Impact of Genetic and Environmental Factors on Peripheral Refraction

Abstract

Purpose: Investigate genetic and environmental influences on refractive errors in monozygotic (MZ) and dizygotic (DZ) twin pairs.

Methods: We assessed foveal and peripheral refractions in 54 MZ and 46 DZ twins, capturing three scans across the retina. The study focused on spherical equivalent (M) at the fovea (M_LOS) and changes in midperipheral (δM_{mid-periphery}), and peripheral (δM_periphery) defocus, along with nasal-temporal asymmetry (root mean squared error [RMSE_ASY]) and image shell contour (RMSE_AVG). Genetic and environmental contributions were analyzed using structural equation models.

Results: No significant differences were observed between MZ and DZ twins for the examined variables. Intraclass correlations (ICC) indicated an important difference in genetic influence between M_LOS, with the MZ twin pairs showing a higher correlation (0.83) than DZ (0.69) pairs, and δM_periphery, because the ICC for the MZ doubled (0.87) that of the DZ (0.42) pairs. Heritability estimates from the ACE model confirmed the large difference on genetic factors’ influence on the variance for M_LOS (0.13) and δM_periphery (0.77) change in refractive error. RMSE_ASY and RMSE_AVG metrics showed significant genetic impact, particularly pronounced in the peripheral measurements, revealing high genetic control.

Conclusions: The study delineates a marked environmental impact on central refractive errors, whereas genetic factors had a more significant influence on peripheral refractive variance and retinal image traits. Findings of the ACE model highlight the intricate genetic and environmental interplay in refractive error development, with a notable genetic dominance in peripheral vision characteristics. This suggests potential genetic targets for interventions in myopia management and emphasizes the need for personalized approaches based on genetic predispositions.

Translational Relevance: Understanding the impact of genetics and environment on peripheral refraction is essential for deepening our fundamental knowledge of myopia and guiding the development of advanced myopia control strategies.

Introduction

The onset of myopia is increasing drastically throughout the world and becoming an alarming threat to visual deprivation, ocular complications, and compromised socioeconomic status.^1,2 Visual information is crucial for the emmetropization process, which depends on the integration of environmental and genetic factors. This complex signaling cascade involves all ocular tissues and multiple pathways, converting visual stimuli into molecular signals that regulate eye growth.^3,4 A crucial aspect is that the retina detects optical defocus through various optical cues and generates specific molecular signals⁵ that either stimulate eye growth (induced hyperopic defocus) or inhibit it (induced myopic defocus). However, the exact processes by which these cues are detected remain not fully understood. In addition to the known optical cues that activate various molecular mechanisms (both sign-dependent and independent, such as brightness, circadian rhythm, chromatic cues, and peripheral defocus), contrast perception has emerged as a potential cue for emmetropization.⁶ Optical defocus causes a proportional degradation of contrast at the luminance edges of images projected onto the retina. The retina may use this contrast at the luminance edges to determine the focal plane and color contrast to identify the sign of defocus. This is possibly the underlying mechanism in the techniques for treating myopia progression as overnight corneal reshaping contact lenses (orthokeratology), soft contact lenses incorporating multifocal or aspheric optics or defocus incorporated multiple segments lenses.⁷ 

Relative peripheral refraction has been a widely discussed parameter in recent studies to understand the root cause of myopia progression and its prevention. Many studies have been conducted to understand the aberrations in the human visual system. However, the study of peripheral refraction came into the limelight when Rempt et al.⁸ and Hoogerheide et al.⁹ in 1971 used peripheral refraction profiles to predict future myopia incidence in young pilots. Many studies showed evidence in favor that myopic individuals have relative hyperopic defocus in the peripheral retina, whereas emmetropic and hyperopic individuals have relative myopic defocus in the peripheral retina.^10–15 Studies in animal models and human eyes support that peripheral hyperopic defocus may be responsible for retinal elongation or development of axial myopia.^16–18 An important contribution of peripheral retina on visual system development has been claimed with an animal study conducted by Smith et al.¹⁶ that showed emmetropization can continue even after central retinal laser ablation. This result showed a substantial contribution of the peripheral retina on the visual feedback mechanism causing emmetropization. However, some studies published contradictory reports on the influence of peripheral refraction on foveal axial length control.^19–22 Despite the debate on the potential importance of peripheral refraction in central myopia development and vice versa, some myopia control optical approaches rely on increasing peripheral refractive power to control myopia progression. 

