Nearwork in Early-Onset Myopia

Abstract

purpose. To determine the relationship of nearwork and myopia in young elementary school-age children in Singapore.

methods. A cross-sectional study of 1005 school children aged 7 to 9 years was conducted in two schools in Singapore. Cycloplegic autorefraction, keratometry, and biometry measurements were performed. In addition, the parents completed a detailed questionnaire on nearwork activity (books read per week, reading in hours per day and diopter hours [addition of three times reading, two times computer use, and two times video games use in hours per day]). Other risk factors, such as parental myopia, socioeconomic status, and light exposure history, were assessed.

results. In addition to socioeconomic factors, several nearwork indices were associated with myopia in these young children. The multivariate adjusted odds ratio of higher myopia (at least −3.0 D) for children who read more than two books per week was 3.05 (95% confidence interval [CI], 1.80–5.18). However, the odds ratios of higher myopia for children who read more than 2 hours per day or with more than 8 diopter hours (1.50; 95% CI, 0.87–2.55 and 1.04; 95% CI, 0.61–1.78, respectively) were not significant, after controlling for several factors.

conclusions. Children aged 7 to 9 years with a greater current reading exposure were more likely to be myopic. This association of reading and myopia in a young age cohort was greater than the strength of the reading association generally found in older myopic subjects. Whether these results identify an association of early-onset myopia with nearwork activity or other potentially confounding factors is discussed.

There has been a dramatic increase in myopia prevalence rates over the past few decades in Singapore and other parts of Asia. ^{1 2 3} The increase in rates has been remarkable in very young Asian children, too, suggesting that early lifestyle risk factors may have a large impact on early myopia development and the overall population prevalence rate of myopia. Because the gene pool has not changed significantly over the decades, the rapid increase in myopia prevalence rates has been attributed to increases in reading activity and other environmental factors. ³ It has been noted that there is an increased prevalence of myopia observed in certain occupations such as microscopists. ^{4 5} Evidence from epidemiologic studies have shown mixed results. Cross-sectional studies in Hong Kong and Newfoundland found rather weak associations of reading or school attendance with myopia, whereas boys in orthodox Jewish schools were found to have a higher rate of myopia (81.3%) compared with boys in general Jewish schools (27.4%). ^{6 7 8} However, no relationship between nearwork and myopia was found for 925 offspring tested in 723 families in Hawaii. ^{9 10}  

The few studies in Asia that have simultaneously explored reading and other environmental risk factors for myopia had relatively small sample sizes, and little is known about the role of reading in the development of myopia in young children. ^{11 12} Therefore, we examined the correlation of potential risk factors such as reading and parental myopia with myopia in 1005 young Singapore children, aged 7 to 9 years. Although a description of the interaction of reading with parental myopia to predict myopia in this population has been presented, ¹³ the present article reports the detailed evaluation of the relation of reading with myopia and ocular biometry measures in these young subjects. 

Methods

Study Population

We report the initial cross-sectional results of a long-term cohort study in Singapore. This report consists of the entry data on children aged 7 to 9 years studied in 1999. The study was approved by the Ethics Committee, Singapore Eye Research Institute and followed the tenets of the Declaration of Helsinki. Permission to conduct the study was obtained from the Ministry of Education of Singapore. The study was supported by the principals and teachers of the two schools. To sample children from schools with different overall academic performance, two elementary schools were selected based on prior National Examination results of their students. One school in the Eastern part of Singapore ranked among the top 20 schools in the country, and the other school in the Northern part of Singapore ranked among the bottom 20 schools. All children in grades 1 and 2, aged 7 to 8 years, in the Eastern school (n = 639), and grades 1 to 3, aged 7 to 9 years, in the Northern school (n = 988) were invited to join the study. Written informed consent was obtained from the parents after the nature of the study was explained. Children who had a serious medical condition, such as leukemia or heart disorders, or a syndrome-associated with myopia or any serious eye disorder, such as congenital cataract, were excluded. Similarly, children allergic to eye drops or who refused the instillation of cycloplegic eye drops were excluded. Three hundred and nine children (48.4%) from the Eastern school and 696 children (70.4%) from the Northern school elected to participate. The proportion of children who reported nearsightedness before the school eye examination was not different in those who elected to participate (27.3%) and those who did not (26.8%). In total, there were 522 children aged 7 years, 321 children aged 8 years, and 162 children aged 9 years. In our sample, there were 729 Chinese (72.5%) and 276 non-Chinese children (19.4% Malays, 5.6% Indians, and 2.5% children of other races). 

