Heritability of Peripheral Refraction in Chinese Children and Adolescents: The Guangzhou Twin Eye Study

Abstract

Purpose.: To estimate the heritability of peripheral refraction in Chinese children and adolescents.

Methods.: The authors examined 72 monozygotic (MZ) twins and 48 dizygotic (DZ) twins aged 8 to 20 years from a population-based twin registry. Temporal and nasal peripheral refraction, each 40° from the visual axis, and axial refraction were measured using an autorefractor. Relative peripheral refractive error (RPRE) was defined as the peripheral refraction minus the axial refraction. Heritability was assessed by structural equation modeling after adjustment for age and sex.

Results.: The mean and SD of temporal refraction (T₄₀), nasal refraction (N₄₀), RPRE-T₄₀, RPRE-N₄₀, and T₄₀-N₄₀ asymmetry were −0.27 ± 2.0 D, 0.36 ± 2.19 D, 1.18 ± 1.39 D, 1.80 ± 1.69 D, and −0.62 ± 1.58 D, respectively. The intraclass correlations for T₄₀ refraction, N₄₀ refraction, RPRE-T₄₀, RPRE-N₄₀, and T₄₀-N₄₀ asymmetry were 0.87, 0.83, 0.65, 0.74, and 0.58 for MZ pairs and 0.49, 0.42, 0.30, 0.41, and 0.32 for DZ pairs, respectively. A model with additive genetic and unique environmental effects was the most parsimonious, with heritability values estimated as 0.84, 0.76, 0.63, 0.70, and 0.55, respectively, for the peripheral refractive parameters.

Conclusions.: Additive genetic effects appear to explain most of the variance in peripheral refraction and relative peripheral refraction when adjusting for the effects of axial refraction.

Myopia is one of the leading causes of visual impairment among East Asian populations. ¹ Longitudinal and cross-sectional epidemiologic studies have suggested that the prevalence of myopia exceeds 70% among teenagers living in urban areas in East Asia. ^{2 –4} In recent population-based studies among adult Chinese living in urban and rural settings, myopic retinopathy has been the second-leading cause of blindness after cataract. ^5,6  

Despite intensive research in recent decades, the etiology of myopia remains elusive. Although on-axis refraction (central refractive error) is the major determinant of central visual acuity, there is increasing evidence suggesting that peripheral defocus also plays an important role in the development of myopia, ^{7 –12} although one study has reported contradictory results. ¹³ In keeping with this notion, previous studies have shown that subjects with relative hyperopic peripheral refraction are more likely to have a prolate posterior eye shape. ^{14 –16} By contrast, those with relative myopic peripheral refraction are more likely to have an oblate posterior eye shape. ^{14 –16} Interestingly, there appears to be ethnic variation in the distribution of peripheral refraction, with East Asians having a greater degree of relative peripheral hyperopia (more prolate ocular shape) than do persons of European descent with similar central refractive power. ^{17 –19} It remains unclear whether this ethnic difference is attributable to inherited genetic susceptibilities or socioenvironmental differences. It is therefore important to understand whether peripheral refraction is genetically determined. Such information may shed light on the mechanisms underlying the development of myopia. 

Twin studies offer a unique opportunity to estimate the relative contribution of genetic and environmental effects to the development of complex traits and diseases. ²⁰ In classic twin studies, it is assumed that monozygotic (MZ) twins share 100% of their genes, whereas dizygotic (DZ) twins share, on average, 50%. The heritability of a specific phenotype can be estimated by comparing the phenotypic concordance within MZ and DZ twin pairs. The purpose of this study was to estimate the distribution and heritability of peripheral refraction in young twins. 

