Prevalence of Refractive Errors in a Multiethnic Asian Population: The Singapore Epidemiology of Eye Disease Study

Abstract

Purpose.: To determine the prevalence of refractive errors in a multiethnic Asian population aged over 40 years and to examine secular trends and racial differences.

Methods.: A total of 10,033 adults (3353 Chinese, 3400 Indians, and 3280 Malays) participated in this study. Refractive error was determined by subjective refraction. Ocular biometric parameters were determined by partial coherence interferometry. Myopia and high myopia were defined as spherical equivalent (SE) of less than −0.5 diopters (D) and −5.0 D, respectively. Hyperopia was defined as SE of more than 0.5 D. Astigmatism was defined as cylinders less than −0.5 D.

Results.: The prevalence of myopia, high myopia, hyperopia and astigmatism in Singapore adults aged over 40 years was 38.9% (95% confidence interval [CI] 37.1, 40.6); 8.4% (95% CI 8.0, 8.9); 31.5% (95% 30.5, 32.5); and 58.8% (95% CI 57.8, 59.9), respectively. Compared with the Tanjong Pagar Survey 12 years ago, there was a significant increase in the prevalence of astigmatism and mean axial length (AL) in Chinese adults aged over 40 years in Singapore. Chinese were most likely to be affected by myopia, high myopia, astigmatism, and had the longest AL among the three racial groups.

Conclusions.: The prevalence of myopia in Singapore adults is lower compared with the younger “myopia” generation in Singapore. The prevalence of astigmatism and mean AL have been increasing significantly within the past 12 years in the Chinese population. Chinese adults had higher prevalence of myopia, high myopia, astigmatism, as well as the longer AL compared with non-Chinese adults in Singapore.

Introduction

Myopia affects 1.6 billion people worldwide and 2.5 billion people are expected to be myopic by the year 2020. ¹ It is also estimated that globally, 153 million people aged over 5 years are visually impaired as a result of uncorrected refractive errors; of that number, 8 million are blind. ² The rapid increase in the prevalence of myopia was first noted in the Inuit population in North America. ^3–6 This trend has also been observed in the United States ⁷ as well as in Asian countries, including Singapore, ⁸ Taiwan, ^9–12 and mainland of China, ^13–14 where the prevalence of myopia in different birth cohorts can be compared. However, myopia results anatomically and physiologically from changes in biometric components including axial length (AL), corneal power, and lens. Understanding the generational change in AL, a major determinant of refractive status, would provide further insights into the increasing rates of myopia observed throughout the world. However, few studies have documented the secular changes in AL. In addition, it is unclear whether the prevalence of other refractive errors such as hyperopia and astigmatism has been changed in addition to myopia during the past decades. 

Singapore, a highly urbanized city-state located in South East Asia with an “epidemic” of myopia, ¹⁵ consists of Chinese, Malaysian, and Indian ancestries. Previous studies have reported the prevalence of myopia in preschool children, ^16–17 adolescents, ⁸ as well as middle-aged to elderly adults in selected racial groups ^18–20 in Singapore. Despite established health differentials between ethnic groups, there has been a lack of study exploring differences in refractive errors across different Asian racial groups using the same study protocol. The Chinese population is thought to have a higher prevalence of myopia than the other racial groups, but previous studies in Beijing (22.9%) ²¹ and Taiwan (19.4%) ²² reported relatively lower prevalence estimates. The inconsistencies may be attributable to the lack of a standardized protocol in prevalence surveys. Thus, there is a need to assess the racial differences in refractive errors using the same study protocols. 

In previous reports, we have described the prevalence of refractive errors and ocular biometric parameters individually for Singapore Chinese in the Tanjong Pagar Study (TPS) ^20,23 ; Singapore Malays in the Singapore Malay Eye Study (SiMES) ^18,24 ; and Singapore Indians in the Singapore Indian Eye Study (SINDI). ^19,25 In this study, we aimed to report the prevalence and risk factors for refractive errors and ocular biometric parameters in Singapore adults aged over 40 years in the Singapore Epidemiology of Eye Disease Study (SEED) of multiethnic cohorts. We also pooled the data of the Chinese cohort in SEED with the TPS, which was conducted 12 years ago, to elucidate the change in the prevalence of refractive errors and AL in Singapore Chinese aged over 40 years. In addition, we explored the racial differences in refractive errors as well as ocular biometric parameters amongst the three major Asian racial groups (Chinese, Malays, and Indians) living in Singapore using the same study protocols. 

