Abstract
purpose. This study investigated the incidence and progression, as well as factors associated with changes in astigmatism in school children.
methods. This was a prospective cohort study. Children 7 to 9 years of age, of Chinese, Malay, and Asian Indian ethnicity, were examined annually over a 4-year period. Cycloplegic autorefraction was performed. A questionnaire was used to evaluate risk factors for incidence and progression of astigmatism.
results. The cumulative 3-year incidence rate of astigmatism was 33.6% (cylinder power of 0.5 D or worse) or 11.5% (cylinder power of 1.0 D or worse). Myopic children had a higher incidence rate of astigmatism than nonmyopes (P < 0.001). The mean J0 change per year was 0.012 D (95% CI: 0.007–0.018), whereas J45 did not show a significant change each year (mean, 0.001 D per year). Chinese children had greater worsening of J0 per year (P < 0.001). Girls also had significantly greater progression of J0 than did boys (P < 0.001). Similarly, myopia at baseline (P < 0.001) and the hours of computer use (P = 0.049) were associated with a greater progression rate of J0. J0 tended to improve in children with compound hyperopic astigmatism, worsen in children with compound myopic astigmatism, and remain stable in mixed astigmatics.
conclusions. Although there was minimal progression of astigmatism in school age children (0.44–0.53 D) over this period of follow-up, incident cases of astigmatism (>1.0 D) were not uncommon. The progression rate of astigmatism was affected by the ethnicity, presence of myopia, axis, and subtype of astigmatism.
Astigmatism is an important refractive condition in children
1 because it imposes considerable optical defocus at all viewing distances during the years of ocular maturation. Although debatable,
2 astigmatism is a possible cause of meridional amblyopia.
3 In addition, there is evidence of difficulty of correcting high astigmatism in clinical practice, leading to rejection of eye wear in children, with a decrease in classroom performance as a result of unsatisfactory vision.
4
The axes of astigmatism differed between races,
5 suggesting a heritable component. A more recent study showed the importance of dominant inheritance in astigmatism.
6 Compared with myopia, the proportion of the variance of astigmatism explainable by genetic factors appeared to be relatively lower.
6 7 In infants, the severity of astigmatism was reduced from 1 to 3 years.
8 9 10 A study of astigmatism from infancy to late childhood showed that the mean cylindrical power remained stable between 7 and 9 years of age, but incidence rates of astigmatism during school age were not reported.
11 The temporal changes in astigmatism within a population over a few decades had been studied,
12 but these studies had not been performed within a cohort. Therefore, the change in astigmatism constituted the least studied aspect of refractive error development in school age children. One population-based incidence study involving a repeat examination after 2 years was performed in the Shunyi District in China
13 in subjects between 5 and 13 years of age. In that study, the mean change in the power of the astigmatic cylinder was found to be 0. However, the progression analysis did not consider the astigmatic cylinders to be vector quantities,
14 therefore limiting the usefulness of the conclusions.
We conducted a study to evaluate the incidence and progression rate of astigmatism in a cohort of school children in Singapore and to investigate the possible epidemiologic associations of these longitudinal changes.
Astigmatism was defined as a cylinder power of at least 0.5 or 1.0 D. Myopia was defined as a spherical equivalent of at least −0.5 D.
J0 and
J45 were Fourier-transformed vectors
18 derived from the average of five refractions, as described in a previous report.
14 Briefly, if
C is the cylindrical power, and
A the axis of the cylinder in the spherical/cylinder notation,
J0 = −(
Ccos2
A)/2 and
J45 = −(
Csin2
A)/2.
The cumulative incidence rate of astigmatism was defined as the proportion of participants with astigmatism present at the 3-year follow-up examination in subjects without astigmatism at the baseline visit. The age-, gender-, and race-specific cumulative incidence rate of astigmatism and 95% confidence intervals (CI) were calculated. The multivariate adjusted hazard ratios and 95% CI were derived with the Cox proportional hazards model.
Change in J0 or J45 was defined as the J0 or J45 at the baseline examination subtracted from the final examination divided by the duration of follow-up in years. Multiple linear regression models were constructed with changes in refraction as the dependent variable and the relevant demographic factors as covariates.