Subjects, Materials, and Methods

Subjects

Subjects were recruited from participants in the Guangzhou Twin Study, an ongoing prospective cohort study of 9709 twin pairs that began in 2006. ²¹ Children older than 8 years of age who visited the Zhongshan Ophthalmic Center between August 1 and August 20, 2010, were invited to participate in the present study. Those with manifest strabismus, amblyopia, nystagmus, or any ocular disease causing best-corrected visual acuity less than 20/20 were excluded. In addition, subjects were excluded if examination with cyclopegia revealed a pupil diameter smaller than 6 mm or the pupillary light reflex was present. The study was conducted in accordance with the tenets of the World Medical Association's Declaration of Helsinki, and the Ethical Review Board of the Zhongshan Ophthalmic Center approved all procedures. Written informed consent was obtained from subjects or from their parents or legal guardians if participants were younger than 18 years of age. 

Examination and Measurement

Cycloplegia was induced by 2 drops of cyclopentolate 1% solution administered 5 minutes apart, followed by a third drop given 20 minutes later. The light reflex was evaluated and the pupil diameter was measured by an ophthalmologist with a ruler and a handheld light after an additional 15 minutes. 

Axial refraction and temporal (T₄₀) and nasal (N₄₀) refraction 40° from the visual axis were performed under cycloplegia by an ophthalmologist using an open-field autorefractor (NVision-K5001; Shin-Nippon Corporation, Tokyo, Japan) in a room with ambient illumination of 150 to 160 lux. Five measurements were taken at the corneal plane of the right eye at T₄₀, axial, and N₄₀, with the subjects fixating on a target located 3 m from the eye. For all measurements, the subjects were asked to maintain straight-ahead binocular fixation while turning their heads in the direction of the fixation target. Care was taken to ensure that subjects turned their heads precisely to the desired angle when fixating on the peripheral targets. 

Statistical Analysis and Genetic Modeling

The right eye was arbitrarily selected to represent the phenotypic characteristics of each subject in the data analysis. To adjust for the effects of axial refraction on peripheral refraction values, we calculated the relative peripheral refractive error (RPRE), which was defined as peripheral refraction minus axial refraction. RPRE-T₄₀ was computed as T₄₀ refraction minus axial refraction, and RPRE-N₄₀ was defined as N₄₀ refraction minus axial refraction. The T₄₀-N₄₀ asymmetry was defined as T₄₀ refraction minus N₄₀ refraction. 

The total phenotypic variance was decomposed into additive (A) or dominant (D) genetic variance and shared (C) or unique (E) environmental variance. The E component also contains measurement error. The correlations for additive and dominant genetic effects are defined as 1.0 in MZ pairs, whereas they are defined as 0.5 and 0.25, respectively, in DZ pairs. The correlation for shared environmental effects is defined as 1.0 and for unique environmental effects it is defined as 0 within both MZ and DZ pairs. As the C and D components confound each other when pairs of twins are reared together, the twin model allows only one of them to be included in a single model. Structural equation modeling was adopted to explain the data clearly and to obtain as few explanatory components as possible. ²² If the DZ pairwise correlation is less than half the correlation in MZ pairs, it suggests a contribution of genetic dominance. In this case, the ADE model is fitted as the full model; otherwise, the model is fitted with an ACE model. The best-fitting model was determined based on minus twice log-likelihood (−2LL) and χ² tests. A significant change in χ² values between the full model and the reduced model indicates that the parameter removed from the full model was significant and therefore should be retained. Alternatively, a nonsignificant change in χ² suggests that the parameter eliminated from the full model was not significant and therefore should be dropped. Because these refractive parameters were significantly associated with age and sex, the model treated these variables as covariates. This age- and sex-adjusted model is equivalent to using linear regression and adjusting for the effect of age as well as sex and then using the residuals to compute the variance-covariance matrices. Data management and preliminary analyses were carried out (Stata 8.0; Stata Corporation, College Station, TX). The Mx program was used for the fitting of genetic models. The level of statistical significance was set at P < 0.05. 

Results

A total of 120 pairs of twins (72 MZ and 48 DZ), ranging in age from 8 to 20 years, were available for this study. Among the MZ twins, 34 pairs were male and 38 pairs were female. Among the DZ twins, 6 pairs were male, 10 pairs were female, and 32 pairs were of mixed sex. The ages of MZ (14.3 ± 2.8 years) and DZ (13.5 ± 3.2 years) twins were not significantly different (t-test, P = 0.12). All subjects had pupil diameters >6 mm with absent light reflex after cycloplegia. 