Methods

Study Cohorts

The SEED program comprises three population-based studies conducted from 2004 through 2011: the SiMES (2004–2006); the SINDI (2007–2009); and the Singapore Chinese Eye Study (SCES, 2009–2011). The detailed methodologies of these studies have been published elsewhere. ^26–27 In brief, an age-stratified random sampling was used to select ethnic Malays, Indians, and Chinese aged 40 to 80 years living in Singapore during each stipulated study period. The overall response rate for SEED was 75.6%. The response rate was 78.7% for SiMES, ¹⁸ 75.6% for SINDI, ¹⁹ and 72.8% for the SCES, respectively. Study participants were older than nonparticipants (P < 0.001), but there was no sex difference (P = 0.68). 

This study was approved by the Singapore Eye Research Institute Institutional Review Board and the conduct of the studies adhered to the Declaration of Helsinki. 

Assessment of Refractive Error and Ocular Biometry

Noncycloplegic autorefraction was performed using an autorefractor (Canon RK-5 Autorefractor Keratometer; Canon, Inc., Tokyo, Japan). Refraction was then subjectively refined by the study optometrists until the best visual acuity was obtained. These subjective refraction results were used in analysis. Spherical equivalent (SE) was defined as sphere plus half cylinder. Myopia and high myopia were defined as SE of less than −0.5 diopters (D) and −5.0 D, respectively. Hyperopia was defined as SE of more than 0.5 D. Astigmatism was defined as a cylinder of less than −0.5 D. ^18,19  

Ocular biometric parameters of AL, anterior chamber depth (ACD), and corneal radius of curvature (CR) were measured using noncontact partial coherence interferometry (IOL Master V3.01; Carl Zeiss Meditec AG, Jena, Germany). ^24,25  

Risk Factor Assessment

Information on educational level (no formal education/primary education/high school/polytechnic/university), country of birth, and lifestyle-related factors was collected via a detailed questionnaire from all participants. Height was measured in centimeters using a wall-mounted measuring tape, after removing shoes. Slit-lamp examination (model BQ-900; Haag-Streit, Köniz, Switzerland) was performed after pupil dilation and included a clinical grading of cataract using the Lens Opacities Classification System (LOCS) III. ²⁸  

Statistical Analyses

Participants with prior cataract surgery were excluded from the analyses related to refractive error. As the Spearman correlation coefficients for SE (r = 0.86) and ocular biometric parameters in the left and right eye were high (r = 0.96 for AL; r = 0.89 for ACD; and r = 0.96 for CR), only right eye data were presented. 

The prevalence of refractive errors and mean ocular biometric parameters was estimated for the overall sample, and then stratified by age group and sex. Since our study included a similar numbers of participants for each racial group and nonresponders were younger compared with responders, the study sample may not be representative of the general population in Singapore. To estimate the prevalence of refractive errors in Singapore, the age-race standardized prevalence rates of refractive errors were calculated by direct standardization of the study samples to the Singapore population, using the 2010 Singapore census (provided in the public domain by http://www.singstat.gov.sg). The specific rates in a study population are averaged, using as weights the distribution of a specified standard population. The directly standardized rate represents what the crude rate would have been in the study population if that population had the same distribution as the standard population with respect to the variable(s) for which the adjustment or standardization was carried out. 

The differences in the prevalence of refractive error and mean AL between the SCES and the TPS were compared and stratified by age and sex. Since height has a significant impact on AL and we want to remove the effect of height, we adjusted for height when comparing AL between the SCES and the TPS by multivariate modeling. The racial differences in refractive errors as well as ocular biometric parameters were estimated initially adjusting for age and sex and further for other potential confounders such as educational level, height, and nuclear cataract, which are well-known risk factors for refractive errors. 

For risk factors, variables of interest were first assessed in univariate models. If the P value was less than 0.10 in univariate models, these possible predictors were included in multiple logistic regression models and manual backward stepwise elimination procedures were performed. 

All probabilities quoted were two-sided and were considered statistically significant when less than 0.05. Data analysis was conducted using statistical analysis software (PASW Statistics 18; SPSS, Inc., Chicago, IL). 

