Longitudinal Study of Anisometropia in Singaporean School Children

Abstract

purpose. The purpose of this study was to document the incidence rates of anisometropia, year-by-year prevalences, changes in the intereye difference in spherical equivalent (SE), and its association with myopia progression and axial length changes in a cohort of Singaporean school children.

methods. This is a prospective cohort study of Singaporean school children (n = 1979) aged 7 to 9 years who were examined annually with cycloplegic refraction and ultrasonography over a 3-year period.

results. In the 1908 children without anisometropia at commencement, the 3-year cumulative incidence rate of anisometropia (difference in SE at least 1.0 D) was 144 (7.55%; 95% CI: 6.42–8.85). The mean intereye difference in SE in all children at baseline was 0.29 ± 0.46 D (SD: 0.46) and increased to 0.44 D (0.59) on the last examination. On the initial examination, 3.6% (95% CI: 2.8–4.4) or 71 children had anisometropia. Of the 59 of 71 children who completed all examinations, only 3 (5.1%) had an increase in the intereye difference in SE by at least 0.5 D, whereas 2 (3.4%) had a decrease of at least 0.5D. The mean intereye difference in SE was stable between visits. The change in intereye difference in SE correlated with the change in intereye axial length (r = 0.43). Compared with the isometropic children, each eye of the anisometropic children had a higher rate of progression of myopia.

conclusions. The 3-year incidence of anisometropia was 7.55% in these young Singaporean children. Although the frequency of anisometropia increased with time, the difference in SE between eyes tended to remain stable.

Anisometropia, a difference in refractive power between the two eyes, is associated with aniseikonia, possible spectacle intolerance, ^{1 2 3 4 5 6 7} and binocular defects. ^{1 8 9 10 11 12} Although the structural basis of anisometropia is related to the intereye difference in axial length, ^{13 14 15} the physiological basis for a differential development of axial lengths between the two eyes of a child remains elusive. A decrease in binocular coordination has been suggested as a possible reason for development of anisometropia, ¹⁶ although it is difficult to show experimentally that the latter did not cause the former. 

Although the severity of myopia is related to the intereye difference in SE, ¹⁷ as shown in a prevalence report by the current authors, ¹⁸ it has yet to be shown that the longitudinal increase in myopia correlates with anisometropia. Similarly, the intereye difference in axial length is associated with the intereye difference in SE, ¹⁸ although the change in the intereye difference in SE has not been linked to a change in the intereye difference in axial length. 

A longitudinal study of 350 Japanese school children aged 6 to 11 years was performed with cycloplegic refraction, but only prevalence figures for anisometropia (4.3%) were reported. ¹⁹ Although a study of anisometropia has been performed on retrospective data, ²⁰ such results have not been reported from a prospectively designed study. This study, ²⁰ which included only subjects with myopia in one eye, reported the change of spherical equivalents in each eye with time but not the change of the difference in SE between eyes. The authors ²⁰ mentioned that anisometropia results from the slower development or lack of development of myopia in the emmetropic eye. It would be interesting to verify this finding in a larger study and also to evaluate, among bilaterally myopic children, whether anisometropia results from too rapid a change of spherical refractive error with time in one eye or a relatively slower process in the other, when compared with eyes of children in whom anisometropia did not develop. 

We sought first to evaluate the incidence rates of new cases of anisometropia, year-by-year prevalences, and the progression rates of established cases of anisometropia, and second, to evaluate the changes in anisometropia and correlations with changes in myopia’s progression and the intereye difference in axial length. 

Methods

Subjects

The cross-sectional results of an ongoing longitudinal observational study ^{21 22} in Singapore have been reported. The study commenced in 1999, and school children in grades 1 to 3, aged 7 to 9 years and attending three Singaporean schools (located on the Eastern, Northern and Western part of Singapore) were recruited. This study was approved by the ethics committee of the Singapore Eye Research Institute, and all procedures adhered to the Declaration of Helsinki. Informed written consent was obtained from parents of the children. 

Exclusion criteria included serious medical conditions such as leukemia or heart disorders and allergies to eye drops but not amblyopia. 