For progression studies, only subjects with at least two visits were included. For studies investigating the sign of J0, the J0 may have been exactly 0 in the baseline visit, in which case the sign of J0 in the second visit was used in the analysis.
The A and V patterns of astigmatism symmetry between right and left eyes were defined as in a previously published report.
19 Briefly, a positive
J45 (axis at 45°) in the right eye and negative axis in the left eye (axis at 135°) refer to a V-pattern symmetry, whereas a negative
J45 (axis at 135°) in the right eye and positive
J45 in the left eye (axis at 45°) produces an A-pattern symmetry.
Compound hyperopic astigmatism was defined as refractions with positive powers in each meridian, and compound myopic astigmatism was defined as those with negative powers in each meridian. Mixed astigmatism referred to cases with positive power in one meridian and negative power in the other.
Statistical significance was set at the level of α = 0.05. As not all children had the same number of follow-up visits, analyses comparing subgroups were repeated with mean values weighted for the number of visits. The number of children eligible for the incidence of astigmatism in each of the visits is shown in
Figure 1 . All analyses referred to the values for the right eyes, except for the symmetry of astigmatism analyses, which naturally involve both eyes. Statistical analysis was conducted using the commercially available software (SAS, Cary, NC).
The children (
n = 1019) were examined at the baseline visit. Among the initial 1019 children, 197 (19.3%) had cylinder powers of 1.0 D or worse. Among the remaining 822 children, 680 children attended a 3-year follow-up visit
(Fig. 1) . For students who were examined in all follow-up visits (
n = 680), the mean cylinder at baseline was 0.44 D (SD: 0.25), and median cylinder was 0.43 D (range, 0.00–1.00). At the last follow-up visit, the mean cylinder was 0.53 D (SD: 0.41) and median cylinder was 0.45 D (range, 0.00–2.80).
The 3-year cumulative incidence rate of astigmatism (cylinder at least 1.0 D) was 11.5% (95% CI: 9.1–13.9;
Table 1 ). Myopic children at baseline had a higher incidence rate of astigmatism than nonmyopes (
P < 0.001). The hazard ratio of the development of astigmatism in myopes compared with nonmyopes was 5.40 (95% CI: 3.59–8.10), when controlling for age, gender, and ethnicity. In the children without myopia, the 3-year incidence rate of astigmatism (at least 1.0 D) was only 4.5% (95% CI: 2.7–6.4), compared with 29.2% (95% CI: 22.7–35.6) in children with myopia. When astigmatism was defined as a cylinder magnitude of 0.5 D or worse, the 3-year cumulative incidence rate was 33.6% (95% CI: 29.1–38.1).
The mean J0 at baseline was 0.28 D (95% CI: 0.25–0.30), whereas the mean J45 at baseline was 0.01 D (95% CI: 0.00–0.02). The mean absolute J45 at baseline was 0.11 D (95% CI: 0.10–0.12).
The
J0 change per year was 0.012 D (95% CI: 0.007–0.018;
Table 2 ). Chinese children had a worsening of
J0 that averaged 0.022 D, whereas non-Chinese children had a reduction in
J0 by 0.014 D. The difference in the progression rate of
J0 between ethnic groups was significant (
P < 0.001). Girls also had significantly greater progression of
J0 than did boys (
P < 0.001). Similarly, myopia at baseline, and the use of computers were also associated with a greater progression rate of
J0 (
P < 0.049).
The absolute
J45 did not appear to change over the period of the study: 0.001 D per year (95% CI: −0.001 to +0.004;
Table 3 ). However, the Chinese children showed a fluctuation of
J45 at 0.004 D per year (95% CI: 0.000–0.007), whereas the non-Chinese children showed −0.005 D per year (95% CI: −0.010–0.000;
P = 0.007). The myopes also showed a significantly increasing slope compared with nonmyopes (
P = 0.003). Ethnicity (
P = 0.02) and myopia status (
P = 0.03) were still significant when the analysis was adjusted for age and gender.
When the J0 and J45 progression rates were weighted according to the number of follow-up visits, the results were essentially the same (not shown).
