required to examine the underlying mechanism. It could be attributed to the accommodation induced by astigmatism which in turn induces myopic shift or it may cause ocular misalignment problems that inhibit accommodation, thereby creating a hyperopic blur signal that is corrected through the induction of myopia.[[26]] Alternatively, ocular growth could induce astigmatism and myopia as the eye elongates and pulls on the ciliary muscles or zonules.[27] Clinically, astigmatism is a major cause of correctable visual impairment in the world and its prevalence was shown in multiple population-based studies to be comparable to myopia.[20],[29],[30] In younger children, it is associated with poor visual function and is a significant risk factor for amblyopia.[[31]] Interestingly, our study also showed an increase in the prevalence of WTR astigmatism and a decrease in the prevalence of ATR astigmatism between the two cohorts of young men. WTR astigmatism is the predominant subtype in the younger population and was shown to be associated with increasing severity of myopia.[[35]] Our study also showed a high prevalence of anisometropia which was consistent with other population-based studies.[30],[38] Similarly, if untreated in children, anisometropia is also a significant risk factor for amblyopia.[32],[39]

Several methodological issues need to be discussed. The strength of this population-based study was its representative sample of the country's multi-ethnic youth male population and the comparison of two cohorts of similar demographics separated by a 13-year period. However limitations need to be considered. First, the study population was comprised of only male subjects. Whether gender plays a role in the prevalence of refractive errors is still unknown with mixed results from other studies.[38],[[40]] As such, the results of our study need to be generalized with caution to the female gender. Second, non-cycloplegic refractions were obtained with different autorefractors. This was due to a long 13-year period between the two cohorts that made it impossible to use the same autorefractor. There have been studies which showed good agreement between different table-mounted autorefractors[[43]] but there was no data directly comparing the two autorefractors used in our study. Although our data showed that the prevalence of myopia and high myopia remained high in Singapore over a 13-year period, the magnitude of increase is small and could partly be attributed to biases from using different autorefractors, in particular incorrect classification of low myopes. However, our data did suggest that the prevalence of myopia and high myopia in this young age group is not decreasing. Third, there was a difference in the way the questions on academic qualifications were structured in the questionnaire for each cohort. In the 1996--1997 survey, the possible choices stated the different levels of education and many would select their current academic level. As for the 2009--2010 survey, the instruction was printed to select the highest level of education completed. As such, a large number of the 2009–2010 cohort who were studying for their Diploma or 'A' levels would have only completed their secondary level education during the point of examination. To overcome this for further analysis, we stratified the education levels into two groups only (primary school education and above primary school education) because the higher levels of education were most affected by the differences in questionnaire design.

In conclusion, our study showed that the prevalence of myopia and high myopia in Singapore remained high and is not decreasing in this young Asian male population. The high prevalence of high myopia may present a significant burden on public health planning as complications associated with pathological myopia may threaten vision and quality of life in this young group of myopes. In addition, the prevalence of astigmatism also showed an increasing trend.

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.