The overall participation rate was 62%. The proportion of children who reported myopia before the school eye examination was not different between participants (27.3%) and nonparticipants (26.8%). 

A total of 1979 children (equal representation of the two genders), with a mean age of 7.8 years (SD 0.83) were studied, with Chinese being the dominant ethnic group (n = 1481), followed by Malay (n = 324) and Asian Indian (n = 174). 

Measurements

The cycloplegic regimen consists of a drop of 1% cyclopentolate repeated twice at 5-minutes interval. All measurements were performed at least 30 minutes after the last eye drop instillation. Cycloplegic refraction has been shown to be the most suitable method of assessing refractive errors in longitudinal epidemiologic studies. ²³  

Cycloplegic autorefraction and autokeratometry were performed in two perpendicular meridians (Canon RK-5; Canon Ltd., Tochigiken, Japan). A-mode ultrasonography was also performed (US-800 EchoScan; Nidek Ltd., Tokyo, Japan) to ascertain the axial length of the eyes. 

Definitions and Data Analysis

The refractive error, in diopters (D), was calculated as the spherical equivalent (SE): the addition of the spherical power and half the magnitude of the cylinder power. Anisometropia was defined as the difference in SE between eyes of at least 1.0 D. The intereye difference in SE was defined as the absolute difference in SE between the right and left eyes in diopters. 

Myopia in this study was defined as an SE of at least −0.5 D, and hyperopia as +0.5 D or more. Intermediate refractions were considered emmetropic. The year-by-year prevalence rates and 95% confidence intervals of anisometropia and 3-year cumulative incidence rate of anisometropia were computed. 

The intereye difference in SE as a continuous variable between subgroups (genders, ages, and ethnic groups) was determined with the nonparametric Mann-Whitney test (in the case of genders) and the Kruskal-Wallis test (in the case of ethnic groups and ages). 

To evaluate astigmatic anisometropia, two more variables were constructed. The derivation of the J0 and J45 vectors has been explained previously. ²⁴ In the current study, the absolute intereye difference in J0 was constructed to evaluate possible anisometropia for with-the-rule or against-the-rule astigmatism, whereas, the absolute intereye difference in J45 was constructed to evaluate possible anisometropia in oblique astigmatism. 

Trend analysis for the rate of change of the absolute difference in SE between both eyes over 3 years was assessed with repeated-measurement analysis. The rate of change of refractive error in each eye of the children was calculated and compared between children in whom anisometropia developed during follow-up and those in whom it did not. In further analyses, the group of children who had anisometropia at any visit was replaced by the subset of those children with any myopia at baseline, and subsequently, the analysis was repeated in the smaller subset of these children with uniocular myopia at baseline. These subgroups have been reported to be unique in the development anisometropia. ²⁰ The rate of change of the intereye difference in SE was calculated for subgroups with one myopic eye, those with both eyes myopic, those with hyperopic eyes in all the children, in the children with preexisting anisometropia, and in the subgroup without preexisting anisometropia. The Spearman correlation coefficient (r) was used to determine the relationship of two continuous variables (e.g., SE of the less ametropic eye and intereye SE difference). For all analyses, statistical significance was set at the level of α = 0.05. Data analysis was performed on computer (SPSS for Windows ver. 12.0; SPSS, Chicago, IL). 

Results

Analysis of All Children

The overall prevalence rate of anisometropia (difference in SE of at least 1.0 D) in all the children (n = 1979) aged 7 to 9 years at the commencement of the study (Table 1)was 3.6% (95% CI: 2.8–4.4). At the last annual visit (n = 1643), the prevalence of anisometropia (using the same definition) was 9.9% (95% CI: 8.5–11.4). Age was significantly associated with the prevalence of anisometropia (P = 0.008): 2.7% (95% CI: 1.6–3.8), 3.0% (95% CI: 1.7–4.4), and 5.8% (95% CI: 3.8–7.9) in the 7-, 8-, and 9-year-old children, respectively, at the baseline visit. Similar trends with age were seen at the first (P = 0.018), second (P = 0.060), and third (P = 0.026) visits, although the trend at the second visit was of borderline significance. The prevalence rate of anisometropia was not associated with gender at any visit. The prevalence rate of anisometropia in myopes (those with at least one myopic eye) was 56 (7.8%) of 719 (95% CI: 5.8–9.7) whereas in nonmyopes it was only 15 (1.2%) of 1260 (95% CI: 0.6–1.8). This difference was statistically significant (P < 0.001). At each visit, the presence of myopia was associated with anisometropia (all P < 0.001). 