The rates of progression of J0 were separately analyzed in children with different subtypes of astigmatism at baseline (i.e.,compound myopic astigmatism, mixed astigmatism, and compound hyperopic astigmatism). In those with compound myopic astigmatism (n = 320), the progression rate was 0.057 D per year (95% CI: 0.047–0.066). In those with mixed astigmatism (n = 86), there was no significant progression of J0 (0.000 D/y; 95% CI: −0.019 to +0.020). In those with compound hyperopic astigmatism (n = 552), the J0 showed a reducing trend at –0.012 D/y (95% CI: −0.017 to −0.007). The analysis for progression of J0 and J45 was repeated in those children with initial cylinder power of 0.5 D or worse and repeated again for those with initial cylinder power of 1.0 D, and essentially the same results were found (not shown).
Most children with positive
J0 (with-the-rule astigmatism) remained with-the-rule at the end of the follow-up, with only 8.3% converting to against-the-rule astigmatism.
(Table 4) . On the contrary, of the children with negative
J0 (against-the-rule astigmatism) at baseline, 41.2% converted to with-the-rule astigmatism.
At the baseline visit, the most common type of oblique astigmatism was V-pattern symmetry, followed by −− (axes at 135° in both eyes), A-pattern symmetry, and ++ (axes at 45° in both eyes;
Table 5 ). This distribution of symmetry persisted at the last visit of the follow-up, with most children retaining the original pattern of astigmatism symmetry.
Because of an increase in the mean cylinder power of only 0.09 D over 3 years (0.44–0.53 D) in this cohort of school children, the 3-year cumulative incidence rate of astigmatism of 33.6% (cylinder power of 0.5 D or worse) or 11.5% (cylinder power of 1.0 or worse) was not expected. This discrepancy could be related to measurement fluctuation and differing trends of astigmatism in some children. Cases of decreasing astigmatic severity may reduce the public health significance of the incidence rate. We found that the incidence of astigmatism was five times greater in myopes than in nonmyopes. With-the-rule astigmatism tended to increase slightly in this longitudinal study, with myopes, girls, and Chinese children showing significantly increasing trends. A greater number of hours of computer use were also associated with this worsening. The magnitude of oblique astigmatism (J45) did not change significantly in the entire cohort, but the Chinese and myopes showed an increasing trend. The effect of ethnicity remained after adjustment for myopic status and other demographic factors. Against-the-rule astigmatism tended to become with-the-rule, rather than vice versa. J0 progression was found in astigmatic children with compound myopia, J0 reduction was found for those with compound hyperopia, and no detectable change in J0 was found in those with mixed astigmatism.
In the Shunyi study,
13 it was found that the astigmatic error showed little change over the 28.5-month period. A decrease of 0.004 cylindrical diopter (SD 0.301) was found. Almost all eyes had cylindrical powers of less than 0.75 D at baseline as well as the repeat measurement. In the eyes with 0.75 D or more (8.4%) at the initial measurement, a general reduction in the magnitude of the astigmatism was found. Increasing astigmatic errors of above 0.50 D or more were associated with older age, and increases of 1.0 D or more were associated with greater cylinder powers at baseline and female gender. However, the Shunyi study found that the reduction of the cylinder was not associated with age and sex on multiple regression. These results of astigmatic change, however, did not treat cylinders as vectors, and therefore could not be directly compared with findings in the present study. It has been mentioned that vectorial analysis is necessary for the algebraic manipulation of astigmatic cylinders,
18 because the direction as well as the magnitude of cylinders could change with time.
The strengths of this study included a reasonable sample size and homogeneity of the method of refraction and biometry, performed in an age group and ethnic background that had not been adequately studied in the past. In the present study, the schools were not randomly selected, and this may limit the ability to generalize the findings to all school children. Because of the different subjects, the results of this study may not be applicable to children Western countries. The 3-year lost-to-follow-up rate was low
(Fig. 1) .
The present findings suggests that severity of astigmatism remained stable in mean cylinder power in school age children. Being myopic was the major risk factor for development of astigmatism. In general, if nonmyopic school children had no astigmatism (using the same definition), the chances of having cylinder powers of above 1.0 D over the next 3 years was fairly low (on the order of 3.4%). Gwiazda et al.
11 reported that cylinder power may increase slightly after 10 years of age, although this trend was not statistically significant. More studies should be performed in older children to determine whether astigmatism progresses beyond the age range in the present study. The association of computer use with progression of
J0 remained even after adjustment for myopia and had not been reported in previous studies on astigmatism. Computer use was found to be associated with
prevalence of astigmatism in a previous report based on children in this same cohort study.