Different approaches have been adopted in the recent past to investigate peripheral retinal alterations with myopia development. These approaches include longitudinal and cross-sectional cohort studies, model eye simulations, animal trials, and retinal contour mapping using magnetic resonance imaging (MRI), computed tomography scanning, and ocular coherence tomography (OCT).^23,24 Classical twin studies are a useful tool to analyze the genetic and environmental influences on a phenotypic trait.^25–27 The extent of genetic and environmental variance of a phenotype can be estimated by comparing the resemblance of the trait in monozygotic (MZ) or identical twins with the concordance of the same trait in dizygotic (DZ) or nonidentical twins.²⁵ A greater similarity between MZ twins compared to DZ twins can be attributed to the additional gene sharing, whereas a high correlation between DZ twins may indicate an important shared environmental effect. Several studies have estimated the heritability (i.e., the proportion of phenotypic variance caused by genetic factors) of refractive error at the fovea, with ample variability across different young and old populations.^28–37 Dirani et al.³⁵ showed a heritability of 88% in men and 75% in women from an Australian population across a wide age range of 18 to 88. Hammond et al.³⁴ found a heritability of 84% in a sample aged 49 to 79 of British population. A similar study in a Danish sample (age range 20–45 years) also revealed high refractive error heritability of 90%.³¹ However, others have found lower heritability estimates, especially for younger samples with larger myopia prevalence. A comprehensive table of all relevant studies from 1962 has been provided in our previous work.³⁸ These discrepancies in heritability estimates could be accounted for, in part, by the environmental changes taking place over the last two decades, which may be responsible for the drastic change in myopia prevalence during that period. Numerous environmental exposures such as excess near work,^39–41 increased indoor activity,⁴² increased level of education^39,43,44 and massive urbanization^45,46 showed a high correlation with the incidence of myopia. On the contrary, increased outdoor activity seems to be delaying the onset of myopia development during childhood.⁴⁷ Such changes could have induced higher environmental variability in myopia development, which would result in decreased heritability. Supporting this interpretation, the heritability of objective refraction in the sample used for the present study was considerably lower compared to a middle-aged twin population from the same geographical origin.³⁸ Analyzing younger samples of people born in the first decade of the current century could help shed light on this question. Moreover, most previous heritability studies were restricted to estimating the variance of on-axis refraction, showing a general trend of substantial heritability, although there is a scarcity of studies exploring the genetic and environmental influences on peripheral refraction. Sample characteristics such as age, years of schooling, and geographical area, combined with the modern lifestyle, may explain these inconsistencies. All previous twin studies have been performed only in the foveal refraction among children, middle-aged, or mixed-twin populations,^34–36,48 except for a study conducted in Chinese children.^49,50 In this twin study, central and peripheral refraction were measured in a group of healthy young adult twin subjects. The primary purpose was to determine to what extent the variance of central and peripheral refraction measurements could be attributable to genetic and environmental influences to increase our knowledge about the etiological architecture of refraction variability. 