Table 2shows the difference in SE between the two eyes in all the children assessed at each visit. The mean difference at baseline was 0.29 D (SD: 0.46), and increased to 0.44 D (SD: 0.59) at the third visit (P < 0.001). The difference in SE between eyes increased within strata of gender, age, and myopia status at baseline. Among the children with at least one myopic eye, the mean intereye difference in SE was 0.33 D (SD: 0.52), the median difference was 0.20 D (range: 0–8.35), whereas among the nonmyopic children, the mean intereye difference in SE was only 0.21 D (0.29), with a median of 0.15 D (range: 0–4.55). The difference between these groups was statistically significant (P < 0.001). Table 2shows the difference in axial length between the two eyes in all the children assessed at each visit. The data in Table 2were very similar to the result (not shown) of analyzing only children with data at all visits (n = 1568). The axial length data agreed with the refractive data, in that the intereye difference increased with age and with each subsequent measurement. 

The intereye differences in J0 and J45 are presented in Table 3 . The mean intereye difference in J0 was 0.16 D (SD: 0.17) at the commencement of the study (n = 1979) and 0.17 D (0.17) at the last visit (n = 1643). The mean intereye difference in J45 was 0.20 D (0.22) at the commencement (n = 1979) and 0.21 D (0.24) at the last visit (n = 1643). The corresponding means for the 1568 children with complete data at every visit were 0.16 D (0.17), 0.17 D (0.17), 0.20 D (0.22) and 0.21 D (0.24) (data not shown). The baseline and last visit data were very similar, and data for the two intervening visits (not shown) were also similar to those of the first and last visits. 

Of the 1908 children without anisometropia at the study’s commencement (difference in SE <1.0 D), 53, 48, and 43 developed anisometropia during a measurement at yearly examinations. When anisometropia developed by the first examination (n = 53) the mean and median difference in SE between eyes were 1.23 D (SD: 0.25) and 1.15 D (range: 1.00–2.43), respectively. For those children who had anisometropia in the second (n = 48) and the third (n = 43) annual examinations, the means of their intereye difference in SE were 1.44 D (0.38) and 1.49 D (0.47), respectively, in the year in which anisometropia developed; and the corresponding medians were 1.33 D (range: 1.00–2.88) and 1.34 D (range: 1.00–3.02), respectively. Therefore, the 3-year cumulative incidence rate of anisometropia was 144 (7.55%) of 1908 (95% CI: 6.42–8.85). In the children without anisometropia at baseline (n = 1908), and the mean difference in SE was 0.22 D (0.20) at commencement and 0.38 D (SD: 0.41) on the last examination. 

Table 4describes the proportion of the children who had anisometropia over the 3-year period for different refractive error categories in the entire study population. At baseline, anisometropia was more common in the children with unilateral myopia in one eye and emmetropia in the other (11.6%), followed by bilateral myopia (7.1%). The prevalence of anisometropia was relatively less in those with hyperopia in both eyes (2.8%). These proportions were not significantly different from one another (P > 0.05). At baseline, there were 56 children with myopic anisometropia and 15 children with hyperopic anisometropia. 

In the study subjects with complete 3-year follow-up at commencement (regardless of anisometropic status, n = 1568) the initial uni- or bilaterality of myopic status did not significantly influence the subsequent rate of change in difference in SE. The rate of change in difference in SE was similar (P = 0.081) in unilaterally myopic (n = 95) and bilaterally myopic subjects (n = 494) at commencement: 0.056 D (SD: 0.145) per year and 0.034 D (0.113) per year, respectively. In children without myopia at baseline (n = 1054), the rate of change in difference in SE was 0.037 (0.10) per year. 