Refractive Error Measurements

The children’s eyes were examined on the school premises during the first 2 weeks of November 1999. Corrected and uncorrected distance visual acuity was measured using log minimum angle of resolution (logMAR) charts, according to a standard protocol. ¹⁴ After instillation of 0.5% proparacaine, cycloplegia was induced in each eye with 3 drops 1% cyclopentolate instilled 5 minutes apart. At least 30 minutes after the last cycloplegic drop, one of two autokeratorefractometers (model RK 5; Canon, Inc. Ltd., Tochigiken, Japan) was used to obtain five consecutive refraction measurements and corneal curvature readings. 

Ultrasound biometry measurements of axial eye length, anterior chamber depth, crystalline lens thickness, and vitreous chamber depth were performed on 979 children (26 children refused examination) using one of two calibrated biometry machines (Echoscan model US-800; Nidek Co., Ltd., Tokyo, Japan; probe frequency of 10 mHz), after 1 drop of 0.5% proparacaine. The average of six axial length measurements was taken, whereby the SD of these six readings was less than 0.12 mm. All refractive error and biometry measurements were conducted without prior knowledge of the child’s questionnaire results. 

Questionnaire on Lifestyle Factors

The parents completed an eight-page questionnaire that was distributed through the school class teacher 2 weeks before the eye examination in the schools. For parents who were not conversant in English, a Chinese or Malay version of the questionnaire was provided. The intraclass correlation coefficient of the reliability of nearwork assessment was 0.87 (95% CI, 0.85–0.91) and the intraclass correlation coefficient for nearwork compared with four 24-hour diaries was 0.50 (95% CI, 0.34–0.66). ¹⁵ The Chinese and Malay versions of the questionnaire were pilot tested in 20 children. Parents were asked to quantify nearwork activity (reading, writing, computer use, playing video games) in hours per day per activity on weekdays and weekends outside school hours and to indicate the number of books read per week. We computed a weighted variable, diopter hours, by adding three times reading, two times computer use, and two times video games use in hours per day for nearwork activity outside school. ¹⁶ During school hours, time spent reading and writing for all children in the same grade in Singapore is rather similar (median = 2.0 hours per day), because there is a common school syllabus. Many children in Singapore receive supplemental instruction outside school, termed “tuition” classes; children spend the majority of the time reading or writing during these classes. We also asked whether the child received tuition classes. Other questions determined basic sociodemographic factors, including father’s and mother’s completed level of education and total family income, outdoor activity, and ambient lighting while sleeping at night (darkness, light from the adjacent room or window, night light or dim light, or room light) before age 2 years. We determined whether the child’s parents were myopic by asking the parents whether they were wearing spectacles or contact lenses for nearsightedness. The age of onset of myopia was assessed by asking when the child first wore spectacles for nearsightedness. 

Data Analysis

The measurements of refraction were analyzed as spherical equivalent (SE; sphere +0.5 negative cylinder). Myopia was defined as a negative refractive error of at least −0.5 D. The distribution of refractive error (SE) was slightly skewed, but axial length was normally distributed. There was a high correlation between right and left eye refractive error data (Pearson correlation coefficient = 0.94). Results in right and left eyes, analyzed separately, were found to be similar, thus only results of the right eye are presented. Subjects were divided into three refractive error groups, based on their SE refractions: higher myopes (SE ≤ −3.0 D), lower myopes (−3.0 < SE ≤ −0.5 D), and nonmyopes (SE > −0.5 D). Higher myopia was defined as myopia greater than −3.0 D, an arbitrary separation to distinguish degrees of myopia for analysis purposes in this study. When comparisons were made across refractive error groups, pair-wise post hoc comparisons were performed, using the Bonferroni test. Multiple logistic regression models were used to examine the relationships between myopia and lifestyle variables, adjusting for other possible confounders. The statistically significant interaction term (books per week × parental myopia) was included in all multivariate models. In addition to the analysis of the entire study population, Chinese and non-Chinese were analyzed separately, because Chinese were the majority ethnic group. A separate analysis of the data from each school was also done. Data analysis was conducted using STATA. ¹⁷  