Results

After excluding participants with previous cataract surgery (n = 985) and no refraction data (n = 276), 8772 adults aged over 40 years contributed to the analysis related to refractive errors. Figure 1 demonstrates the distribution of refraction among myopic adults (SE < −0.5 D) by age groups in SEED. Table 1 shows the age-race standardized prevalence of refractive errors. The overall prevalence of myopia (SE < −0.5 D) and high myopia (SE < −5.0 D) was 38.9% (confidence interval [CI] 37.1, 40.6) and 8.4% (95% CI 8.0, 8.9), respectively. When myopia was defined as SE < −1.0 D, overall prevalence of myopia was 31.4%. The prevalence of myopia (SE < −0.5 D) was 37.9% in men and 39.8% in women while the prevalence of high myopia was 8.0% in men and 8.7% in women with no sex differences (P = 0.44 for myopia and P = 0.67 for high myopia). The age-specific prevalence of myopia (SE < −0.5 D) was 47.4%, 35.9%, 30.1%, and 32.7% in adults aged 40 to 49, 50 to 59, 60 to 69, and over 70 years, respectively. The overall age-race prevalence of hyperopia and astigmatism was 31.5% and 58.8%, respectively. The prevalence of hyperopia was 30.9% in men and 32.1% in women while the prevalence of astigmatism was 60.1% in men and 57.9% in women. There was no sex difference in the prevalence of hyperopia (P = 0.45) or astigmatism (P = 0.11). The prevalence of hyperopia was highest in adults aged 60 to 69 years (47.2%, 95% CI 45.1, 49.3) while the prevalence of astigmatism increased with increasing age (P < 0.001). 

Figure 1

 

Distribution of refraction among myopes in SEED.

Table 1. 

 

Prevalence of Refractive Errors in Singaporean Adults Aged Over 40 Years

		Myopia, <−0.5 D			Myopia, <−1.0 D			High Myopia, <−5.0 D			Hyperopia, >+0.5 D			Asigmatism, <−0.5 cylinder
			%	95% CI		%	95% CI		%	95% CI		%	95% CI		%	95% CI
Overall
Crude rate	8772	2641	30.1	29.1, 31.1	2027	23.1	22.2, 24.0	459	5.2	4.8, 5.7	3179	36.2	35.2, 37.2	4729	53.9	52.9, 55.0
Age-race standardized rate*			38.9	37.1, 40.6		31.4	30.5, 32.2		8.4	8.0, 8.9		31.5	30.5, 32.5		58.8	57.8, 59.9
Men
Crude rate	4321	1251	29.0	27.6, 30.3	954	22.1	20.8, 23.3	201	4.7	4.0, 5.3	1541	35.7	34.2, 37.1	2375	55.0	53.5, 56.4
Age-race standardized rate*			37.9	36.5, 39.2		30.7	29.4, 31.9		8.0	7.3, 8.6		30.9	29.4, 32.3		60.1	58.6, 61.5
Women
Crude rate	4451	1390	31.2	29.9, 32.6	1073	24.1	22.9, 5.4	258	5.8	5.1, 6.5	1638	36.8	35.4, 38.2	2354	52.9	51.4, 54.4
Age-race standardized rate*			39.8	38.5, 41.2		32.8	31.6, 34.0		8.7	8.0, 9.4		32.1	30.7, 33.5		57.9	56.4, 59.4
40–49 y
Crude rate	2437	925	38.0	36.0, 39.9	712	29.2	27.4, 31.0	184	7.6	6.5, 8.6	359	14.7	13.3, 16.1	921	37.8	35.9, 39.7
Race standardized rate*			47.4	45.4, 49.3		37.7	35.9, 39.5		11.5	10.4, 12.5		14.2	12.8, 15.6		44.0	42.1, 45.9
50–59 y
Crude rate	3013	840	27.9	26.3, 29.5	645	21.4	19.9, 22.9	155	5.1	4.4,5.9	1169	38.8	37.1, 40.5	1498	49.7	47.9, 51.5
Race standardized rate*			35.9	34.3, 37.5		28.8	27.3, 30.3		6.7	6.0, 7.5		33.0	31.3, 34.7		56.0	54.2, 57.8
60–69 y
Crude rate	2160	501	23.2	21.4, 25.0	379	17.6	15.9, 19.2	80	3.7	2.9, 4.5	1145	53.0	50.9, 55.1	1425	66.0	64.0, 68.0
Race standardized rate*			30.1	28.3, 31.9		23.2	21.5, 24.8		4.8	4.0, 5.6		47.2	45.1, 49.3		74.3	72.3, 76.3
70+ y
Crude rate	1162	375	32.3	29.6, 35.0	291	25.0	22.6, 27.5	40	3.4	2.4, 4.5	506	43.5	40.7, 46.4	885	76.2	73.7, 78.6
Race standardized rate*			32.7	30.0, 35.4		26.6	24.2, 29.1		4.5	3.5, 5.6		44.7	41.9, 47.6		87.1	84.6, 89.5

*  Standardized to the Singapore 2010 census population.