The rates of refractive change in individual eyes were compared in the children with and without anisometropia at any visit (Table 5) . There was no difference in the change in refraction in the more ametropic eye compared with the less ametropic eye, but, in each of the eyes, the change in refraction was greater in children in whom anisometropia developed compared with those without any anisometropia (P < 0.001). The rate of change of refraction in each eye of unilateral myopes in whom anisometropia developed (n = 20) was compared with the corresponding eye in children in whom it did not (n = 1381). The same conclusion was obtained: The rate of change of refraction was greater in the children who had anisometropia. 

Analysis of Children with Preexisting Anisometropia

In the children with anisometropia at commencement who attended follow-up visits during the whole study period (n = 59), only 3 (5.1%) of 59 had an increase in anisometropia by at least 0.5 D whereas 2 (3.4%) of 59 had a decrease of at least 0.5 D. Therefore, more than 90% stayed the same. No child had a change in anisometropia of 1.0 D or more. 

Table 6describes the change in anisometropia in the children with anisometropia at baseline by type of anisometropia (myopic in one eye, myopic in both eyes, hyperopic in at least one eye). Among the children with anisometropia at baseline (n = 59) the mean difference in SE between eyes was 1.97 D (SD: 1.46) at commencement and 1.91 D (1.68) on the last examination. In these children, the mean intereye difference in axial lengths was 0.76 mm (SD: 0.59) at baseline and 0.70 mm (0.65) in the final measurement, the rate of change being −0.02 mm (0.24) per year. The change in anisometropia correlated with the change in the intereye difference in axial length (r = 0.43). The rate of change of SE in the less ametropic eye was also associated with the anisometropia progression rate (r = 0.40), especially in the subgroup with bilateral myopia (r = 0.45). These correlations were also significant for children without anisometropia at commencement (n = 1908) except that the correlation coefficient (r) values were lower (data not shown). The means in the table differed considerably from the medians, suggesting deviation of the distribution from normality. When the analysis was repeated and four outliers with baseline difference in SE greater than 4.0 D were excluded, the same conclusions were obtained regarding the significant correlations (data not shown). 

Discussion

The prevalence rates of anisometropia in the cohort at the four visits over 3 years were 3.6%, 5.6%, 7.4%, and 9.9%. The corresponding prevalence rates in the 1568 subjects with complete 3-year data were 3.6% (95% CI: 2.7–4.6), 5.7% (95% CI: 4.6–6.9), 7.3% (95% CI: 6.1–8.7) and 9.7% (95% CI: 8.3–11.3). Our study shows that new cases of anisometropia occurred at a rate of 7.55% over 3 years. Cases of anisometropia, however, tended not to progress significantly with time, and in some cases even regressed. The rate of change of difference in SE between eyes was similar in the children with myopia in one eye and those with myopia in both eyes. This study also showed that the longitudinal change in anisometropia was related to a change in intereye difference in axial length. Myopic anisometropia was more prevalent than hyperopic anisometropia, though this study population was largely emmetropic or myopic. In children with anisometropia, each eye had a faster change in refraction toward more myopia than in the children without anisometropia. 

In this study, the means of the intereye difference in J0 and J45 were all rather low and were not associated with significant change between initial and last visits. This suggests that cylindrical anisometropia may not be of great magnitude and does not fluctuate much in these children with time. 

In a previous report, we discussed the prevalence rates of anisometropia in various studies. ¹⁸ To our knowledge, the incidence rates of anisometropia in children have not been reported elsewhere. 