Methods

This study was conducted at the Laboratory of Optics on a sample of young university students collected from the Murcia Twin Registry (Murcia, Spain). All subjects agreed with the research protocol and willingly signed written informed consent. The study was designed according to the tenets of the Declaration of Helsinki⁵¹ and the ethical approval was granted by the Research Ethics Committee of the University of Murcia (ID no. 1108/2015). Twin pairs were excluded in the presence of any active ocular pathology, corrected visual acuity below 0.9 or history of ocular surgical procedures. A summary of demographic and refractive data is shown in Table 1. Two hundred university students from the Murcia Twin Registry^52,53 were included in this study, including 54 MZ and 46 DZ pairs. There was a significantly higher female participants in both twin groups: 83% in MZ and 69% in DZ pairs. However, all our results were corrected for age and sex effects. A considerably high prevalence of myopia among the study participants was found, with 77% of students showing manifest myopia (≤−0.50 D). 

Table 1.

 

Subject Demographics and Distribution of Refractive Error

 

The open-view fast-scanning Hartmann-Shack (HS) peripheral wavefront sensor (VPR; Voptica SL, Murcia, Spain) was used to measure the peripheral wavefront profile. A 780 nm near-infrared laser light source (model S1FC780; Thorlabs, Newton, NJ, USA) was used in the instrument considering ocular conform with a minimal pupillary light response. Laser power (10 µW/cm²) was kept much lower than the permissible ocular safety limit standard for the selected wavelength. Further detail about the peripheral sensor can be found elsewhere.⁵⁴ The VPR includes a rotating camera that, starting at the line of sight, performs four consecutive scans. For each scan, 81 high-resolution HS images, corresponding to a 1° retinal area, are recorded to cover the fovea and a range of 40° in both nasal and temporal directions in just 1.8 seconds. This results in a total set of 324 HS images, which were later processed for a 4 mm pupil and up to fifth-order Zernike polynomials. For each angle, objective spherical equivalent (M; D) was calculated by averaging values from the four independent HS images. Although HS images were available for up to ±40°, only those up to ±35° were considered to avoid frequent noise in extreme peripheral measurements. Subjects were asked to fixate at a laser fixation point at 3 meters distance to control accommodation response. Figure 1 shows an example of the three different measurements recorded by moving vertically the fixation laser point accordingly: a central scan at primary gaze (central scan) and two peripheral scans along 20° superior and 20° inferior retina. 

Figure 1.

 

Example of the scans registered in one subject: primary-gaze scan across the central field of view (red), superior 20° (black), and inferior 20° (green). Error bars: 1 SD for each angle.

Refractive error at the fovea (M_LOS) was compared between twin groups. To compare the peripheral refractive error between subjects with different M_LOS values, relative values (δM) were calculated by subtracting M_LOS from the actual M value for each angle. Mid-peripheral value  (δM_{mid-periphery)} was calculated by averaging data of the central scan for an arc between 20° temporal and 11° nasal (n = 32 data points). The peripheral relative value  (δM_periphery) included the rest of the central scan data beyond 20° for both temporal and nasal sides, along with all data from both superior and inferior scans (n = 172 data points). For central scan measurements, data from 12° to 19°, corresponding generally to the optic disk area, were not considered in our analysis (Fig. 2 left). 

Figure 2.

 

Left: Example of RMSE_AVG value (yellow) obtained by doing the root-mean-square error of the difference of δM values between a reference profile (black line) and each individual refraction profile (red line); shaded box represents the optical disk area. Right: example of RMSE_ASY (cyan) obtained by doing the root-mean-square error of the difference between nasal and flipped-temporal profile of δM values (red line) for each subject.

We were also interested in checking the genetic and environmental influences over the defocus changes at the central scan. A spherical equivalent profile for the central scan was obtained using a mean δM profile obtained by averaging the whole set of data (black line, Fig. 2; N = 200 eyes). As could be seen in Figure 2, for each subject, a deviation from the mean (root mean squared error [RMSE_AVG]) value was obtained by doing the root-mean-square error between the actual δM (red line) compared to the average profile δM (black line). 

Possible asymmetries in the variation of δM throughout the horizontal meridian were also checked by means of a nasal-temporal refractive profile asymmetry value (RMSE_ASY) calculated, as shown in Figure 2 (right graph), by doing the root-mean-square error of the difference between temporal and nasal refractive error (cyan shaded area). For this parameter, nasal values within the optic disk area were interpolated from surrounding data points. 