Table 2 shows the mean ocular biometric parameters by age and sex. The age-race standardized mean AL, ACD, and CR were 23.88 mm (95% CI, 23.85, 23.90), 3.27 mm (95% CI 3.26, 3.28) and 7.65 mm (95% CI 7.64, 7.66). Men had significantly longer AL (24.17 mm vs. 23.60 mm; P < 0.001); deeper ACD (3.33 mm vs. 3.21 mm; P < 0.001); and flatter CR (7.71 mm vs. 7.59 mm; P < 0.001) than women. There was a significant trend of decreasing AL with increasing age in both men and women (P < 0.001). ACD was deepest in adults aged 40 to 49 years. CR did not vary significantly across different age groups. 

Table 2. 

 

Mean Ocular Biometric Parameters in Singaporean Adults Aged Over 40 Years

		AL, mm		ACD, mm		CR, mm
		Mean	95% CI	Mean	95% CI	Mean	95% CI
Overall
Crude	9380	23.65	23.62, 23.67	3.22	3.21, 3.23	7.64	7.63, 7.65
Age-race standardized*		23.88	23.85, 23.90	3.27	3.26, 3.28	7.65	7.64, 7.66
All men
Crude	4617	23.90	23.86, 23.94	3.27	3.25, 3.28	7.70	7.69, 7.71
Age-race standardized*		24.17	24.13, 24.21	3.33	3.31, 3.34	7.71	7.70, 7.72
All women
Crude	4763	23.40	23.37, 23.43	3.17	3.15, 3.18	7.58	7.57, 7.59
Age-race standardized*		23.60	23.57, 23.63	3.21	3.19, 3.22	7.59	7.58, 7.60
Men
40–49 y
Crude	1166	24.00	23.93, 24.07	3.34	3.32, 3.36	7.70	7.69, 6.72
Race standardized*		24.29	24.22, 24.36	3.38	3.36, 3.40	7.71	7.70, 7.73
50–59 y
Crude	1414	23.96	23.89, 24.03	3.29	3.27, 3.31	7.70	7.69, 7.71
Race standardized*		24.21	24.14, 24.28	3.34	3.32, 3.36	7.70	7.69, 7.71
60–69 y
Crude	1222	23.88	23.81, 23.95	3.22	3.19, 3.24	7.70	7.68, 7.71
Race standardized*		24.14	24.07, 24.21	3.28	3.25, 3.30	7.72	7.70, 7.74
70+ y
Crude	815	23.68	23.60, 23.76	3.21	3.17, 3.25	7.68	7.66, 7.70
Race standardized*		23.84	23.76, 23.92	3.29	3.25, 3.33	7.69	7.67, 7.71
Women
40–49 y
Crude	1257	23.63	23.56, 23.70	3.26	3.24, 3.27	7.60	7.59, 7.61
Race standardized*		23.91	23.84, 23.98	3.31	3.29, 3.32	7.61	7.60, 7.62
50–59 y
Crude	1634	23.49	23.43, 23.56	3.16	3.15, 3.18	7.59	7.57, 7.60
Race standardized*		23.75	23.69, 23.82	3.31	3.20, 3.23	7.60	7.58, 7.62
60–69 y
Crude	1157	23.20	23.14, 23.26	3.06	3.04, 3.09	7.56	7.54, 7.57
Race standardized*		23.44	23.38, 23.50	3.11	3.09, 3.14	7.56	7.54, 7.57
70+ y
Crude	715	23.11	23.03, 23.18	3.18	3.13, 3.23	7.56	7.54, 7.58
Race standardized*		23.25	23.17, 23.32	3.24	3.19, 3.30	7.57	7.55, 7.59

*  Standardized to the Singapore 2010 census population.