Results

Characteristics of the Study Population

For the sample as a whole (N = 1005), the mean refractive error was −0.33 D (range, −9.1–4.8). The distribution of refractive errors in the sample is shown in Figure 1 . The proportion of children with higher myopia was 8.1% and of those with lower myopia, 24.3%. The prevalence rates of overall myopia were higher in Chinese (37.0%) than in non-Chinese (19.9%) children. For higher myopia in particular, the prevalence rates were similarly higher in Chinese children (10.0%) than in non-Chinese (2.9%) in the total sample and, separately, in the two schools. Chinese children had higher total family incomes (P < 0.001) and were more likely to attend tuition classes (P < 0.001). 

Ocular Components

Eyes in children with higher myopia were more likely to have higher cylinder power, longer axial lengths, deeper anterior chambers, longer vitreous chambers, steeper corneas, and a higher ratio of axial length (AL) to corneal radius (CR) than were eyes in children with lower myopia or no myopia (Table 1) . The relationships of refractive error and ocular components were similar when children from the two schools or Chinese and non-Chinese children were analyzed separately. 

Risk Factors

Consistent with the development of myopia, the prevalence rates increased with age (27.6% at age 7 years, 34.6% at age 8, and 43.2% at age 9). There were positive associations between higher myopia prevalence rates and larger housing type, higher family income, more advanced father’s and mother’s education (P < 0.001, for each). Housing type, family income, and parental education are all likely surrogates for socioeconomic status. Parental history of myopia was positively related to the refractive error group (P < 0.001; Table 2 ). For Chinese children, the proportion of children with myopia in one or two myopic parents was higher than in those with no myopic parents; whereas for non-Chinese children, the proportion with myopia in two myopic parents was higher than in those with one myopic parent or none. Night lighting before age two was not associated with the level of myopia (P = 0.74). 

Some associations of nearwork activities and myopia grouping emerged, but chiefly in children with higher myopia (Table 3) . The strongest association related to the number of books read per week. The overall population read a median of two (range, 0–20) books per week, with no differences between Chinese and non-Chinese subjects (P = 0.94). Children with higher myopia read more books per week (median = 3) compared with those with lower myopia or nonmyopes (median = 2; P < 0.001). Post hoc pair-wise comparisons revealed significant differences in the number of books read per week and reading in hours per day in children with higher myopia versus those with lower myopia and higher myopia versus no myopia. The crude odds ratio of higher myopia for reading more than two books per week was 3.15 (95% CI, 1.96–5.04), whereas the odds ratio adjusted for age, gender, race, night light, parental myopia, and school, was 3.05 (95% CI, 1.80–5.18). Thus, books read per week was an independent risk factor for higher myopia. The odds ratio of higher myopia in children who read more than 2 hours per day (median = 2) was 2.16 (95% CI, 1.34–3.47), and the multivariate adjusted odds ratio was 1.50 (95% CI, 0.87–2.55). The crude and multivariate adjusted odds ratio of higher myopia in children with higher diopter hours (>8; median = 8.7) were 1.65 (95% CI, 1.02–2.68), and 1.04 (95% CI, 0.61–1.78), respectively. Children with tuition classes were twice as likely to have higher myopia. Similarly, children who used the computer on a regular basis had a two times higher rate of higher myopia. In Chinese children alone, books read per week was significantly associated with myopia and a positive but nonsignificant association was found in non-Chinese children. Significantly positive associations between books read per week, computer use, music classes, and higher myopia were found in Chinese but not in non-Chinese children. Books read per week, tuition classes, and computer use positively associated with higher myopia, when each school was analyzed separately. 

Children with higher myopia had an earlier mean age of myopia onset (6.4 years; range, 3–9), compared with 7.2 years (range, 4–9) in children with lower myopia (P < 0.001). The mean age of myopia onset in children who read more than two books per week also was younger at 6.7 years (range, 3–9) compared with 7.1 years (range, 4–9) in children who read fewer books (P = 0.007). Children who used the computer regularly (P = 0.05) and children who had at least one myopic parent (P = 0.01) had an earlier age of onset of myopia. However, reading, diopter hours and tuition lessons were not significantly related to age of onset of myopia. Similar relationships were found in the two different schools and in Chinese or non-Chinese children, when analyzed separately. 