19 The exact explanation for this association was unclear and could be explored in further studies.
To summarize, in Singaporean school children, the change in mean cylinder power was slight over 3 years. The 3-year incidence rate of astigmatism was 11.5% (cylinder power of 1.0 D or worse). The presence of myopia at baseline was the most important risk factor for incident cases of astigmatism. Progression of astigmatism was generally not severe, but was shown to be affected by ethnicity, presence of myopia, the axis, and subtype of the astigmatism. In addition, the progression of with-the-rule astigmatism may be affected by the number of hours of computer use.
Supported by Singapore Eye Research Institute Grant MG/97-04/005 and National Medical Research Council Grant 0695/2003.
Submitted for publication May 4, 2004; revised June 1 and June 23, 2004; accepted July 1, 2004.
Disclosure:
L. Tong, None;
S.-M. Saw, None;
Y. Lin, None;
K.-S. Chia, None;
D. Koh, None;
D. Tan, None
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.
Corresponding author: Louis Tong, Singapore National Eye Center, 11 Third Hospital Avenue, Singapore 168751;
[email protected].
Table 1. Incidence Rates of Astigmatism
Table 1. Incidence Rates of Astigmatism
| Nonmyopes | | Myopes | | 3-Year Cumulative Incidence of Astigmatism* , † | Multivariate Hazard Ratios of Astigmatism* , † |
| n (Baseline) | 3-Year Cumulative Incidence of Astigmatism | Multivariate Hazard Ratios of Astigmatism* | n (Baseline) | | |
| All | 680 | 11.5 (9.1–13.9) | | 268 | 3.4 (1.2–5.5) | |
| Age at baseline (y) | | | | | | |
| 7 | 342 | 9.6 (6.5–12.8) | 1.00 | 130 | 3.9 (0.5–7.2) | 1.00 |
| 8 | 232 | 14.7 (10.1–19.2) | 1.24 (0.82–1.85) | 97 | 2.1 (0.0–4.9) | 1.64 (0.50–5.37) |
| 9 | 106 | 10.4 (4.6–16.2) | 0.69 (0.38–1.26) | 41 | 4.9 (0.0–11.5) | 1.27 (0.24–6.59) |
| P | | 0.41 | 0.23 | | 0.99 | 0.78 |
| Race | | | | | | |
| Chinese | 502 | 13.9 (10.9–17.0) | 1.00 | 165 | 3.6 (0.8–6.5) | 1.00 |
| Non-Chinese | 178 | 4.5 (1.5–7.5) | 0.81 (0.48–1.37) | 103 | 2.9 (0.0–6.2) | 1.02 (0.33–3.14) |
| P | | <0.001 | 0.42 | | 0.75 | 0.97 |
| Gender | | | | | | |
| Male | 336 | 10.7 (7.4–14.0) | 1.00 | 142 | 3.5 (0.5–6.6) | 1.00 |
| Female | 344 | 12.2 (8.8–15.6) | 1.10 (0.75–1.60) | 126 | 3.2 (0.1–6.2) | 0.98 (0.33–2.93) |
| P | | 0.54 | 0.62 | 88 | 0.015 | 0.97 |
| Myopia at baseline | | | | | | |
| Nonmyopia | 488 | 4.5 (2.7–6.4) | 1.00 | — | — | — |
| Myopia | 192 | 29.2 (22.7–35.6) | 5.40 (3.59–8.10) | — | — | — |
| P | | <0.001 | <0.001 | | | |
Table 2. J0 Progression for All Students with at Least One Follow-up Visit Over 3 Years
Table 2. J0 Progression for All Students with at Least One Follow-up Visit Over 3 Years
| n (Baseline) | Astigmatism (J0) | Adjusted Astigmatism (J0)* |
| All | 981 | 0.012 (0.007–0.018) | 0.018 (0.012–0.024) |
| Age at baseline (y) | | | |
| 7 | 510 | 0.010 (0.003–0.017) | 0.019 (0.012–0.027) |
| 8 | 315 | 0.015 (0.006–0.024) | 0.019 (0.011–0.028) |
| 9 | 156 | 0.013 (0.005–0.026) | 0.015 (0.003–0.026) |
| P | | 0.71 | 0.47 |
| Race | | | |