Table 1 summarizes the distribution of peripheral refraction in 240 participants (112 male, 128 female). Among first-born twins, no significant differences were found between MZ and DZ individuals in any parameters, nor were there any differences between first- and second-born twins,. The distributions of axial and peripheral refraction in right eyes are shown in Figure 1. 

Table 1.

 

Phenotypic Characteristics of Twin Pairs by Zygosity

	T40 Refraction	N40 Refraction	RPRE-T40	RPRE-N40	T40-N40 Asymmetry
Monozygosity
First born	−0.50 ± 2.10	0.29 ± 2.38	0.91 ± 1.33	1.70 ± 1.90	−0.79 ± 1.65
Second born	−0.39 ± 1.85	0.20 ± 2.20	1.16 ± 2.34	1.75 ± 1.72	−0.59 ± 1.62
Dizygosity
First born	0.03 ± 1.89	0.81 ± 1.81	1.24 ± 1.39	2.02 ± 1.49	−0.79 ± 1.36
Second born	−0.01 ± 2.15	0.24 ± 2.23	1.57 ± 1.51	1.82 ± 1.54	−0.25 ± 1.63
Total
First born	−0.29 ± 2.03	0.50 ± 2.18	1.04 ± 1.36	1.83 ± 1.74	−0.79 ± 1.53
Second born	−0.24 ± 1.98	0.22 ± 2.20	1.32 ± 1.42	1.78 ± 1.64	−0.46 ± 1.62

 Values are mean ± SD and are expressed in diopters. T₄₀ refraction, temporal field refraction at 40° angle; N₄₀ refraction, nasal field refraction at 40° angle; RPRE-T₄₀, T₄₀ refraction minus the axial refraction; RPRE-N₄₀, N₄₀ refraction minus the axial refraction; T₄₀-N₄₀ asymmetry, T₄₀ refraction minus N₄₀ refraction.

Figure 1.

 

The distribution of axial, temporal, and nasal peripheral refraction.

The intraclass correlation coefficients (ICC) between MZ and DZ twin pairs for the various refraction outcomes are shown in Table 2. Within-pair correlations were significant (P < 0.05 for all) for both MZ and DZ twins, and were greater among MZ than DZ twins, for all these variables. This indicates substantial genetic influences on all these parameters. The scatterplots for refractive outcomes within MZ and DZ pairs are shown in Figure 2. 

Table 2.

 

Intraclass Correlation Coefficients (95% CI) for All Parameters

	T40 Refraction	N40 Refraction	RPRE-T40	RPRE-N40	T40-N40 Asymmetry
Monozygosity	0.87 (0.80–0.92)	0.83 (0.74–0.89)	0.65 (0.49–0.77)	0.74 (0.62–0.83)	0.58 (0.40–0.71)
Dizygosity	0.49 (0.24–0.68)	0.42 (0.16–0.63)	0.30 (0.02–0.54)	0.41 (0.15–0.62)	0.32 (0.04–0.55)
Total	0.71 (0.61–0.79)	0.68 (0.57–0.77)	0.51 (0.37–0.63)	0.63 (0.51–0.73)	0.48 (0.33–0.61)

 T₄₀ refraction, temporal field refraction at 40° angle; N₄₀ refraction, nasal field refraction at 40° angle; RPRE-T₄₀, T₄₀ refraction minus the axial refraction; RPRE-N₄₀, N₄₀ refraction minus the axial refraction; T₄₀-N₄₀ asymmetry, T₄₀ refraction minus N₄₀ refraction.

Figure 2.

 

Pairwise correlation for peripheral refraction in MZ and DZ twin pairs. (A) T₄₀ refraction, temporal field refraction at a 40° angle. (B) N₄₀ refraction, nasal field refraction at a 40° angle. (C) RPRE-T₄₀, T₄₀ refraction minus axial refraction. (D) RPRE-N₄₀, N₄₀ refraction minus axial refraction. (E) T₄₀-N₄₀ asymmetry, T₄₀ refraction minus N₄₀ refraction.