Statistical Analysis

Descriptive analyses were performed using SPSS v28.0 (SPSS Inc., Chicago, IL, USA) statistical software. Normal distribution was checked using the Kolmogorov-Smirnov test. Differences between variables were obtained by means of the Student t-test for normally distributed and the U Mann-Whitney test for non-normally distributed variables. The significance level for P value was set at 0.05. The intraclass correlation coefficient was used instead of the Pearson correlation coefficient to avoid problems with twin data dependence when making the comparison between twins. Within a classical twin study, using structural equation modeling (SEM), the variance of any phenotypic trait can be decomposed into four latent factors: Additive (A) genetic variance represents the combined individual effects of alleles influencing a phenotype; Dominant (D) or non-additive genetic effects capture the variance due to interactions between genes, including dominance and, possibly, epistasis; Common (C) or shared environmental influences are those that are shared by the twins and act to make them similar to each other; and unique or non-shared Environment (E) impacts on each individual separately making twins in a pair different (it also includes measurement error). C and D cannot be estimated at the same time in a classical twin study because they are negatively correlated. Hence, the selection of a model including ACE or ADE components is based on the pattern of twin correlations. An ADE model is usually selected when the DZ correlation is lower than half of the MZ correlation. In contrast, an ACE model is selected if the DZ correlation is greater than half of the MZ twin correlation.^55,56 

In order to estimate the variance components (A, D/C and E) the data were analyzed by SEM, using the OpenMx package in R (R Core Team, Vienna, Austria).⁵⁷ Mean effects of age and sex were corrected by including them as covariates in all models, in order to avoid inflating twin estimates of shared environment, as it is standard procedure in twin modelling.⁵⁸ To be able to use all data from complete and incomplete pairs, Full Information Maximum Likelihood estimation with raw data was used. In this method, twice the negative log-likelihood (−2LL) of the data for each twin pair was calculated, and parameters were estimated so that the likelihood of the raw data was maximized. Nested models (AE, CE, E) were compared against the full model (ACE/ADE) with likelihood ratio tests, which were was obtained by subtracting −2LL for a restricted nested model from that of a less-restricted model (χ² = [−2LL₀] – [−2LL₁]). The resulting test statistic had a χ² -distribution with degrees of freedom (df) equal to the difference in df between the two models. When the fit of a more-restrictive (nested) model differs significantly from that of the less restrictive, it implies that the restriction imposed in the nested model does not hold for the available data. The best-fitting model was chosen in each case by deducting the residual deviance of the compared models and by comparing Akaike's information criterion. 

The power of the experimental design to detect heritability based on the current sample size was determined by testing full models (ADE/ACE) versus restricted nested models dropping the genetic components (A + D or A, according to the model) with a 2 or 1 df test, respectively, and alpha of 5%. The power to detect a broad sense heritability (A + D) of 0.5, 0.6, or 0.8 was 82%, 95%, and 99.9%, respectively, when the contribution of additive and nonadditive effects was equal. The power to detect a narrow sense heritability (A) of 0.4 was 49% when the contribution of additive genetic and shared-environmental effects was equal.⁵⁹ 

Results

In exploring the complex interplay between genetic predisposition and environmental factors in the development of myopia, this study analyzes the variance of refractive errors within a cohort of 54 monozygotic and 46 dizygotic twins. Given the structured assessment of both central and peripheral refractive errors, we postulate the influence of shared environments and genetic makeup on these phenotypes. Figure 3A shows the correlation between members of the twin pair for M_LOS was similar for both twin groups: 0.83 for MZ (95% confidence interval [CI], 0.73–0.90), and 0.69 for DZ (95% CI, 0.51–0.82) pairs, suggesting that variance in shared environmental influences might play a significant role for explaining the variance of foveal objective refraction. 

Figure 3.

