Subfoveal scleral thickness is associated with peripheral retinal changes in high myopia in children and adolescents

This study aims to identify the risk factors in peripheral retinal changes (PRC) associated with high myopes among children and adolescents. This is a cross-sectional study on children and adolescents diagnosed with high myopia. The subjects involved underwent a series of ocular examinations, including the dilated fundus examination for PRC and the swept-source optical coherence tomography for foveal retinal, choroidal and scleral thickness measurement. Then, the variables were compared among the eyes with high risk, low risk, and no PRC. Spearman correlation was applied to evaluate the relationship between the parameters and the extent of PRC. Logistic regression was performed to identify the potential risk factors. A total of 117 eyes from 117 subjects were recruited. The prevalence of PRC was 57.3% (67 eyes), while that of high-risk PRC was 22.2% (26 eyes). Significant differences were found in the mean subfoveal scleral thickness, spherical equivalent refraction, and axial length among the eyes with high-risk, low-risk, and no PRC (p < 0.01, p < 0.01, p = 0.048, respectively). Compared with spherical equivalent (r = 0.32, p < 0.01) and axial length (r = 0.18, p = 0.05), subfoveal scleral thickness exhibited higher correlation coefficient with PRC (r =  − 0.38, p < 0.01). Subfoveal scleral thickness and spherical equivalent refraction were identified as the independent risk factors for PRC and high-risk PRC. It was demonstrated that there was a correlation between subfoveal scleral thickness and PRC. The eyes with thinner subfoveal scleral thickness carried a higher risk of PRC.


Introduction
At present, myopia is known as the most common refractive error worldwide [1]. It is estimated that the number of myopia will reach 4.8 billion by 2050, with nearly one billion being highly myopic [2]. High myopia is not only a refractive error. Instead, it carries an increased risk of complications, including cataracts, glaucoma, and chorioretinopathy. In some cases, they may even occur in youngsters, for example, retinal detachment, which may lead to vision impairment [3,4].
Peripheral retinal change (PRC) is a group of disorders occurring at the peripheral retina. It is commonly found in highly myopic eyes. It was once reported that as high as 61.7% of the adolescents with highly myopic eyes suffered PRC in Hong Kong [5]. Some types of PRC, such as lattice degeneration (LD) and retinal breaks, are at high risk for retinal detachment [6][7][8]. It was previously reported that the prevalence of LD reached 16.9% in highly myopic Singaporeans [9] and 3.3% among the highly myopic adolescents in Hong Kong [5]. Given the high prevalence of PRC in high myopes, understanding the risk factors in the development of PRC would be conducive to identifying these high-risk individuals and taking measures to prevent the occurrence of retinal detachment for them.
Currently, the sclera is suggested to play a crucial role in the progression of myopia. In histological studies, it has been found out that scleral thinning can occur to the posterior pole of highly myopic eyes in both human beings and animals, and posterior scleral thinning is considered as contributory to axial length elongation [10,11]. In general, the retinal changes in high myopia are believed to result from the mechanical elongation of the scleral caused by the excessive axial elongation of the eyeball. In several studies, it has been demonstrated that the increased prevalence of PRC was associated with the rise of either axial length (AL) or spherical equivalent refraction (SER) [5,9,[12][13][14]. Therefore, it is speculated in this study that the thinning of the posterior scleral causes the increase in axial length and elongation of the retinal, thus leading to peripheral retinal changes. Sweptsource optical coherence tomography (SS-OCT) could be applied to achieve an in vivo visualization of the sclera [15][16][17]. Up to now, there have been plenty of studies investigating the scleral morphological changes in myopic eyes with SS-OCT [15, [18][19][20]. However, none of them focus attention on the correlation between sclera and PRC.
This study is purposed to identify the risk factors in PRC among children and adolescents with high myopia, including subfoveal scleral thickness (SST), subfoveal choroidal thickness (SCT), foveal retinal thickness (FRT) and other ocular parameters.