In maximum-likelihood modeling, the full model started from ACE for T₄₀ refraction, N₄₀ refraction, RPRE-N₄₀, and T₄₀-N₄₀ asymmetry because the ICC in DZ pairs was greater than half the ICC in MZ pairs. The full model started from ADE for RPRE-T₄₀ because the ICC in DZ pairs was smaller than half the ICC in MZ pairs. Statistical modeling suggested that the AE model (additive genes and unique environment) was the best fitting and most parsimonious. Common environmental (C) and dominant genetic (D) effects were dropped from the model without significant change (P > 0.05) (Table 3). 

Table 3.

 

Genetic and Environmental Effects Estimated By Age- and Sex-Adjusted Maximum-Likelihood Model

Variables/Models	A (95% CI)	C or D (95% CI)	E (95% CI)	−2LL	df	Δχ2	Δdf	P
T40 refraction, D
ACE	0.843 (0.480–0.892)	0.000 (0.000–0.356)	0.157 (0.108–0.232)	881.771	234
AE*	0.843 (0.768–0.892)	—	0.157 (0.108–0.232)	881.771	235	0.000	1	1
CE	—	0.665 (0.553–0.754)	0.335 (0.246–0.447)	908.157	235	26.386	1	<0.001
E	—	—	1.000 (1.000–1.000)	978.037	236	96.266	2	<0.001
N40 refraction, D
ACE	0.763 (0.439–0.836)	0.000 (0.000–0.305)	0.237 (0.164–0.344)	937.054	234
AE*	0.763 (0.656–0.836)	—	0.237 (0.164–0.344)	937.054	235	0.000	1	1
CE	—	0.595 (0.466–0.699)	0.405 (0.301–0.534)	955.091	235	18.038	1	<0.001
E	—	—	1.000 (1.000–1.000)	1007.107	236	70.053	2	<0.001
RPRE-T40, D
ADE	0.247 (0.000–0.736)	0.393 (0.000–0.747)	0.360 (0.251–0.518)	786.527	234
AE*	0.629 (0.468–0.744)	—	0.371 (0.256–0.531)	787.133	235	0.607	1	0.436
E	—	—	1.000 (1.000–1.000)	823.465	236	36.939	2	<0.001
RPRE-N40, D
ACE	0.505 (0.064–0.782)	0.193 (0.000–0.595)	0.302 (0.213–0.426)	865.752	234
AE*	0.700 (0.581–0.787)	—	0.300 (0.213–0.419)	866.231	235	0.479	1	0.489
CE	—	0.622 (0.498–0.721)	0.378 (0.279–0.503)	870.991	235	5.239	1	0.022
E	—	—	1.000 (1.000–1.000)	928.066	236	62.314	2	<0.001
T40-N40 asymmetry, D
ACE	0.507 (0.000–0.677)	0.041 (0.000–0.510)	0.452 (0.323–0.625)	864.816	234
AE*	0.550 (0.386–0.677)	—	0.450 (0.322–0.614)	864.837	235	0.021	1	0.884
CE	—	0.456 (0.303–0.586)	0.544 (0.414–0.697)	868.076	235	3.260	1	0.071
E	—	—	1.000 (1.000–1.000)	896.013	236	31.197	2	<0.001

 T₄₀ refraction, temporal field refraction at 40° angle; N₄₀ refraction, nasal field refraction at 40° angle; RPRE-T₄₀, T₄₀ refraction minus the axial refraction; RPRE-N₄₀, N₄₀ refraction minus the axial refraction; T₄₀-N₄₀ asymmetry, T₄₀ refraction minus N₄₀ refraction; A, additive genetic; D, dominant genetic; C, common environment; E, unique environment; −2LL, twice the negative log-likelihood; Δχ², difference of χ² values; Δdf, difference of degree of freedom; P, P of χ² test in model fitting.