Table 4 shows the relationship between the different risk factors and axial length (biometry measures were available for 979 of the 1005 subjects). Children who read more than two books per week (P < 0.001) or who had parents with myopia (P < 0.001) had eyes with longer axial lengths. However, there were no significant relationships between reading in hours per day, diopter hours, tuition classes, or computer use and axial length. Separate assessments by school and race (Chinese and non-Chinese) revealed similar patterns. For every book read per week, there was a corresponding 0.04-mm increase in axial length, when controlling for age, gender, race, parental history of myopia, night light, and school. 

Reading, Parental Myopia, and Myopia

Children who read more than two books per week and had two parents with myopia had an adjusted mean refractive error of −1.33 D and the longest eyeballs (23.78 mm), whereas children who read two or fewer books per week and had no myopic parents had an adjusted mean refractive error of −0.19 D and the shortest eyeballs (23.20 mm; Fig. 2 ). There was a statistically significant interaction effect of parental history of myopia and books read per week on axial length (P < 0.001). Similar results were found when each school or when Chinese and non-Chinese children were analyzed separately. 

Type of School

There were 167 boys (54.0%) and 142 girls (46.0%) from the Eastern school, and 332 boys (47.7%) and 364 girls (52.3%) in the Northern school. The Eastern school included 305 Chinese (98.7%), 3 Indians (1.0%), and 1 child of another race (0.3%); whereas the Northern school included 424 Chinese (60.9%), 196 Malaysians (28.2%), 52 Indians (7.5%), and 24 children of other races (3.5%). The proportion of children with higher myopia was 17.2% in the Eastern school, compared with 4.0% in the Northern school (P < 0.001). These refractive differences between the two schools remained, even after controlling for the associations in age, gender, race, books read per week, and parental myopia (described later). A higher proportion of children from the Eastern school read more than two books per week (48.5% vs. 32.3%) and had two myopic parents (45.2% vs. 11.1%), compared with those from the Northern school. 

Discussion

This large epidemiologic study of the association of myopia and reading in young Asian children provides particularly comprehensive estimates of nearwork activity, including books read per week, reading in hours per day, and diopter hours. An unusual feature of our study, particularly in studies of 7- to 9-year-old children, is inclusion not only of cycloplegic refractions but also of the essential ocular components of refraction through ocular biometry and keratometry. 

Ocular Components of Myopic Children

The eyes of even these very young children had morphologic characteristics found in older subjects. The more highly myopic eyes had longer axial lengths, larger vitreous chambers, deeper anterior chambers, and thinner lenses. ^{18 19 20 21 22} Whether the lens becomes mechanically flattened in the expanded myopic eye or somehow participates in an emmetropization process is indeterminate from our data. ²³ The school entrant myopes in our study had higher cylinder power and higher rates of astigmatism, suggesting that often myopia and astigmatism coexist and similar risk factors may affect their development. Both astigmatism and myopia may result from failure of proper emmetropization, or perhaps astigmatism-induced blur may disrupt the normal emmetropization process and contribute to myopia. ²⁴ Another possibility is that both astigmatism and myopia are a result of uncoordinated eye growth due to a failure in the emmetropization mechanism. Children with myopia also have higher AL/CR ratios, demonstrating that an inappropriately steep cornea in myopia occurs early in the development of the condition. ^{25 26}  

Nearwork and Early-Onset Myopia

Despite much study, it has been difficult to associate quantitative measures of nearwork activity with myopia in epidemiologic studies. ^{3 12} Compared with most prior reports, stronger correlates between nearwork and myopia emerged. The chief finding: Children who read more books per week spent more time reading in hours per day, reported higher diopter hours, or had extra tuition classes (involves additional reading and writing), were more likely to have higher myopia. However, after controlling for potential confounders, only books read per week was independently related to higher myopia. 

As a nearwork marker, reading of books per week may be independently linked to parental attitude and education (families who value reading encourage the children to read), family income (ability to purchase books), or access to facilities and transport means (the proximity of a library or bookstore and the affordability of various modes of transportation). Chinese children in Singapore are generally from families with higher income and therefore with better access to books and libraries, as well as from families with a higher educational level with generally higher parental expectations for their child’s school performance. Because the risk factors may be interrelated and statistical adjustment may not explain or completely remove the influence of one environmental risk factor on another, we examined whether the results were explained simply by differences in ethnicity. The positive association between books read per week and higher myopia remained in Chinese and non-Chinese children (although the association was slightly weaker in non-Chinese) after stratification by ethnic group, suggesting that reading of books may be associated independently with myopia. In fact, previously hypothesized risk factors for myopia, such as indicators of socioeconomic status and affluence, may be surrogate markers of reading activity. 