 

Correlation of objective refraction at various retinal zones up to ±35° horizontal and ±20° vertical retinal eccentricity from the line of sight, comparing MZ (red), and DZ (blue) siblings: (A) line of sight (M_LOS), (B) relative mid-peripheral value (δM_{mid-periphery}) calculated by averaging data within an arc spanning from 20° temporal to 11° nasal retina, and (C) relative periphery (δM_periphery), incorporating data beyond 20° on both the temporal and nasal sides in the central scan, as well as all data from both superior and inferior scans. Plot A has a different scale because it is based on absolute refraction values, whereas plots B and C share the same scale, which is relative to the line of sight. Schematic yellow lines on the fundus images indicate the analyzed retinal areas (not to scale).

The ACE model for M_LOS showed a limited effect of additive genetic factors on this phenotype (A = 0.13; 95% CI, 0–0.46; C = 0.67, 95% CI, 0.35–0.83; E = 0.20; 95% CI, 0.13–0.30). Hence, the best fit for this trait was provided by a CE model (C = 0.78, 95% CI, 0.68–0.84; E = 0.22, 95% CI, 0.16–0.32; Table 2). 

Table 2.

 

Structural Equation Model Fitting for M_LOS, δM_{mid-periphery}, and δM_periphery

 

Average values of δM_{mid-periphery} were small for both MZ (−0.13 ± 0.26 D) and DZ (−0.11 ± 0.18 D) twin groups, indicating little change in refractive error in the vicinity of the macular area. Intrapair correlations (Fig. 3B) were modest: 0.51 for MZ (95% CI, 0.27–0.69), and 0.32 for DZ (95% CI, 0.03–0.56) twin pairs. The SEM analysis of this phenotype (Table 2) showed that the additive genetic component in the ACE model was smaller (A = 0.22) than environmental effects (C = 0.24; E = 0.54). The CE model was again the best fitting model (C = 0.42; E = 0.58). All models for this phenotype showed a large effect of nonshared environmental factors. 

Average (δM_periphery) values were much higher for both MZ (−1.64 ± 0.85 D) and DZ (−1.58 ± 0.77 D) twins. Intraclass correlation coefficient values (Fig. 3C), being more than twice for MZ (0.87; 95% CI, 0.79–0.92) than for DZ (0.42; 95% CI, 0.15–0.63) pairs, indicating a high influence of genetic factors on the variance of this phenotype. This was confirmed by SEM models (Table 2): in the ACE model, the additive genetic component explained most of the variance of this phenotype (A = 0.76), with small contribution of environmental effects (C = 0.08; E = 0.16). The nested AE model provided the best fit (A = 0.84; E = 0.16). Thus variance of difference in objective refractive error between fovea and peripheral retina (δM_periphery) was mainly driven by genetic factors. 

The peripheral image shell analysis (RMSE_AVG) showed similar average values for both MZ (0.65 ± 0.46 D) and DZ (0.62 ± 0.31 D) twin pairs. Despite having similar image shell contour, the correlation between members of the twin pair (Fig. 4) was two-fold higher for MZ (0.52; 95% CI, 0.29–0.69) than for DZ (0.23; 95% CI, 0–0.41) twins, suggesting a genetic impact. The best fit was provided by an AE model for this phenotype (Table 3), with additive genetic (A) and non-shared environmental (E) factors, accounting each for approximately half of its variance. 

Figure 4.

 

Correlation of peripheral image shell analysis (RMSE_AVG) between MZ (red) and DZ (blue) siblings.

Table 3.

 

Structural Equation Models for the Difference Between the RMSE_AVG and RMSE_ASY

 

Results for peripheral asymmetry phenotype (RMSE_ASY; Fig. 5) showed similar average values for both MZ (0.77 ± 0.51 D) and DZ (0.78 ± 0.50 D) twin groups, although it showed a fourfold higher correlation between MZ (0.65; 95% CI, 0.47–0.78) than between DZ (0.16; 95% CI, 0–0.43) pairs. 