Compared with the Chinese subjects in the TPS, Chinese adults in the SCES were less likely to be female (50.4% vs. 55.1%, P < 0.001); more educated (P < 0.001); had higher income (P < 0.001) and better housing (P < 0.001); less likely to be affected by nuclear cataract (9.2% vs. 34.2%, P < 0.001); and more likely to undergo cataract surgery (10.5% vs. 8.4%, P < 0.001). Table 3 compares the age and sex-specific prevalence of refractive errors in the TPS (1996–1997) and the SCES (2009–2011). The prevalence of myopia was higher in Chinese adults aged 40 to 49 years (P = 0.03 for men and 0.01 for women) and 50 to 59 years (P < 0.001 for both men and women) in the SCES compared with the TPS. The prevalence of astigmatism in the SCES was significantly higher in all age groups compared with the TPS (all P < 0.001). The major increase in astigmatism rate is due to the increase in myopic astigmatism rate, which accounted for 72% of the observed increasing rate in astigmatism, while hyperopic astigmatism only accounted for 16%. Figure 2 shows the height-adjusted mean AL by age and sex in the SCES and the TPS. After adjusting for the effect of height, Chinese in the SCES had significant longer AL in all age groups compared with their counterparts in the TPS (all P < 0.001). 

Figure 2

 

Height-adjusted AL in the TPS (1997–1998) and the SCES (2009–2011) by age and sex.

Table 3. 

 

Comparisons on the Prevalence of Refractive Errors Between the TPS (1997–1998) and the SCES (2009–2011)

	Myopia, %			High Myopia, %			Hyperopia, %			Astigmatism, %
	TPS	SCES		TPS	SCES		TPS	SCES		TPS	SCES
Men
40–49 y	45.2	47.5	0.03	11.3	12.7	0.08	12.1	14.7	0.44	25.8	47.2	<0.001
50–59 y	25.2	37.6	<0.001	4.3	6.8	0.11	42.6	32.2	0.01	33.9	60.1	<0.001
60–69 y	29.2	35.9	<0.01	3.8	5.6	0.01	42	40.2	0.68	49	78.4	<0.001
70+ y	31.7	34.7	0.05	1.0	5.8	<0.001	38.5	42.2	0.13	60	91.6	<0.001
Women
40–49 y	51.7	56.3	0.01	15.9	14.1	0.51	10.6	13.8	0.01	34.4	48.5	<0.001
50–59 y	27.1	40.8	<0.001	4.8	7.8	<0.001	46.8	29.4	<0.001	36.2	59.2	<0.001
60–69 y	30	28.8	0.67	6.0	4.8	0.04	52	51.1	0.88	53.3	76.6	<0.001
70+ y	40.3	30.6	<0.001	7.3	3.8	<0.001	38.7	48.1	<0.001	62.9	88.1	<0.001

Table 4 shows racial differences in the prevalence of refractive errors and ocular biometry among the three major racial groups in SEED. After adjusting for age and sex, Chinese had higher odds ratios of myopia, high myopia, and astigmatism as compared with Malays or Indians. There were no significant differences in the prevalence of myopia or high myopia between Malays and Indians, whereas Indians were more astigmatic than Malays. Hyperopia was most common in Indians among the three racial groups. After further adjusting for educational level, height and the presence of nuclear cataract did not alter the racial variations in all refractive errors. In addition, after adjusting for age, sex, height, educational level, and the presence of nuclear cataract, Chinese had 0.20 mm longer in AL and 0.09 mm deeper in ACD on average compared with Malays while Malays had 0.33 mm longer in AL compared with Indians on average. 

Table 4. 

 

Racial Differences in the Prevalence of Refractive Errors and Ocular Biometry Among the Three Major Racial Groups in Singapore