| Chinese | 714 | 0.022 (0.017–0.028) | 0.030 (0.024–0.037) |
| Non-Chinese | 267 | −0.014 (−0.024–−0.005) | 0.005 (−0.004–0.015) |
| P | | <0.001 | <0.001 |
| Gender | | | |
| Male | 488 | 0.003 (−0.004–0.011) | 0.007 (−0.000–0.015) |
| Female | 493 | 0.021 (0.014–0.028) | 0.028 (0.021–0.036) |
| P | | <0.001 | <0.001 |
| Myopia at baseline | | | |
| Nonmyopia | 660 | −0.008 (−0.014–−0.003) | −0.014 (−0.020–−0.007) |
| Myopia | 321 | 0.055 (0.047–0.063) | 0.049 (0.040–0.058) |
| P | | <0.001 | <0.001 |
| Computer (h/d) | | | |
| ≤0.5 | 549 | 0.009 (0.002–0.015) | 0.013 (0.006–0.020) |
| >0.5 | 416 | 0.019 (0.011–0.027) | 0.022 (0.015–0.030) |
| P | | 0.04 | 0.049 |
Table 3. Progression of J45 for All Children with at Least One Follow-up Visit Over 3 Years
Table 3. Progression of J45 for All Children with at Least One Follow-up Visit Over 3 Years
| n (Baseline) | Absolute Astigmatism (J45) | Adjusted Absolute Astigmatism (J45)* |
| All | 981 | 0.001 (−0.001–0.004) | 0.002 (−0.001–0.006) |
| Age at baseline (y) | | | |
| 7 | 510 | −0.003 (−0.006–0.001) | −0.003 (−0.007–0.001) |
| 8 | 315 | 0.006 (0.001–0.011) | 0.005 (−0.000–0.010) |
| 9 | 156 | 0.005 (−0.002–0.012) | 0.004 (−0.002–0.011) |
| P | | 0.06 | 0.06 |
| Race | | | |
| Chinese | 714 | 0.004 (0.000–0.007) | 0.006 (0.002–0.010) |
| Non-Chinese | 267 | −0.005 (−0.010–0.000) | −0.002 (−0.007–0.004) |
| P | | 0.007 | 0.02 |
| Gender | | | |
| Male | 488 | 0.003 (−0.001–0.007) | 0.003 (−0.001–0.008) |
| Female | 493 | −0.000 (−0.004–0.004) | 0.001 (−0.003–0.005) |
| P | | 0.27 | 0.40 |
| Myopia at baseline | | | |
| Nonmyopia | 660 | −0.001 (−0.005–0.002) | −0.001 (−0.005–0.003) |
| Myopia | 321 | 0.007 (0.003–0.012) | 0.006 (0.000–0.011) |
| P | | 0.003 | 0.03 |
Table 4. Changes in the Direction of J0 (with-the-Rule Astigmatism) in Children with One Follow-up Visit Over 3 Years
Table 4. Changes in the Direction of J0 (with-the-Rule Astigmatism) in Children with One Follow-up Visit Over 3 Years
| Baseline J0 | n (Baseline) | J0 Positive at Final Visit* | J0 Negative at Final Visit, † |
| All | 975 | 819 (84.0) | 156 (16.0) |
| J0 positive | 827 | 758 (91.7) | 69 (8.3) |
| J0 positive | 148 | 61 (41.2) | 87 (58.8) |
Table 5. Changes in Right–Left Eye Symmetry of J45 (Oblique Astigmatism) in Children with at Least One Follow-up Visit over 3 Years
Table 5. Changes in Right–Left Eye Symmetry of J45 (Oblique Astigmatism) in Children with at Least One Follow-up Visit over 3 Years
| J45 Symmetry of Right and Left Eyes at Baseline | n | J45 Symmetry of Right and Left Eyes at Final Visit | | | |
| | ++ | +− | −+ | −− |
| All | 976 | 118 (12.1) | 433 (44.4) | 177 (18.1) | 248 (25.4) |
| ++ | 104 | 40 (38.5) | 31 (29.8) | 12 (11.5) | 21 (20.2) |
| +− | 407 | 32 (7.9) | 283 (69.5) | 24 (5.9) | 68 (16.7) |
| −+ | 204 | 28 (13.7) | 32 (15.7) | 91 (44.6) | 53 (26.0) |
| −− | 261 | 18 (6.9) | 87 (33.3) | 50 (19.2) | 106 (40.6) |
Miller JM, Dobson V, Harvey EM, Sherrill DL. Comparison of preschool vision screening methods in a population with a high prevalence of astigmatism. Invest Ophthalmol Vis Sci
. 2001;42:917–924.