Reading Accomplishment: a Novel Approach to Quantifying Nearwork?

The typical epidemiologic assessments of nearwork attempt to measure time spent in nearwork activities. As in many studies in older subjects, the time-based nearwork measures evaluated herein largely failed to account for myopia’s developing. Little if any information is available in the myopia epidemiology literature about how survey-derived time-based nearwork measures actually correlate with ocular use. Complex issues such as subject attentiveness or patterns of temporal interruptions may be physiologically important but are essentially unstudied. Number of books read differed from conventional nearwork indices in that it assessed nearwork accomplishment instead of assessing nearwork time. It also differs from the diopter hours, a time-weighted score devised to weigh the amount of accommodation required for different nearwork tasks. Besides the need to assess time, the concept of the score, diopter hours, also may be limited by intersubject differences in task-specific accommodative needs. Myopes with uncorrected vision, but also those with corrected vision, accommodate less during reading than nonmyopes, and diopter hours may thus underestimate accommodative activity. Furthermore, diopter hours is an inaccurate measure of accommodation in noncorrected hyperopia. As a proximity index, diopter hours may be a more useful parameter, if it provides associations of potential mechanistic value. Because increasing myopia severity associated with more books read per week, a nearwork measure that quantifies nearwork accomplishment rather than nearwork time, may prove a useful physiologic parameter to assess myopia risk. However, we acknowledge that the measure “books per week” may be confounded by other varying factors such as the size of font, the type of font, or type of characters (Chinese versus non-Chinese). Because books read per week provided the strongest association in the present study, assessing nearwork accomplishment by this or other possible novel approaches would seem a useful methodologic strategy in trying to unravel myopia risk factors. 

Is Nearwork a Risk Factor for Myopia?

Because the oldest children in this study were 9 years of age, all myopia recorded was early in onset. However, young children with lower myopia often progress to higher myopia at a later age. Because many of the putative risk factors considered may be interrelated and because reading and myopia were both measured at one time point in this cross-sectional study, we cannot conclude that there is a cause–effect relationship. An important factor to consider is the number of books read per week before the age of onset of myopia. One assumption in our study is that current books read per week reflect the reading habits of the child before the onset of myopia. This is likely, because the time interval between the age of onset of myopia and age of entry into the study is somewhat short (median = 1; range, 0–5 years). However, the children with higher myopia and the most reading also had the earliest onset of myopia, raising the question of whether books read constitutes a surrogate marker for some aspect of neurocognitive development or intelligence in these young children. 

Despite these qualifications, the nearwork–myopia relationship appears stronger in the present study of young children with early-onset myopia than in surveys of older subjects. The biological mechanisms for early-onset compared with late-onset myopia may differ, confounding studies in older subjects if the influence of nearwork in precipitating myopia is age-related and most closely linked in early childhood. 

Nearwork–Myopia Relationship in Chinese and Non-Chinese Children

In the Asian cities of Hong Kong, Taiwan, and Singapore, with predominantly Chinese populations, the increasingly high myopia prevalence rates have been attributed to the extremely competitive schooling systems implemented in the past few decades. Although some racial differences in myopia occur in Singapore, data from military conscripts reveal high myopia prevalence rates in all major racial groups in the country: Chinese (82.2%), Malays (65.0%), and Indians (68.7%). ²⁷ We found that the nearwork–myopia association was stronger in Chinese than in non-Chinese children. This suggests that nearwork and other environmental factors may differentially affect refractive development in Chinese compared with non-Chinese children. 

Interaction of Reading and Parental Myopia as Determinants of Myopia

The link with parental myopia suggests either a familial predisposition to myopia from shared genes, shared environment, or both. ²⁸ Children with two myopic parents and who read more than two books per week had eyes with axial lengths that were 0.7 mm longer (high myopia, 23.7%) than children with no parental history of myopia and who read two or fewer books per week (high myopia, 2.5%). ¹³ This interaction of parental myopia and books read with axial length parallels the similar interaction in determining refraction in the same population. Data on parental myopia was limited by rather indirect estimates from a questionnaire rather than refractive error measurements, and this may have lead to misclassification bias. 