Subject recruitment
This cross-sectional study was conducted in Joint Shantou International Eye Center, Shantou University and Hongkong Chinese University, from Nov 2018 to Apr 2020. The children and adolescents who were highly myopic and aged from 7 to 17 were recruited. High myopia was defined as SER (spherical power ? 1/2*cylindrical power) of -5.00D or higher [21]. The subjects with either past or present severe ocular diseases, such as manifest deviation, ptosis, glaucoma, cataracts, other retinopathy except for PRC, media opacities or a medical history of ocular surgery were excluded from this study. In addition, the subjects unfit for the fundus examination were excluded as well. For the subjects with bilateral high myopia, only the eyes with higher SER were included in this study.
In line with the tenets of the Declaration of Helsinki, the study was granted approval by the Institutional Review Board. Written consent was obtained from their parents in advance.

Data collection and ocular examination
The data used in this study, including age, gender, ethnicity, ocular history, systemic diseases, and the family history of high myopia, was recorded in detail. Besides, their heights and weights were also measured. All of the participants underwent a range of comprehensive ocular examinations, including uncorrected and best-corrected visual acuity (BCVA), intraocular pressure (IOP) measured by noncontact tonometer (TX-F, Canon, Japan), slit-lamp examination (BM900, Haag-Streit AG, Koeniz, Switzerland), axial length and corneal curvature measurement, refraction, fundus examination and SS-OCT.
Corneal curvature, axial length and refractive error assessment Axial length and corneal curvature were measured before cycloplegia using ocular biometry (IOL Master 500, ZEISS, German). Refractive error was determined by an autorefractor machine after cycloplegia (KR-8 Topcon, Japan). Cycloplegia was induced by adding four drops of 0.5% Tropicamide, with a 5-min interval [22]. The pupillary light reflex and pupil size were examined at least 20 min after the last drop of Tropicamide was administered. The criterion of cycloplegia was determined as the absence of a pupillary light reflex and a pupil size of at least 6 mm. Subsequently, subjective cycloplegic refraction tests were conducted by the trained, certified optometrists to achieve BCVA, with the results of subjective refraction taken as the final refraction for analysis.
Peripheral retinal changes assessment Dilated fundus examination was performed with binocular indirect ophthalmoscopy by an ophthalmologist (WZ). Then, widefield laser scanning ophthalmoscope (Daytona P200T, Optomap, German) imaging of the central fundus and the orientations of PRC were performed. Peripheral retinal changes were recorded and confirmed by the second ophthalmologist (TS) before being classified as high-risk or lowrisk PRC, depending on the predisposition of retinal detachment. High-risk PRC included LD, snail-track degeneration, retinal hole or tear, while low-risk PRC included white without pressure (WWOP), snowflake degeneration, microcystoid degeneration, peripheral pigmentary degeneration [8,[23][24][25][26].

SS-OCT examination and measurements
All the eyes were examined by SS-OCT (DRI OCT-1 Atlantis, Topcon, Japan) according to the 12 mm line scan protocol with an average of 128 consecutive, overlapping single B-scan OCT images. Besides, the line scan was made horizontal. An image quality score ranging from 0 to 100 was given by software for each volumetric OCT scan, with any scores below 60 excluded. There are also other significant image artifacts excluded from analysis, including(1) motion artifacts, (2) blur affecting 20% or more of the image (e.g., due to tilted images, defocus, or axial movement), (3) signal loss (e.g., due to eye blinking), or (4) poor centration (i.e., fovea displaced from the center). After cycloplegia, the same examiner performed all the SS-OCT examinations. We recognized the fovea to be at the center of the vessel-free area while scanning the images. The fovea in the OCT images was defined as the area where the inner retinal layers (the nerve fibre layer, ganglion cell layer, inner plexiform layer and inner nuclear layer) were absent. FRT, SCT and SST were measured manually by a trained grader without knowledge of information about the participants. Retinal thickness was defined as the vertical distance between the internal limiting membrane and the outer border of the retinal pigment epithelium (RPE), choroidal thickness was defined as the vertical distance between the outer border of the RPE and the choroidal-sclera interface, while scleral thickness was defined as the vertical distance between the choroidalsclera interface and the outer scleral border, as shown in Fig. 1. To determine the repeatability and reproducibility of SST measurement, a total of 23 OCT images randomly selected were measured by two ophthalmologists independently and twice by an ophthalmologist.