The mean difference in SE between eyes in 350 Japanese children ²⁰ (school-based) aged 7, 8, and 9 years were 0.22 D (95th percentile, 0.60), 0.21 D (95th percentile 0.58), and 0.25 D (95th percentile 0.62). This result was lower than the 0.25, 0.29 and 0.34 D, respectively, in our study. The same conclusion could be drawn by comparing the median values. In the Japanese study, ²⁰ unlike the present study, the age trend was not statistically significant (P = 0.117), although there was a slight increase in difference in SE between eyes with age. The proportion of children with anisometropia of 1.0 D or greater was 1.43% for children 6 to 8 years of age, and 2.3% for 9-year-old children. ²⁰ In the present study, only 5 (0.32%) of 1568 of the children with complete 3-year follow-up had a change in anisometropia of 0.5 D or more. In contrast, 15.7% of the Japanese children had a significant change in the magnitude of anisometropia (the criterion for change to be significant was not clear). ²⁰ Of interest was that 23 hyperopic children (6.6% of study population) had a reduction of anisometropia on sequential refractions. ²⁰ In our study, among those children without anisometropia, hyperopes are less prone to development of anisometropia than were other refractive groups (P < 0.011). 

In a 3-year longitudinal study in 238 schoolchildren with myopia, ¹⁷ anisometropia increased in 27% and decreased in 6%. The mean anisometropia increased from 0.30 to 0.51 D, ¹⁷ and in the present study a similar magnitude of increase was observed, from 0.29 to 0.44 D. 

Previously, the intereye difference in spherical equivalent was found to be associated with differences in axial lengths. ^{13 14 15} However, the correlation between the longitudinal changes in these two parameters have not been reported. In our study, the correlation between change in anisometropia and the change of intereye difference in axial length was significant (r = 0.429). 

There have been no prior reports of the different frequencies of anisometropia in different refractive error subgroups. The clinical implication of anisometropia in myopes differs from that in hyperopes. The latter is more predisposed to amblyopia with the same magnitude of anisometropia. In our study, anisometropia was slightly more frequent in myopes than in hyperopes, although the trend was not statistically significant for any visit (all P > 0.05). 

A previous study showed that cases of unilateral myopia in which anisometropia developed were related to a slower rate of change of SE in the companion eye rather than a accelerated rate in the initially myopic eye, relative to other myopes without anisometropia. ²⁰ Children in whom anisometropia did not develop had no change or an equal rate of SE change in each eye. A unilaterally slower rate of SE change compared with isometropic children has been proposed to be a cause of the development of anisometropia. ²⁰ However, our results did not support this hypothesis. We found that the rate of SE change in the less ametropic eye in children who became anisometropic was not slower than the rate encountered in children who did not. 

In our study, the SE change in each eye was accelerated compared with nonanisometropic children. This result suggests that the nonsymmetrically accelerated SE change between the left and right eye is the cause of anisometropia. 

We also found that baseline uniocular myopes did not have a significantly increased rate of difference in SE on follow-up, compared with bilateral myopes. This suggests that uniocular myopes may not be a unique subgroup in terms of development of anisometropia, contrary to the finding in the previous report. ²⁰  

The different conclusions between these studies may be related to differences in study methodology, such as the method of sampling, ethnicity of the population, and choice of refractive error groups and sample size. 

The strength of the present study is the relatively large sample size, availability of longitudinal data and uniform methods of biometry and refraction, as well as the use of cycloplegic refraction. A limitation is that we did not perform systematic best corrected refraction with subjective refinement of visual acuity. This limits the potential usefulness of our data in the evaluation of the possibility of the development of amblyopia. 

What could be the reason for the mismatch in the growth rates of the eyes of children with incident anisometropia? One explanation may be that the scleral growth or the output in the response to an abnormal visual stimulus is not associated with a constant gain in tissues from different eyes. Another explanation could be that subtle changes in lenticular or corneal properties could result in different image qualities in the two eyes. 

In the future, studies should investigate possible asymmetry in the visual images between eyes, such as those due to astigmatism and higher order aberrations. For now, most of the children who had preexisting anisometropia in this age group could be reassured that their anisometropia is generally not a progressive phenomenon. 

In summary, the incidence rate of new anisometropia is low among school children in Singapore, and preexisting anisometropic children tend to have a stable intereye difference in SE over time. The differential rate of increase of axial length is a probable biometric basis for anisometropia. 