Other Correlates with Early Childhood Myopia

In multivariate modeling, the strongest risk factor for higher myopia was books read per week, followed by age, parental myopia, and school. These risk factors independently predict higher myopia, after several other confounders were controlled for. Of note, among non-Chinese children the likelihood of myopia was higher in children with two myopic parents than those with one myopic parent or none, but the likelihood of myopia was similar between those with one myopic parent or none. These findings are consistent with studies in the United States in which the odds ratio was 6.42 for children with two myopic parents compared with those with no or one myopic parent. ²⁹ In the present study, the relationship between other risk factors and myopia is again similar to other studies: Myopic children are more likely to have myopic parents and are from families with higher socioeconomic status. ^{16 30 31} It is postulated that socioeconomic status may be a surrogate for environmental lifestyle factors such as academic achievement, intellectual ability, or nearwork. ^{32 33} The type of school was also associated with myopia, even after controlling for nearwork, but school may be a surrogate for other environmental factors related to myopia: intelligence, personality, or an unidentified environmental risk factor. Another factor to consider is the dissimilar ethnic compositions of the two schools, with a higher proportion of Chinese in the Eastern school. As lifestyle and cultural habits are closely linked to ethnicity in Singapore, perhaps the difference in myopia prevalence rates in the two schools could be explained by a larger proportion of Chinese in the Eastern school. More important, we still observed the nearwork–myopia relationship among Chinese and non-Chinese children (weaker association present) alone, even when each school was analyzed separately. 

Conclusions

In summary, the number of books read per week, a novel index of nearwork activity, was associated with higher myopia and earlier onset myopia in young Asian children, independent of several other related factors. Other quantitative measures of nearwork (reading in hours per day) were related with higher myopia (at least −3.0 D), but the associations did not remain after multivariate adjustment. In the context of available quantitative nearwork studies in the myopia literature, the present results provide somewhat stronger nearwork correlates with myopia, but do not unambiguously resolve whether nearwork is a risk factor for the development of myopia or a surrogate for other environmental or genetic factors. 

 Supported by Grant SERI/MG/97-04/0005 from the National Medical Research Council, Singapore.

 Submitted for publication May 29, 2001; revised September 19, 2001; accepted October 8, 2001.

 Commercial relationships policy: N.

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked“ advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

Distribution of refractive error in the sample.

Table 1.

 

Adjusted Means of Cylinder, Biometry, and Corneal Curvature Measurements in the Right Eye of Study Subjects

	Higher Myopes (n = 81; SE ≤ −3.0 Diopters)	Lower Myopes (n = 244; −3.0 < SE ≤ −0.5 D)	Nonmyopes (n = 680; SE > −0.5 D)	P *
Cylinder (D)	1.1 (1.0, 1.3)	0.9 (0.8, 1.0)	0.6 (0.5, 0.6)	<0.001
Axial length (mm)	24.81 (24.67, 24.96)	23.76 (23.68, 23.85)	22.99 (22.94, 23.05)	<0.001
Anterior chamber depth (mm)	3.73 (3.67, 3.79)	3.72 (3.68, 3.75)	3.61 (3.59, 3.63)	<0.001
Lens thickness (mm)	3.42 (3.38, 3.46)	3.44 (3.42, 3.46)	3.47 (3.46, 3.48)	0.015
Vitreous chamber depth (mm)	17.65 (17.51, 17.80)	16.60 (16.52, 16.68)	15.92 (15.87, 15.97)	<0.001
Average corneal radius of curvature (mm)	7.67 (7.62, 7.72)	7.71 (7.68, 7.74)	7.76 (7.75, 7.78)	<0.001
AL/CR ratio	3.23 (3.22, 3.25)	3.08 (3.07, 3.09)	2.96 (2.96, 2.97)	<0.001

 Data are adjusted mean with 95% CI in parentheses. N = 1005.

*  By two-tailed analysis of covariance, adjusted for age, gender, and race for total population.

Table 2.