	Age-Sex Adjusted			Multivariate Adjusted*
	OR, β	95% CI		OR, β	95% CI
Myopia, <−0.5 D
Chinese	2.04	1.82, 2.28	<0.001	2.04	1.79, 2.32	<0.001
Indians	1.04	0.93, 1.18	0.48	0.97	0.85, 1.11	0.65
Malays		Reference			Reference
High myopia, <−5.0 D
Chinese	2.27	1.80, 2.27	<0.001	1.84	1.42, 2.39	<0.001
Indians	1.05	0.80, 1.37	0.22	0.84	0.63, 1.11	0.22
Malays		Reference			Reference
Hyperopia, >+0.5 D
Chinese	0.90	0.81, 1.01	0.07	0.84	0.74, 0.95	0.005
Indians	1.53	1.36, 1.71	<0.001	1.54	1.37, 1.74	<0.001
Malays		Reference			Reference
Astigmatism, <−0.5 cylinders
Chinese	3.28	2.93, 3.67	<0.001	3.6	3.18, 4.07	<0.001
Indians	2.62	2.34, 2.93	<0.001	2.79	2.48, 3.15	<0.001
Malays		Reference			Reference
AL, mm
Chinese	0.43	0.38, 0.48	<0.001	0.20	0.14, 0.25	<0.001
Indians	−0.14	−0.20, −0.09	<0.001	−0.33	−0.39, −0.28	<0.001
Malays		Reference			Reference
ACD, mm
Chinese	0.21	0.16, 0.26	<0.001	0.09	0.04, 0.15	0.001
Indians	0.12	0.08, 0.17	<0.001	−0.01	−0.06, 0.05	0.75
Malays		Reference			Reference
CR, mm
Chinese	0.01	−0.04, 0.06	0.77	−0.03	−0.09, 0.02	0.25
Indians	−0.04	−0.09, 0.02	0.19	−0.07	−0.12, −0.01	0.02
Malays		Reference			Reference

 OR, odds ratio; β, regression coefficient.

*  Adjusted for age, sex, educational level, height, and nuclear cataract.

Table 5 shows the associates of myopia, high myopia, and AL in SEED. After adjusting for age, sex and race, the odds ratio of any myopia was 4.86 (95% CI 3.88, 6.10) for those with a university education compared with no formal education; 3.11 (95% CI 2.64, 3.66) for those with nuclear cataract compared with non-nuclear cataract and 1.26 (95% CI 1.12, 1.41) for those born in Singapore compared with those born outside of Singapore. The associates for high myopia were similar with myopia. After adjusting for age, sex and race, each centimeter of height increase was associated with a 0.03-mm increase in AL. Adults with a university educational level had 0.96 mm longer in mean AL than those with no formal education. Singapore born adults had 0.12 mm longer in mean AL than those born outside Singapore. 

Table 5. 

 

Associates of Myopia, High Myopia, and AL in the SEED Study

	Myopia, <−0.5 D			High Myopia, <−6 D			AL, mm
	OR	95% CI	P Value	OR	95% CI	P Value	β	95% CI	P Value
Age, per year increase	0.98	0.97, 0.99	<0.001	0.97	0.96, 0.99	<0.001	−0.001	−0.004, 0.001	0.33
Female vs. male	1.29	1.17, 1.43	<0.001	1.53	1.26, 1.87	<0.001	−0.02	−0.08.0.05	0.60
Race
Malay		Reference			Reference			Reference
Indian	1.02	0.89, 1.16	0.80	0.88	0.66, 1.17	0.38	−0.32	−0.37, −0.26	<0.001
Chinese	2.07	1.83, 2.34	<0.001	1.78	1.38, 2.30	<0.001	0.19	0.14, 0.25	<0.001
Educational level
No formal education		Reference			Reference			Reference
Primary education	1.07	0.92, 1.25	0.32	1.15	0.79, 1.67	0.48	0.10	0.03, 0.16	0.003
Secondary education	2.12	1.80, 2.50	<0.001	2.22	1.52, 3.24	<0.001	0.38	0.31, 0.45	<0.001
Polytechnic	3.07	2.52, 3.73	<0.001	4.45	2.98, 6.66	<0.001	0.69	0.61, 0.78	<0.001
University	4.86	3.88, 6.10	<0.001	8.40	5.50, 12.8	<0.001	0.96	0.86, 1.07	<0.001
Born in Singapore vs. born outside Singapore	1.26	1.12, 1.41	<0.001	1.56	1.22, 2.00	<0.001	0.12	0.07, 0.17	<0.001
Nuclear cataract vs. non-nuclear cataract	3.11	2.64, 3.66	<0.001	2.37	1.68, 3.36	<0.001		–
Height, per cm increase							0.029	0.025, 0.033	<0.001

Discussion

The prevalence of myopia in our middle-aged to elderly cohort was much lower compared with the younger “myopia” generation in Singapore with a myopia prevalence of 81% (Saw et al. IOVS 2011;52:ARVO E-Abstract 2490). As expected, Chinese had the longest AL and were most likely to be affected by myopia, high myopia, and astigmatism while Indians were most likely to be hyperopic. Furthermore, we observed a significant increase in the prevalence of astigmatism and mean AL in Singapore Chinese adults aged over 40 years within the past 12 years. 