 Supported by National Medical Research Council Grant NMRC/0695/2002.

 Submitted for publication July 14, 2005; revised September 6 and November 15, 2005, and March 16, 2006; accepted May 23, 2006.

 Disclosure: L. Tong, None; Y.-H. Chan, None; G. Gazzard, None; D. Tan, None; S.-M. Saw, 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.

 

Table 1.

 

Visit-by-Visit Assessment of Anisometropia Prevalence Rates

	Baseline Visit (n = 1979)	First Visit (n = 1876)	Second Visit (n = 1788)	Third Visit (n = 1643)
Overall	71 (3.6, 2.8–4.4)	105 (5.6, 4.6–6.6)	133 (7.4, 6.2–8.7)	163 (9.9, 8.5–11.4)
Gender
Male	36/1000 (3.6, 2.4–4.8)	55/946 (5.8, 4.3–7.3)	60/900 (6.7, 5.0–8.3)	74/827 (8.9, 7.0–10.9)
Female	35/979 (3.6, 2.4–4.7)	50/930 (5.4, 3.9–6.8)	73/888 (8.2, 6.4–10.0)	89/816 (10.9, 8.8–13.0)
P	0.976	0.680	0.211	0.184
Age at baseline (y)
7	23/851 (2.7, 1.6–3.8)	32/805 (4.0, 2.6–5.3)	45/777 (5.8, 4.1–7.4)	56/708 (7.9, 5.9–9.9)
8	19/630 (3.0, 1.7–4.4)	37/597 (6.2, 4.3–8.1)	47/562 (8.4, 6.1–10.7)	54/523 (10.3, 7.7–12.9)
9	29/498 (5.8, 3.8–7.9)	36/474 (7.6, 5.2–10.0)	41/449 (9.1, 6.5–11.8)	53/412 (12.9, 9.6–16.1)
P	0.008	0.018	0.060	0.026

 Anisometropia was defined as an SE difference of at least 1.0 D. Data are n (%, 95% CI). n = 1979.

Table 2.

 

Absolute Difference in Spherical Equivalence and Axial Length between Eyes

	Baseline Visit	First Visit	Second Visit	Third Visit
Spherical equivalence
Overall	0.29 (0.46) 0.18 (0–8.35)	0.33 (0.49) 0.20 (0–8.67)	0.39 (0.55) 0.23 (0–8.85)	0.44 (0.59) 0.25 (0–8.82)
Gender
Male	0.30 (0.52) 0.20 (0–8.35)	0.34 (0.56) 0.20 (0–8.67)	0.39 (0.59) 0.23 (0–8.85)	0.42 (0.59) 0.25 (0–8.82)
Female	0.27 (0.38) 0.18 (0–4.9)	0.32 (0.42) 0.20 (0–4.88)	0.39 (0.51) 0.23 (0–5.65)	0.45 (0.58) 0.25 (0–5.35)
P	0.202	0.291	0.897	0.294
Age at baseline (y)
7	0.25 (0.41) 0.18 (0–5.63)	0.28 (0.40) 0.18 (0–5.28)	0.35 (0.48) 0.20 (0–5.65)	0.40 (0.52) 0.25 (0–5.35)
8	0.29 (0.45) 0.20 (0–8.35)	0.34 (0.53) 0.20 (0–8.67)	0.40 (0.59) 0.22 (0–8.85)	0.43 (0.60) 0.25 (0–8.80)
9	0.34 (0.54) 0.20 (0–7.70)	0.40 (0.58) 0.25 (0–8.10)	0.45 (0.62) 0.30 (0–8.65)	0.49 (0.68) 0.30 (0–8.82)
P	0.003	< 0.001	0.006	0.051
Axial length
Overall	0.22 (0.24) 0.17 (0–3.35)	0.26 (0.26) 0.19 (0–3.48)	0.25 (0.26) 0.18 (0–3.50)	0.29 (0.28) 0.22 (0–3.70)
Gender
Male	0.22 (0.24) 0.17 (0–2.99)	0.26 (0.28) 0.19 (0–3.48)	0.25 (0.28) 018 (0–3.50)	0.29 (0.28) 0.22 (0–3.70)
Female	0.22 (0.23) 0.16 (0–3.35)	0.25 (0.24) 0.20 (0–2.07)	0.25 (0.25) 0.19 (0–2.30)	0.29 (0.28) 0.21 (0–2.15)
P	0.897	0.586	0.819	0.868
Age at baseline (y)
7	0.21 (0.24) 0.16 (0–3.35)	0.24 (0.23) 0.18 (0–2.14)	0.24 (0.24) 0.18 (0–2.23)	0.26 (0.24) 0.19 (0–2.15)
8	0.22 (0.22) 0.18 (0–2.99)	0.27 (0.28) 0.20 (0–3.48)	0.25 (0.27) 0.18 (0–3.39)	0.30 (0.30) 0.24 (0–3.70)
9	0.24 (0.25) 0.17 (0–2.77)	0.27 (0.28) 0.21 (0–3.33)	0.27 (0.29) 0.21 (0–3.50)	0.32 (0.31) 0.24 (0–3.10)
P	0.267	0.012	0.161	< 0.001