Figure 5.

 

Correlation of peripheral asymmetry phenotype (RMSE_ASY) between MZ (red) and DZ (blue) siblings.

The large difference in correlation between twin groups confirms the importance of additive genetic factors on the variance of this phenotype, as shown by SEM analysis, as shown in Table 3. An AE model provided the best fit for this phenotype, where additive genetic accounted for 61% of its variance, and nonshared environments and errors accounted for nearly 39% of the variance. 

Discussion

Phenotypic analysis of the objective refraction over a wide retinal area was performed in a group of 100 young twin pairs with similar age and manifest refraction, for a wide range of retinal eccentricities, using data collected from three different measures. We found a high myopia prevalence (77%) and higher shared visual exposures at the central retina in our study sample, which most likely include myopigenic factors connected to modern lifestyle and massive urbanization, as suggested by recent myopia studies.^{42–46,60–62} However, the environmental contribution to the variance of objective refraction at the fovea (M_LOS) within our sample of millennials opposes most previous literature done in earlier generations from different ethnicities.^{23–25,28–31} This sudden increase in the environmental impact on foveal refraction is directly supported by our previous work, which demonstrated a high heritability for foveal refractive error in an older twin sample from the same geographical location as the present study.³⁷ As our study population is from an urban area of a developed country, the university students were certainly exposed to these myopia causative visual exposures. The probable cause of higher myopia prevalence among our study sample was further explained in our previous article.³⁸ 

Interestingly, despite the lower heritability of foveal refraction, we found strong additive genetic control over the variance of relative peripheral refractive error (δM_periphery). In the mid-peripheral retinal zone, we observed that the variance in relative refraction (δM_{mid-periphery}) suggested an exaggerated influence from unique environmental factors (44%). This is likely because the variation within individuals was greater than the average value of the phenotype being tested, indicating a reduced impact from genetic factors. Previous work, where the relative weight of genetic and environmental factors was estimated at every degree of the horizontal scan up to ±35°, showed that whereas the phenotypic variance in the fovea and the central retina (up to around ±20°) was mainly explained by environmental factors (CE), the best fit for values in the nasal or temporal retina was provided by an AE model, where a substantial proportion of the variance was explained by genetic factors.⁶³ There is a significant scarcity of relevant literature about heritability of relative objective refraction across a wide retinal eccentricity. The only relevant work found is the study by Ding et al.,⁶⁴ conducted in a Chinese twin sample of age group between 8 to 20 years, showed similar findings as the additive genetic component plays a significant role in explaining the peripheral refraction variation. However, their study methodology did not include an assessment of refraction inheritance in mid-peripheral, superior, and inferior retinal areas. Consequently, our work suggests that the impact of environmental factors tends to concentrate in the central retina, producing changes in eyeball structure because of its functional adaptation to environmental needs. In contrast, eccentric zones of the eyeball are likely less sensitive to environmental effects and more dependent on genetic architecture.⁶³ 

Peripheral defocus has usually been linked to myopia in previous studies. Recent studies suggest that the relative peripheral hyperopic defocus in myopia is a consequence of central myopia development. Myopia control optical approaches, such as orthokeratology, rely on producing myopic defocus as a treatment for myopia progression. However, the required magnitude of myopic defocus and the most responsive treatment region on the retina remain questionable. 

Studies have shown that the axial growth of the eyeball in myopia is unlikely to be confined solely to the macular area but extends to an unspecified region of the posterior ocular wall, often resulting in a contralateral distribution of retinal contour.⁶⁵ Longitudinal and cross-sectional cohort studies have been evaluated to investigate the peripheral retinal alteration with myopia development using different imaging devices such as MRI and OCT.^23,24 We used an alternative approach to study the peripheral image shape contour (RMSE_AVG) between MZ and DZ siblings. We found that, despite having similar averaged image shell contours, the variance of RMSE_AVG showed twice as high a correlation among MZ twins. However, genetic modeling indicated an “AE” model fit, with half of the variance controlled by additive genetics and the other half by unique environmental factors. 