[PubMed]Ingram RM, Walker C, Wilson JM, Arnold PE, Dally S. Prediction of amblyopia and squint by means of refraction at age 1 year. Br J Ophthalmol
. 1986;70:12–15.
[CrossRef] [PubMed]Abrahamsson M, Fabian G, Andersson AK, Sjostrand J. A longitudinal study of a population based sample of astigmatic children. I. Refraction and amblyopia. Acta Ophthalmol (Copenh)
. 1990;68:428–434.
[PubMed]Garber JM. High corneal astigmatism in Navajo school children and its effect on classroom performance. J Am Optom Assoc
. 1981;52:583–586.
[PubMed]Thorn F, Held R, Fang LL. Orthogonal astigmatic axes in Chinese and Caucasian infants. Invest Ophthalmol Vis Sci
. 1987;28:191–194.
[PubMed]Hammond CJ, Snieder H, Gilbert CE, Spector TD. Genes and environment in refractive error: the twin eye study. Invest Ophthalmol Vis Sci
. 2001;42:1232–1236.
[PubMed]Mash AJ, Hegmann JP, Spivey BE. Genetic analysis of indices of corneal power and corneal astigmatism in human populations with varying incidences of strabismus. Invest Ophthalmol
. 1975;14:826–832.
[PubMed]Ingram RM, Barr A. Changes in refraction between the ages of 1 and 3 1/2 years. Br J Ophthalmol
. 1979;63:339–342.
[CrossRef] [PubMed]Gwiazda J, Scheiman M, Mohindra I, Held R. Astigmatism in children: changes in axis and amount from birth to six years. Invest Ophthalmol Vis Sci
. 1984;25:88–92.
[PubMed]Abrahamsson M, Fabian G, Sjostrand J. Changes in astigmatism between the ages of 1 and 4 years: a longitudinal study. Br J Ophthalmol
. 1988;72:145–149.
[CrossRef] [PubMed]Gwiazda J, Grice K, Held R, McLellan J, Thorn F. Astigmatism and the development of myopia in children. Vision Res
. 2000;40:1019–1026.
[CrossRef] [PubMed]Maples WC, Atchley J, Ashby W, Ficklin T. An epidemiological study of the ocular and visual profiles of Oklahoma Cherokees and Minnesota Chippewas. J Am Optom Assoc
. 1990;61:784–788.
[PubMed]Zhao J, Mao J, Luo R, et al. The progression of refractive error in school-age children: Shunyi district, China. Am J Ophthalmol
. 2002;134:735–743.
[CrossRef] [PubMed]Tong L, Carkeet A, Saw SM, Tan D. Corneal and refractive error astigmatism in Singaporean school children: a vector based Javal’s rule. Optom Vis Sci
. 2001;78:881–887.
[CrossRef] [PubMed]Saw SM, Chua WH, Hong CY, et al. Nearwork in early onset myopia. Invest Ophthalmol Vis Sci
. 2002;43:332–339.
[PubMed]Saw SM, Chua WH, Hong CY, et al. Height and its relationship to refraction and biometry parameters in Singapore Chinese children. Invest Ophthalmol Vis Sci
. 2002;43:1408–1413.
[PubMed]Zadnik K, Mutti DO, Adams AJ. The repeatability of measurement of the ocular components. Invest Ophthalmol Vis Sci
. 1992;33:2325–2333.
[PubMed]Thibos LN, Wheeler W, Horner D. Power vectors: an application of Fourier analysis to the description and statistical analysis of refractive error. Optom Vis Sci
. 1997;74:367–375.
[CrossRef] [PubMed]Tong L, Saw SM, Carkeet A, et al. Prevalence rates and epidemiological risk factors for astigmatism in Singapore school children. Optom Vis Sci
. 2002;79:606–613.
[CrossRef] [PubMed]