Our study showed the prevalence of myopia in middle-aged to elderly adults aged over 40 years is much lower compared with adolescents aged 16 to 26 years (Saw et al. IOVS 2011;52:ARVO E-Abstract 2490). The latest data showed that the prevalence of myopia in Singapore male military conscripts aged 16 to 26 years (n = 29,170) was 81.3% (95% CI 81.2, 82.1) in 2009 (Saw et al. IOVS 2011;52:ARVO E-Abstract 2490). This pattern of decreasing myopia prevalence with increasing age may be partially explained by an intrinsic age-related decline in myopia prevalence such as an increase in refractive power of lens, ²⁹ changes in lens position, ³⁰ or increased AL. ³¹ Meanwhile, this trend may also be explained by the cohort effect in Singapore (Fig. 3). The higher prevalence of myopia in younger generations as compared with older ones may be driven by higher levels of reading exposure with a large amount of near-work activity—correspondingly lower levels of outdoor physical activity, and other unmeasured factors among different birth cohorts. Since our study is cross-sectional, we cannot determine whether the cohort effect or the age-related effect plays a more important role. There has been long concern that visual impairment and blindness resulting from myopia would lead to major health concern for Asians. These data may have implications for many Asian countries where dramatic changes in social environmental variables have been taking place during urbanization. 

Figure 3

 

Mean SE in different birth cohorts in Singapore. Data taken from: 0.5 to 6 years: The Strabismus, Amblyopia and Refractive Error Study (Chinese only) ¹⁶ ; 7 to 9 years: The Singapore Cohort Study of the Risk Factors for Myopia (SCORM) ¹⁷ ; 10 to 14 years: (SCORM follow-up, unpublished); 15 to 26 years: Singapore male military conscripts survey (Saw et al. IOVS 2011;52:ARVO E-Abstract 2490); 27 to 39 years: The Singapore Prospective Program (SP2), unpublished; more than 40 years: SEED.

The prevalence of myopia in Singapore is comparable with other urbanized Asian communities such as Hong Kong ³² and Tajimi city in Japan, ³³ with approximately 40% of the general population aged over 40 years being affected by myopia (Fig. 4). However, lower prevalence of myopia was also reported in some Asian countries such as India ³⁴ and mainland China, ²¹ with approximately only 20% of the adults aged over 40 years being affected by myopia. In contrast, recent data revealed that the prevalence of myopia in Caucasians in the United States has been increasing significantly and reached the same level reported in urbanized Asian societies. The 1999–2004 National Health and Nutrition Examination Survey (NHANES) reported that 33% of the whites aged over 40 years in the United States have been affected by myopia even using a more stringent criterion of −1 D to define myopia, ^7,35 which was not lower than the figures reported in most Asian populations. Using the definition of less than −1.0 D, our study reported a similar prevalence of myopia (31.4%) compared with the whites in NHANES. Thus, we believe that there is little evidence supporting an intrinsically higher prevalence or myopia, or a greater susceptibility to environmental risk factors in populations of Asian origin. Myopia is most common in large urban Asian cities but lower in other parts of Asia. 

Figure 4

 

Prevalence of myopia in adults in East Asian urban cities. Data taken from: Hong Kong: n = 335, aged over 40 years ³² ; Beijing: n = 4319, aged over 40 years ²¹ ; Taiwan: n = 1361, aged over 65 years ²² ; Tajimi (Japan): n = 3021, aged over 40 years. ³³

When comparing the data of the Chinese cohort (SCES) in our study with an earlier Chinese cohort of the same age range (Tanjong Pagar Survey, ^20,23 1996–1997), we observed an increase in the prevalence of myopia in younger age groups within the past 12 years (Table 3). It has been reported that the rapid increase in the prevalence of myopia has been well documented throughout the world. ^7,36–38 However, nuclear cataract has been found to result in a myopic shift, reflecting the increased power of the more sclerotic lens rather than increased AL. We found that the prevalence of nuclear cataract was much higher in the Tanjong Pagar Survey than in SCES (34.2% vs. 9.2%, P < 0.001), indicating that most myopic adults have lens-induced myopia rather than axial myopia in the Tanjong Pagar Survey. Therefore, the change in AL would be more informative than refraction to the understanding of the increasing myopia rates in the past decades. Our study documented a significant increase in AL in all age groups even after adjusting for the effect of height (Fig. 2), supporting an increasing trend in the prevalence of axial myopia in Singapore Chinese aged over 40 years within the past 12 years. In addition, the change in AL might have been related to the prevalence of other ocular complications including primary open glaucoma, ³⁹ retinal detachment, ⁴⁰ and myopic retinopathy. ⁴¹  