Statistical analysis
All statistical analyses were conducted using SPSS version 19 (SPSS, Chicago, IL, USA). Bland-Altman analyses were carried out to assess the intraobserver and the interobserver consistency in SST, SCT and FRT measurement. One-way analysis of variance (ANOVA) and Chi-square was conducted to analyze continuous and categorical variables, respectively. Bonferroni method was adopted for the post hoc test. The correlation between FRT, SCT, SST with AL and SER, as well as the correlation between PRC with AL, SER and SST were analyzed using Spearman correlation. Stepwise logistic regression was also performed for the potential risk factors of PRC and highrisk PRC. The level of p \ 0.05 was considered statistically significant.

Baseline characteristics
Initially, 145 participants met the inclusion criteria. Then seven participants were excluded because of their inability to cooperate with the examinations. Besides, another twenty-one were excluded for the outer scleral border beyond recognition. Finally, 117 eyes of 117 participants were included in the data analysis. Among them, 52 were right eyes (44.4%), and 65 were left (55.6%). There were 56 boys (47.9%) and 61 girls (52.1%). All of the participants were Chinese. The baseline characteristics of these eyes are shown in Table 1. The mean age was 11.4 ± 3 years. Figure 2 shows that the intraobserver and interobserver agreement of SST, SCT, FRT measurements were all good. The outer borders of the sclera drawn manually were verified as reliable. Comparisons among eyes with high-risk, low-risk and without PRC Table 1 shows the mean values of parameters in the eyes with high-risk PRC, eyes with low-risk PRC and the eyes without PRC, respectively. The differences in mean SER (-10.6 ± 3.6, -9.0 ± 2.6, -8.2 ± 2.2), AL (27.22 ± 1.74, 26.86 ± 1.33, 26.47 ± 1.09), SST (328.1 ± 45.1, 337.2 ± 42.9, 375.3 ± 52.0) among the three groups were statistically significant (one-way ANOVA, p \ 0.01, p = 0.048, p \ 0.01, respectively). Although SCT in the eyes of high-risk PRC and low-risk PRC was thinner than those without PRC, the difference was not statistically significant. There was no significant difference observed for age, gender, height, weight, BMI, the family history of high myopia, IOP, BCVA, corneal curvature, and FRT among the three groups (one-way ANOVA, p [ 0.05).
The results of the post-hoc analysis are presented in Table 2. The eyes with high-risk PRC and those with low-risk PRC had significantly thinner sclera than the eyes without PRC (328.1, 337.2 V.S 375.3, both p = 0.001). No significant difference was observed between the high-risk PRC group and the low-risk PRC group in SST or AL. The eyes with high-risk PRC Fig. 1 A SS-OCT image of a highly myopic eye, depicting the measurement of foveal retinal thickness, subfoveal choroidal thickness and subfoveal scleral thickness showed significantly higher SER and longer AL than without PRC (p \ 0.001 and p = 0.04, respectively). In contrast, there was no significant difference found in SER and AL between the eyes with low-risk PRC and the eyes without PRC. However, a significant difference was observed in SER between the high-risk and low-risk PRC groups (p = 0.02).
SCT exhibited moderate correlation with both SER (r = 0.59, p \ 0.01) and AL (r = -0.624, p \ 0.01). While there was only a weak correlation between SST and AL (r = -0.19, p = 0.04). There was no correlation observed between FRT and SER or AL.
According to stepwise logistic regression, SST and SER were independently and significantly associated with both PRC and high-risk PRC. Every 1 lm increase in SST contributed to 1.8% and 1% reduction to the risk of PRC and high-risk PRC, respectively. Additionally, every one diopter increase in SER was associated with 16.6% and 21.1% decrease in the risk of PRC and high-risk PRC, respectively, as shown in Table 3.

