Interocular Symmetry in Retinal Nerve Fiber Layer Thickness in Children: The Hong Kong Children Eye Study

The Hong Kong Children Eye Study (HKCES) is a population-based cross-sectional study targeting primary school children from 6 to 8 years old in Hong Kong who were invited to visit the Hong Kong Eye Centre for comprehensive ophthalmological examinations [10‐16]. It aims to evaluate various eye conditions among children in this age group and to determine the prevalence of different eye diseases. The sample selection was based on a stratified and clustered randomized sampling frame. Further details on sample size calculation and the study protocol were mentioned in previous reports. The participants were invited to attend in comprehensive ophthalmic examinations, physical examinations, and standardized interviews [10]. This study excluded children with congenital malformations, previous ocular trauma, a history of ocular surgery, or any ocular disorders other than refractive errors. Additionally, children with systemic diseases, except for obesity, or those who were unable to cooperate were also excluded from the study. The project adhered strictly to the Declaration of Helsinki, and approval was obtained from the Ethics Committee Board of the Chinese University of Hong Kong. Participating children and their parents gave informed consent prior to their participation in the study. This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

OCT imaging was performed with the Spectralis SD-OCT (Heidelberg Engineering, Heidelberg, Germany) system by a well-trained ophthalmic photographer. All the averaged B-scans in this study had a signal quality of at least 15 dB [15]. Imaging of the peripapillary retinal nerve fiber layer (p-RNFL) was performed at a central wavelength of 870 nm. For each participant, a scan circle with a diameter of approximately 3.45 mm (consisting of 1536 A-scans) was manually positioned at the center of the optic disc, with the eye-tracking system activated. Fifteen B-scans were captured at the same location and automatically averaged using the built-in software (Heidelberg Eye Explorer, version 1.6.1.0; Heidelberg Engineering) to enhance the signal-to-noise ratio of the images. The software’s automatic segmentation feature was employed to delineate the p-RNFL on each averaged B-scan. The measurements obtained from the SD-OCT scans included the mean circumpapillary p-RNFL thickness as well as the sectorial peripapillary p-RNFL thickness in six specific regions: nasal, temporal, superonasal, inferonasal, inferotemporal, and supratemporal. These measurements were derived from the segmented p-RNFL using the aforementioned built-in software. To ensure high-quality scans, all averaged B-scans included in this study had a signal quality of ≥ 15 dB, as per the manufacturer's recommendation.

Visual acuity was measured by using a logarithm of the minimum angle of resolution (log MAR) chart (Nidek Inc., Gamagori, Aichi, Japan). Cycloplegic autorefraction was performed using an autorefractor (ARK-510A, Nidek, Gamagori, Japan) after a cycloplegic regime, which consisted of at least two cycles of eye drops. At the first cycle, two separate eye drops, cyclopentolate 1% (Cyclogyl, Alcon-Convreur, Rijksweg, Belgium) and tropicamide 1% (Santen, Osaka, Japan), were administered to both eyes 5 min apart. A second cycle of the same cycloplegic drops would be administered 10 min after the first cycle. A third cycle of the same cycloplegic eye drops would be given 30 min after the second cycle if pupillary light reflex was still present or the pupil size was less than 6.0 mm. Further cycles of cycloplegic eye drops would be administered if necessary to ensure the pupils are well dilated. Five readings, all of which had to be less than 0.25 D apart, were obtained and averaged. Ocular axial length (AL) was measured on a Zeiss IOL Master unit (Carl Zeiss Meditec Inc., Dublin, CA, USA) using noncontact partial coherence interferometry. A non-contact tonometer, Corvis ST (Corvis ST; Oculus, Arlington, WA, USA) was used to measure the subjects’ intro-ocular pressure (IOP) and central corneal thickness. Trained ophthalmologists conducted complete ocular examinations for each participant, including examinations of the anterior segment, posterior segment, and ocular motility [13‐16]. Spherical equivalent refraction (SER) was defined as spherical diopters (D) plus one-half of the value for cylindrical diopters.

Blood pressure (BP) was measured in the seated position after a 5-min rest using a digital autonomic BP monitor (Vital Signs Monitor; Heal Force Bio-Meditech, Shanghai, China). Body height and weight were measured using a professional integrated set (seca; Hamburg, Germany). A tape measure was used to measure the child’s waist and head circumference, the most prominent part of the forehead-1 to 2 fingers above the eyebrows and to the widest part of the back of the head. Three measurements were taken, with the average result taken for subsequent analysis. Body mass index (BMI) was calculated as body weight (in kilograms) divided by the square of body height (in meters).

Stata Version 14.0 (StataCorp LLC, College Station, TX, USA) was used for all statistical analyses. Confidence intervals (CIs) and p values (significant at the < 0.05 level) were derived for difference estimates. Mean differences for all parameters were calculated by subtracting the left eye parameters from the right eye parameters, and their respective p values were reported using the paired sample t test. Therefore, mean differences were stated as positive when right eyes had higher values than left eyes, and vice versa. To measure interocular correlation, intraclass correlation coefficients (ICCs) were computed. Absolute interocular differences were log-transformed due to the skewed distribution. Univariable and multiple linear regression analyses were conducted to determine associations, if any, between systemic, demographic, and ocular variables (as independent variables) with log-transformed absolute interocular p-RNFL thickness differences (as dependent variables). We interpreted the effect of each independent variable on the outcome and reported the result as the unstandardized correlation coefficients (β). We also calculated the standardized regression coefficients (sβ) for each independent variable. The dataset utilized in the current study was sourced from the same dataset as our previous studies [15], emphasizing different research questions. Any data with a p value < 0.05 were considered as statistically significant.

Among the 4273 children who participated in the study, 159 either declined later or were not able to complete the OCT examination, and another 80 children were also excluded due to suboptimal imaging quality. Consequently, 8068 eyes from 4034 children were included in the final analysis, including 2067 boys (51.2%) and 1967 girls (48.8%), with a mean age of 7.61 ± 0.98 years. Among them, 1009 had myopia with the mean SER of − 1.95 ± 1.33 D (Table 1).

The means for global p-RNFL thickness of right and left eyes were, respectively, 106.60 ± 9.41 μm and 105.99 ± 9.30 μm, with ICC 0.866 (95% CI 0.858 to 0.873, p < 0.001) (Table 2). The absolute difference of global p-RNFL thickness was difference was 3.58 ± 3.84 μm (Table 3).

Among the differences of p-RNFL in various quadrants, the temporal superior quadrant had the largest degree of asymmetry, with an ICC value of 0.567 (95% CI 0.546–0.588) and an absolute mean difference of 13.30 μm; while the least asymmetry was found to be in the nasal inferior quadrant, with a ICC being 0.736 (95% CI 0.721–0.750) and an absolute mean difference of 12.89 μm (Tables 2, 3).

After adjusting for confounders such as age, sex, BMI, systolic and diastolic BP, waist and head circumference, VA, IOP, CCT, and central macular thickness, we found greater axial length differences was significantly associated with increased absolute global p-RNFL thickness difference (β = 0.514; 95% CI = [0.342, 0.687]; p < 0.001). This association with AL was significant across all subfields except temporal superior region (β from 0.051 to 0.649; p < 0.001 to 0.039) (Table 4). The univariable analysis echoed these observations, pinpointing axial length and SE differences as the predominant factors influencing the absolute global p-RNFL thickness difference, whereas variables such as age, sex, and IOP did not exhibit significant associations (Supplementary Table 1).