We further analyzed peripheral refraction asymmetry by performing a peripheral image shell analysis to understand the inheritance of phenotypic variance in nasal-temporal shape asymmetry (RMSE_ASY) between the eyes of twin siblings who share both genetic inheritance and environmental exposures. We found a higher correlation in MZ twins compared to DZ twins, which was later supported by an AE genetic model fit (Table 3), where half of the variance was explained by additive genetics. Despite having similar average asymmetry values for both twin groups (RMSE_ASY: 0.77 ± 0.51 for MZ twins and 0.78 ± 0.50 for DZ twins), the correlation was much higher between MZ siblings, and the variance of RMSE_ASY was mostly influenced by variance in genes. Peripheral refraction asymmetry may be linked to myopia development, leading to unequal elongation of the eyeball. Refraction across retinal eccentricities is commonly asymmetric, with nasal-temporal asymmetry associated with extended axial length⁶⁶ and higher astigmatism.⁶⁷ A recent study in the Indian myopic population found an asymmetric type of peripheral refraction with relative hyperopic defocus in temporal retina and myopic defocus in the nasal retina.⁶⁸ An animal study on marmosets also showed asymmetries in the peripheral refraction after myopia development, after triggered by compensating mechanism by imposing a full-field defocus.⁶⁷ 

Our study presents important strengths that contribute to our knowledge regarding the progress of myopia and the development of myopia control strategies. It is based on a genetically informative sample, allowing for the analysis of the relative impact of genetic and environmental factors in the studied traits. Also, the selected sample of millennials is a window into the processes that are driving the prevalence of myopia in the population to pandemic levels. Additionally, it makes use of cutting-edge technology for measuring peripheral refraction at separate points of the eyeball, revealing the possibility of different architectures in refractive error development. In this study, we were limited by sample size, which resulted in large confidence intervals for some of our estimates, that hindered the possibility of clearly differentiating between components of genetic variance. Although we detected a fourfold increase in the correlation of the asymmetry metric used in our study sample, we could not precisely quantify the contribution of dominant genetic impact, likely because of the limited sample size affecting precision and producing very large confidence intervals when both additive and nonadditive genetic factors were included in the model at the same time. Obviously, a much larger sample could have increased the precision of the estimates and boosted our ability to perform additional analyses comparing refractive error groups. Nonetheless, the significance of the results regarding the influence of relative genetic and environmental factors remains unchanged. Moreover, a detailed questionnaire covering parental and personal history of our study subjects could have provided additional information and insights into their impact on refractive error, especially in extreme cases. 

Finally, the classical twin design used in this study is a powerful method for disentangling the genetic and environmental architecture of outcome phenotypes and provides useful estimates about the distribution of variance between these broad categories. However, it offers limited information about the distinct effects of specific exposure variables (whether genetic variants or environmental conditions). A different, complementary design would be needed to further explore these aspects. 

Conclusions

We have analyzed the phenotypic variance of refraction across different retinal zones in a sample of 100 pairs of young university twins. In our sample of high myopic prevalence, the foveal objective refraction variance was mainly linked to shared environmental influences. Whereas the variance of peripheral refraction was mainly driven by genetic factors. Moreover, the naso-temporal refraction asymmetry variance also showed a genetic dependency. These results suggest that the variance of peripheral defocus remains largely influenced by our genes, even when the subject develops under myopigenic influences. 

Acknowledgments

Supported by MyFUN, EU-ITN 675137, Agencia Estatal de Investigación, Spain (PID2019-105684RB-I00/AEI/10.13039/501100011033). MyFUN receives funding from the European Union's Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie grant agreement No 675137 

Disclosure: D. Pusti, None; A. Benito, None; J.J. Madrid-Valero, None; J.R. Ordoñana, None; P. Artal, None