Discussion
To our knowledge, this is the first study exploring the correlation between PRC and foveal retinal, choroidal, scleral thickness in both children and adolescents. The eyes with PRC were found to have significantly thinner SST, higher SER and longer AL than without PRC. In contrast, SCT and FRT were identified as irrelevant to PRC. The correlations between PRC and SER, AL and SST were confirmed in univariate analyses. In stepwise logistic regression, SST and SER were identified as the independent risk factors for both PRC and high-risk PRC.
Different from most published studies focusing on the adult population, the study conducted on a young adult population in Singapore revealed that the prevalence of PRC in high myopes was 67.1%, and the most frequent PRC was WWOP (57.2%), followed by LD (16.9%) and retinal hole (4.4%) [9]. Lai et.al. reported pigmentary peripheral retinal degeneration, WWOP, LD and retinal break in 37.7%, 21.1%, 13.2% and 6.2% highly myopic eyes, respectively, in an adult  [13]. In another study on an adolescent population in Hongkong, it was found out that the four most prevalent findings were WWOP (51.7%), microcystoid degeneration (5%), peripheral pigmentary degeneration (4.2%) and LD (3.3%) with PRC prevalence of 61.7%. Compared with prior literature, the frequency of PRC (57.3%) in our study was lower than in other reports, which may be attributed to the younger age of our study population compared with previous studies. The prevalence of PRC increases with the progression of myopia over time. Likewise, this also explains the lower prevalence reported that an AL of 26.5 mm or longer was the significant risk factor for PRC, while it was not the case for SER [5], which is consistent with the results of a study conducted in Singapore [9]. In the research of Zhang et al., PRC was significantly associated with a longer AL and a higher SER [27]. In contrast, Celorio et al. reported that the peak prevalence of LD was in the eyes with a range of AL from 26.0 mm to 26.9 mm, which suggests an inverse relationship between the prevalence of LD and AL when the AL [ 27.0 mm [14]. The conflicting results of these studies might be attributable to the differences in the age spectrum of the study population. In our study, higher SER and longer AL were associated with the increase in PRC in univariate analyses. Differently, SER rather than AL was the independent risk factor in stepwise logistic regression. So far, the underlying mechanism of PRC in patients with high myopia is still not fully understood. A widely accepted view is that the mechanical axial elongation of the eyeball may be contributory to the development of PRC. Starting at the equator, the sclera thinning is most marked at the posterior pole owing to the axial elongation occurring in the course of myopization [28]. It is speculated that scleral thinning results from the mechanical axial elongation since the scleral volume showed no obvious increase during the process of myopization, which is accompanied by the remodeling of existing scleral tissue [29]. Our study found out that the eyes with PRC had a thinner SST than the eyes without PRC. The correlation between PRC and SST was stronger than that between PRC and SER or AL. In stepwise logistic regression, SST was the risk factor associated with both PRC and high-risk PRC, which has not been reported yet. Considering the findings of our study and others, it is speculated that the equatorial expansion resulting from axial elongation produces mechanical tractional forces on a continued basis for the peripheral retina, thus promoting the development of PRC.
While choroidal thinning potentially results in the reduction of choroidal perfusion and ischemia, which may be related to PRC. The retina will suffer damage when choroidal perfusion is reduced. Nonetheless, SCT in eyes with PRC was thinner than without PRC, despite no statistical significance. The findings of our study are consistent with that of Chen et al. [9]. As we know, the choroidal blood vessels become increasingly thinner from the posterior pole to the periphery [30], which may make the peripheral blood vessels less able to withstand mechanically tractional forces. Thus, it is necessary to conduct a further study on peripheral choroidal thickness for confirming our speculation. In our study, the FRT was shown not to  be significantly different among the three groups. It was reported that the occurrence of retinal thinning in the equatorial and retroequatorial region was attributed to a tube-like enlargement of the eyeball, and retinal thickness in the macular region was irrelevant to axial length [28], which may explain the lack of association between FRT and PRC. It should be emphasized that choroidal thickness of all subjects was measured after cycloplegia in our study. Choroidal thickness might be modified by cycloplegia reported by some studies, although the results were inconsistent, one of them found a decrease in choroidal thickness with cycloplegia [31]; another demonstrated an increase [32]. There are several limitations in our study. Firstly, only the children with high myopes were recruited, and the representation of AL and SST was restricted, which might be related to the weak correlation between SST and AL. Secondly, the exclusion of subjects with unrecognizable outer scleral border may lead to the underestimate of mean SST, especially in the eyes without PRC. Thirdly, this is a cross-sectional study, as a result of which causality cannot be confirmed. Lastly, in this study, we did not measure systemic parameters, such as blood pressure or circadian rhythm which may affect the measurements of choroidal thickness.
In conclusion, SST is an independent risk factor for peripheral retinal changes. High myopes with thinning sclera may require detailed fundus examination, particularly in the peripheral retina. The SS-OCT measurement of subfoveal scleral thickness in high myopes would assist the identification of people at risk of peripheral retinal changes.