 Values are mean (SD), median (range). n = 1979.

Table 3.

 

Absolute Difference in J0 and J45 between Eyes

	Baseline Visit	Final Visit
Intereye J0 difference
Overall	0.16 (0.17) 0.11 (0–1.98)	0.17 (0.17) 0.12 (0–1.52)
Male	0.16 (0.16) 0.11 (0–1.22)	0.17 (0.16) 0.13 (0–1.05)
Female	0.15 (0.17) 0.11 (0–1.98)	0.17 (0.17) 0.12 (0–1.52)
p-value	0.794	0.862
7 y	0.16 (0.16) 0.11 (0–1.62)	0.16 (0.15) 0.12 (0–1.17)
8 y	0.16 (0.18) 0.11 (0–1.98)	0.16 (0.17) 0.11 (0–1.41)
9 y	0.15 (0.16) 0.11 (0–1.69)	0.18 (0.18) 0.13 (0–1.52)
P	0.896	0.307
Intereye J45 difference
Overall	0.20 (0.22) 0.14 (0–2.30)	0.21 (0.24) 0.14 (0–2.76)
Male	0.21 (0.24) 0.14 (0–2.15)	0.22 (0.27) 0.14 (0–2.76)
Female	0.19 (0.20) 0.14 (0–2.30)	0.21 (0.21) 0.14 (0–1.84)
P	0.084	0.221
7 y	0.21 (0.23) 0.14 (0–2.30)	0.20 (0.23) 0.14 (0–2.41)
8 y	0.20 (0.22) 0.14 (0–1.87)	0.21 (0.24) 0.15 (0–1.88)
9 y	0.19 (0.20) 0.15 (0–1.94)	0.22 (0.26) 0.14 (0–2.76)
P	0.471	0.400

 Data are mean diopters (SD), and median (range). n = 1979 at baseline; n = 1643 at the final visit.

Table 4.

 

Distribution of Anisometropia in Different Refractive Error Categories

	Baseline Visit	First Visit	Second Visit	Third Visit
Both eyes myopic	43/607	29/607	24/607	23/607
(Myopic anisometropia)	(7.1%)	(4.8%)	(4.0%)	(3.8%)
Both eyes hyperopic	5/181	5/181	3/181	3/181
(Hyperopic anisometropia)	(2.8%)	(2.8%)	(1.7%)	(1.7%)
1 eye myopic; 1 eye emmetropic	13/112	12/112	10/112	8/112
(Unilateral myopic anisometropia)	(11.6%)	(10.7%)	(8.9%)	(7.1%)
1 eye hyperopic; 1 eye emmetropic	10/119	8/119	8/119	6/119
(Unilateral hyperopic anisometropia)	(8.4%)	(6.7%)	(6.7%)	(6.7%)

 Anisometropia was defined as SE difference of at least 1.0 D. The denominator is the number of children at the baseline visit who had has the specified type of refractive error, with or without anisometropia. Myopia defined as SE at least −0.5 D and hyperopia defined as SE at least +1.0 D.