 

Proportion of Children with Higher Myopia or Lower Myopia and without Myopia, by Risk Factors

	Chinese					Non-Chinese					Total					Total (n)
		Higher Myopes (n = 73)	Lower Myopes (n = 197)	Nonmyopes (n = 459)	Higher Myopes (n = 8)	Lower Myopes (n = 47)	Nonmyopes (n = 221)	Higher Myopes (n = 81)	Lower Myopes (n = 244)	Nonmyopes (n = 680)
		%
Total family income per month * †
Sin$2,000	6.2	24.4	69.4	3.2	13.2	83.6	4.7	18.9	76.4	382
Sin$2,001–5,000	10.6	27.2	62.2	1.4	25.7	72.9	8.5	26.9	64.6	316
>Sin$5,000	12.5	28.3	59.2	0.0	28.6	71.4	12.1	28.3	59.6	272
P		0.14			0.14			<0.001
Parental myopia * †
None	4.3	18.2	77.5	2.4	14.5	83.1	3.5	16.5	80.0	372
One	8.7	29.0	62.3	3.0	16.4	80.6	7.6	26.7	65.7	360
Two	17.8	33.5	48.7	0.0	36.8	63.2	16.2	33.8	50.0	207
P		<0.001			0.16			<0.001
Lighting while sleeping before 2 years of age, †
In the dark	7.8	26.9	65.4	5.0	13.8	81.3	7.2	24.2	68.6	389
Light from adjacent room or window	12.8	32.0	55.2	1.5	18.5	80.0	9.0	27.4	63.7	190
Night light	13.8	26.7	59.5	1.8	17.4	80.7	9.7	23.5	66.8	319
Room light	6.8	22.0	71.2	3.6	28.6	67.9	6.2	24.1	70.1	87
P		0.13			0.62			0.74

 Data are expressed as percentages. High myopia, SE ≤ −3.0 D; low myopia, −3.0 < SE ≤ −0.5 D; without myopia, SE > −0.5 D.

*  By two-tailed test for trend across risk factors for total population. Post hoc pair-wise comparisons using Bonferroni’s test where P < 0.05 for: income for higher myopes versus nonmyopes and higher myopes versus lower myopes in the total sample; and parental myopia for higher myopes versus lower myopes, higher myopes versus nonmyopes, and lower myopes versus nonmyopes in the total sample.

^†  There were 35 missing values for income, 66 for parental myopia, and 20 for night lighting before 2 years of age.

Table 3.

 

Comparison of Children with Higher Myopia or Lower Myopia and without Myopia, by Different Nearwork Activities

	Chinese					Non-Chinese					Total					Total (n)
		Higher Myopes (n = 73)	Lower Myopes (n = 197)	Nonmyopes (n = 459)	Higher Myopes (n = 8)	Lower Myopes (n = 47)	Nonmyopes (n = 221)	Higher Myopes (n = 81)	Lower Myopes (n = 244)	Nonmyopes (n = 680)
Books read per week (n)*
Mean ± SD	4.4 ± 6.0	2.6 ± 2.4	2.4 ± 2.0	3.4 ± 1.9	2.5 ± 2.8	2.6 ± 2.4	4.3 ± 5.8	2.6 ± 2.5	2.5 ± 2.2	1,005
Median (range)	3.0 (0, 50)	2.0 (0, 20)	2.0 (0, 14)	3.0 (1, 7)	2.0 (0, 15)	2.0 (0, 20)	3.0 (0, 50)	2.0 (0, 20)	2.0 (0, 20)
P		<0.001			0.77			<0.001
Reported reading in hours per day (n)*
Mean ± SD	2.8 ± 1.1	2.5 ± 1.1	2.8 ± 1.3	2.2 ± 1.0	1.6 ± 0.8	1.6 ± 0.8	2.8 ± 1.1	2.3 ± 1.1	2.4 ± 1.3	981
Median (range)	2.7 (0.3, 5.6)	2.3 (0, 6.5)	2.4 (0, 8)	2.3 (1, 3.4)	1.7 (0, 4)	1.7 (0, 5.4)	2.7 (0.3, 5.6)	2.0 (0, 6.5)	2.3 (0, 8)
P		0.78			0.31			<0.001
Reported diopter hours (n)*
Mean ± SD	10.4 ± 4.1	9.4 ± 4.5	10.3 ± 4.9	9.8 ± 4.0	7.0 ± 3.4	7.3 ± 4.2	10.3 ± 4.0	9.0 ± 4.4	9.3 ± 4.8	957
Median (range)	10.1 (1.8, 23.4)	8.7 (0, 34)	9.8 (2.1, 30.1)	10.2 (3.9, 14.2)	7.0 (0, 14.9)	6.9 (0, 21.4)	10.1 (1.8, 23.4)	8.4 (0, 34)	8.6 (0, 30.1)
P		0.98			0.37			0.08
Computer use (%), †
Yes	12.5	26.1	61.4	2.5	19.4	78.1	10.0	24.4	65.6	404
No	6.2	28.5	65.3	3.4	13.8	82.8	5.4	24.3	70.3	582
P		0.023			0.45			0.03
Tuition classes (%), †
Yes	11.2	26.0	62.8	4.1	12.1	83.8	10.0	23.5	66.5	424
No	7.8	28.4	63.8	2.4	20.4	77.2	5.7	25.2	69.1	561
P		0.30			0.19			0.05