*  Best-fitting model.

Additive genetic effects (A) explained 84.3% (95% confidence interval [CI], 76.8%–89.2%), 76.3% (95% CI, 65.6%–83.6%), 62.9% (95% CI, 46.8%–74.4.%), 70.0% (95% CI, 58.1%–78.7%), and 55.0% (95% CI, 38.6%–67.7%) of the variation in T₄₀ refraction, N₄₀ refraction, RPRE-T₄₀, RPRE-N₄₀, and T₄₀-N₄₀ asymmetry, respectively, with the remaining variance attributed to unshared environment (E) (Table 3). 

Discussion

To the best of our knowledge, this is the first classical twin study exploring the distribution and heritability of peripheral refraction. Peripheral and axial refraction appear to play different roles in the etiology of myopia. Peripheral refraction is an important determinant of the progression of myopia, based on evidence from animal and cross-sectional human studies. ^7,8,11 Our study results are consistent with a significant genetic contribution to peripheral refraction, with genetic effects accounting for more than three-quarters of the variation in peripheral refraction and more than half the variation in relative peripheral refraction. 

The distribution and pattern of peripheral refraction we found were similar to those of previous reports derived from East Asian populations. In our study, RPRE-T₄₀ and RPRE-N₄₀ were 1.18 ± 1.39 D and 1.80 ± 1.69 D, respectively, similar to the values of 0.91 ± 1.15 (nasal retina) and 1.68 ± 0.95 (temporal retina) reported in the Chen et al. study. ¹⁷ This is different from what has been found in European eyes. ^{14 –16}  

The high heritability of peripheral refraction observed in our study is consistent with estimates for axial refraction. ^23,24 In the Hammond et al. ²³ study in adult twins, the heritability of axial refraction was 0.86 in right eyes, and the best-fitting model was also the AE model. In the Dirani et al. ²⁴ twin study, the heritability of axial refraction was similar, at 0.88 in males and 0.75 in females. In our study, we found that the heritability of peripheral refraction was 0.84 and 0.76 for the temporal and nasal retina, respectively. We also observed that the heritability remained in the range of 0.6 to 0.7 for relative peripheral refraction, which adjusted for the effects of axial refraction. 

Our study also investigated temporal-nasal refractive asymmetry. Similar to other studies, we found that axial-nasal refractive differences were greater than axial-temporal disparities. ^16,17,19 Interestingly, the correlation of temporal-nasal asymmetry was higher among MZ than DZ twin pairs, suggesting that ocular peripheral refraction asymmetry may also be under genetic control to a considerable extent. In the best-fitting variance model, variance in temporal-nasal asymmetry was influenced primarily by additive genetic effects. 

Our study is the first to assess the distribution and heritability of peripheral refraction and its asymmetry. However, our conclusions should be interpreted in the context of the study's limitations: first, for practical reasons, we were only able to choose peripheral refraction at 40° as the outcome of interest and did not measure peripheral refraction at other angles. Therefore, the results may be valid only for this angle, although 40° has typically been used in other clinical studies of peripheral refraction. ^17,25 Second, our sample size was relatively small; therefore, we might have lacked the power to detect the impact of common environmental effects and to estimate additive genetic effects more precisely. 

In conclusion, our study demonstrates a high heritability of peripheral refraction and its nasal-temporal asymmetry. Given that peripheral refraction is recognized as an important feature in the onset and progression of myopia, our study findings are consistent with an important role of genetic factors in myopia. Future work is needed to identify the genes contributing to peripheral refraction and its development. 

Footnotes

 Supported by the Fundamental Research Funds of the State Key Laboratory, National Natural Science Foundation of China Grant 30772393, and Natural Science Foundation of Guangdong Province for Doctoral Scholars Grant 3030901005155.

Footnotes

 Disclosure: X. Ding, None; Z. Lin, None; Q. Huang, None; Y. Zheng, None; N. Congdon, None; M. He, None