Table 5.

 

Changes in Refraction in the Study of Children with and without Anisometropia (n = 1568)

	Children without Anisometropia at Any Visit (SE Diff <1.0 D) (n = 1381)	Children with Anisometropia (SE Diff ≥ 1 D at Any Visit) (n = 187)	P *
Changes in refraction in eye with less ametropia (D/y)
Mean (SD)	−0.33 (0.26)	−0.44 (0.28)	<0.001
Median	−0.29	−0.44
Range	−1.25 to 0.14	−1.20 to 0.24
Changes in refraction in eye with worse ametropia D/y
Mean (SD)	−0.34 (0.26)	−0.43 (0.30)	<0.001
Median	−0.29	−0.41
Range	−1.25 to 0.11	−1.45 to 0.10

*  Mann-Whitney test

Table 6.

 

Changes in Spherical Equivalent and Axial Length in Children with Preexisting Anisometropia

	Initially Myopic in One Eye (n = 12)	Initially Myopic in Both Eyes (n = 35)	Initially Hyperopic in at Least One Eye (n = 12)	Overall (n = 59)
Beginning intereye difference in SE (absolute)
Mean (SD)	2.71 (2.60)	1.78 (0.95)	1.80 (0.98)	1.97 (1.46)
Range	1.03 to 8.35	1.0 to 4.9	1.02 to 4.55	1.0 to 8.35
Median	1.39	1.33	1.47	1.35
Final intereye difference in SE (absolute)
Mean (SD)	3.36 (2.91)	1.33 (0.71)	2.16 (1.21)	1.91 (1.68)
Range	0.28 to 8.82	0.16 to 2.7	0.22 to 4.53	0.16 to 8.82
Median	2.41	1.27	2.06	1.61
Rate of change of intereye SE difference (D/y)
Mean (SD)	0.16 (0.24)	−0.11 (0.16)	0.09 (0.32)	−0.02 (0.24)
Range	−0.21 to 0.55	−0.62 to 0.2	−0.54 to 0.68	−0.62 to 0.68
Median	0.16	−0.098	0.11	−0.056
Beginning intereye difference in AL
Mean (SD)	1.12 (0.91)	0.66 (0.45)	0.71 (0.43)	0.76 (0.59)
Range	0.26–2.99	0.0–2.16	0.26–1.59	0.0–2.99
Median	0.89	0.56	0.58	0.60
Final intereye difference in AL
Mean (SD)	1.36 (1.14)	0.48 (0.25)	0.68 (0.34)	0.70 (0.65)
Range	0.26–3.7	0.12–1.1	0.3–1.4	0.12–3.7
Median	0.95	0.5	0.54	0.54
Rate of change of intereye AL difference (mm/year)
Mean (SD)	0.60 (0.13)	−0.05 (0.10)	−0.01 (0.13)	−0.017 (0.12)
Range	−0.15 to 0.27	−0.37 to 0.15	−0.3 to 0.23	−0.37 to 0.27
Median	0.05	−0.03	0.02	−0.003
Rate of change of SE of the less ametropic eye (D/y)
Mean (SD)	−0.35 (0.27)	−0.59 (0.31)	−0.36 (0.31)	−0.50 (0.32)
Range	−0.96 to 0.04	−1.08 to 0.02	−0.87 to 0.01	−1.08 to 0.04
Median	−0.32	−0.62	−0.18	−0.45
Correlation: rate of change of SE difference vs. rate of change of intereye difference in axial length
r (P)	0.55 (0.063)	0.20 (0.249)	0.51 (0.110)	0.43 (0.001)*
Correlation: rate of change of SE difference vs. rate of change of SE of the less ametropic eye
r (P)	0.52 (0.080)	0.45 (0.007)*	−0.40 (0.199)	0.40 (0.002)*

  n = 59.

*  Statistically significant.

The authors thank Angela Cheng for the coordination and collection of data in this study and Ruthie McNeil for proofreading the revised manuscript.