It is also interesting to find that the prevalence of astigmatism has increased significantly during the past 12 years (Table 3). The explanation for the increase trends is unclear. Compared with the TPS, our study showed that the major increase in astigmatism rate is due to the increase in myopic astigmatism rate, which accounted for 72% of the observed increasing rate in astigmatism in Singapore Chinese. Therefore, our postulation is that there is an increase in myopic astigmatism due to the rise in environmental factors such as a more competitive education environment. 

It was not unexpected to find that Chinese had the highest prevalence of myopia among the three major racial groups living in Singapore. The result of this study is consistent with previous studies in children or teenagers. In the latest survey in 2009 on Singapore male conscripts, the prevalence of myopia in Chinese, Indian, and Malay was 85.9%, 74.5%, and 70.7%, respectively (P = 0.03, aged 16–26 years; Saw et al. IOVS 2011;52:ARVO E-Abstract 2490). In the Gombak district of Malaysia, Chinese children had the highest prevalence of myopia (46.4%) among the ethnic groups, followed by Indians (16.2%) and Malays (15.4%) across all ages (n = 4634, aged 7–15 years). ⁴² However, our study provided more accurate comparisons by using the same study protocols to collect data and filled the gap of knowledge in the middle-aged to elderly population. In addition, our study provided further insights into the racial differences in refractive errors by assessing the racial variations in ocular biometric parameters. 

Some implications from our study should be noted. First, it is of great public health significance to quantify and compare the burden of myopia and other refractive errors across the three major ethnic groups in Asia in a single setting using the same methodology. The results of our study may be related to many countries where a key objective of public health system reform is to redress racial inequality. Second, our studies now demonstrate that astigmatism is extremely prevalent in Singapore and has increased rapidly within the past decade. Since vision clarity is so critical for many aspects of daily living related to public safety, including vehicular driving and electronic device usage, health policymakers should be made aware of this trend and take actions to prevent not only myopia but also astigmatism. Last but not least, the study samples were drawn from population-based surveys of Singapore Chinese, Malays, and Indians. The findings from Chinese, India, and Malays in Singapore may be extrapolated to Chinese in China, Indians in India, and Malays in Indonesia, which have the greatest population of each cohort throughout the world. 

The study's strengths included a large sample size, a reasonable response rate, and a standardized refraction and ocular biometry assessment. There were also some limitations for this study. First, we were unable to identify the exact explanatory factors for the observed racial differences since some important myopia-related exposures such as time spent outdoors in childhood, parental myopia, and parental educational level were not captured by our study. In addition, the cross-section design is limited to establish a temporal relationship and thus cannot infer any casual relationships between these risk factors and any of the refractive errors. Finally, nonresponders were older than responders, leading to an overestimation in myopia prevalence and an underestimation in hyperopia or astigmatism prevalence due to the age distributions of refractive errors. 

In conclusion, this population-based study of adults aged over 40 years in Singapore showed a significant lower prevalence of myopia compared with younger generations reported previously. The prevalence of astigmatism and mean AL in Singapore Chinese adults aged over 40 years have been increasing within the past 12 years. Chinese adults had a higher prevalence of myopia, high myopia, astigmatism as well as longer AL as compared with Indians or Malays living in the same environment. Further studies are warranted to identify the exact factors explaining the observed racial differences in refractive errors. 

Acknowledgments

Supported by the Biomedical Research Council (BMRC), 08/1/35/19/550 & National Medical Research Council (NMRC), STaR/0003/2008, Singapore. The authors alone are responsible for the content and writing of the paper. 

Disclosure: C.-W. Pan, None; Y.-F. Zheng, None; A.R. Anuar, None; M. Chew, None; G. Gazzard, None; T. Aung, None; C.-Y. Cheng, None; T.Y. Wong, None; S.-M. Saw, None