*  By two-tailed test for trend across refractive error groups for total population. Post hoc pair-wise comparisons using Bonferroni’s test where P < 0.05 for: books read per week for higher myopes versus lower myopes and higher myopes versus nonmyopes in the total sample; books read per week for higher myopes versus lower myopes and higher myopes versus nonmyopes in Chinese sample; and reading for higher myopes versus lower myopes and higher myopes versus nonmyopes in the total sample.

^†  By two-tailed χ² test for total population. Post hoc pair-wise comparisons using Bonferroni’s test where P < 0.05 for: computer use for higher myopes versus nonmyopes in total sample; computer use for higher myopes versus lower myopes and higher myopes versus nonmyopes in Chinese sample.

Table 4.

 

Adjusted Means of Axial Length of Eyes in Children with Different Characteristics

	Chinese* (n = 729)	Non-Chinese* (n = 276)	Total Sample, † (n = 979)	Total (n)
Total family income per month, ‡
Sin$2,000	23.26 (23.14, 23.39)	23.09 (22.98, 23.20)	23.21 (23.12, 23.30)	373
Sin$2,001–5,000	23.35 (23.23, 23.46)	23.13 (22.95, 23.31)	23.30 (23.20, 23.40)	305
>Sin$5,000	23.56 (23.45, 23.67)	23.39 (22.83, 23.96)	23.52 (23.41, 23.63)	267
Number of books read per week
≤2 books per week	23.31 (23.23, 23.40)	23.08 (22.97, 23.20)	23.25 (23.18, 23.31)	615
>2 books per week	23.57 (23.46, 23.68)	23.17 (23.02, 23.32)	23.49 (23.39, 23.58)	364
Reported reading in hours per day
≤2 hours per day	23.40 (23.30, 23.51)	23.07 (22.96, 23.17)	23.30 (23.22, 23.37)	507
>2 hours per day	23.41 (23.33, 23.50)	23.25 (23.08, 23.42)	23.36 (23.28, 23.44)	472
Reported Diopter hours
≤8	23.45 (23.33, 23.56)	23.04 (22.92, 23.16)	23.32 (23.24, 23.41)	422
>8	23.39 (23.30, 23.47)	23.19 (23.03, 23.34)	23.33 (23.25, 23.40)	557
Parental myopia, ‡
None	23.20 (23.08, 23.32)	23.05 (22.94, 23.17)	23.17 (23.08, 23.26)	360
One	23.40 (23.30, 23.50)	23.20 (23.02, 23.38)	23.35 (23.26, 23.44)	351
Two	23.65 (23.53, 23.78)	23.12 (22.78, 23.45)	23.57 (23.45, 23.69)	202

 Data are adjusted means with 95% CI in parentheses.

*  Gender-age–adjusted means.

^†  Gender-age-race–adjusted means.

^‡  There were 34 missing values for income and 66 for parental myopia.

Figure 2.

 

Age-gender-race adjusted axial lengths by books per week and parental myopia. (♦) Two or fewer books per week; (▪) more than two books per week; *Two-tailed P < 0.001 for interaction (books per week × parental myopia) by multiple linear regression.

The authors thank the staff of the Singapore Eye Research Institute for help in data collection and Wallace S. Foulds for critical reading of the manuscript. RAS was supported by grants from Research to Prevent Blindness and the Paul and Evanina Bell Mackall Foundation Trust.