Abstract
Purpose
To investigate longitudinal changes in choroidal thickness (ChT) after 1-year treatment of repeated low-level red-light (RLRL) and their predictive value in efficacy on myopia prevention and myopic shift among pre-myopic children.
Methods
278 pre-myopic (-0.50 D < spherical equivalent refraction, SER ≤ 0.50 D) participants were assigned to the RLRL group and control group randomly and evenly. The OCT, visual acuity, AL, SER and other parameters were measured before enrollment and every 3 months after intervention. The data from both eyes of the included participants were analyzed.
Results
A total of 463 eyes were analyzed. Due to the COVID-19 pandemic, 176 participants in the RLRL group had treatment interrupted. The continued RLRL group, interrupted RLRL group and control group were well balanced in baseline characteristics. In the continued and interrupted RLRL group, the average ChT increased significantly at 3-month visit (all P < 0.001) and the subfoveal ChT thickened evidently. The area under the curve (AUC) for the models including gender and 3-month change in ChT to predict satisfactory myopia prevention at 12 months was 0.983. The efficacy of the models that also used the combined indicators of baseline age, gender and the 3-month change in ChT to predict AL progression control over 12 months reached 0.944.
Conclusions
Continued RLRL intervention induced notable thickening of ChT in premyopic population, especially at the subfoveal sector. For participants received RLRL treatment, the 3-month change in ChT combined with other baseline factors have acceptable predictive discrimination of myopia prevention efficacy.
Introduction
Myopia, a refractive disorder characterized by optical myopic defocus and blurred vision, has emerged as the most prevalent ocular condition globally, especially prevalent in East Asia. Furthermore, an earlier myopia onset and prolonged progression of myopia significantly increase the risk of macular degeneration and vision impairment, which is often associated with a decrease in the choroidal thickness (ChT). Consequently, myopia prevention is a critical area of focus in the refractive development of children and adolescents. International Myopia Institute defined pre-myopia as a refractive state in which the spherical equivalent refraction (SER) is in the range of −0.50–0.75 diopter (D) inclusive in children and adolescents. This condition, when assessed alongside baseline SER, age, and other measurable risk factors, indicates a substantial likelihood of myopia onset, thereby warranting preventative interventions.
Recent advances have introduced repeated low-level red-light (RLRL) as a promising intervention to control the myopic shift by slowing axial length (AL) elongation and controlling refraction. Our previous randomized controlled trial (RCT) provided evidence that RLRL controlled myopic shift and effectively prevented the development of myopia in pre-myopic children. In addition, our RCT showed that subfoveal ChT increased in the intervention group which was consistent with prior studies.
A vital role in emmetropization is played by the choroid, which thickens or attenuates itself to enable external light to focus accurately on the retina and influences the synthesis and degradation of the scleral extracellular matrix. Furthermore, short-term variations in subfoveal ChT have been linked to long-term changes in AL among myopic children undergoing treatments such as atropine eye drops and orthokeratology. Xiong et al. reported a rapid increase in subfoveal ChT among children with myopia exposed to RLRL and it could predict long-term myopia control efficacy. We speculate this predictive efficacy could extend to pre-myopic children and provide reference to clinical practice. In addition, changes across different choroidal regions post-RLRL treatment and the possible predictive value of other regions other than the central sector remain unexplored.
To test this hypothesis, we investigated the changes in different regions of the choroid after RLRL treatment and the predictive value of myopia prevention and control efficacy.
Methods
Study design
This study is a secondary analysis of data from a RCT. In brief, the RCT was a 12-month, two-group, single-blind, school-based study conducted in Shanghai, China (ClinicalTrials.gov identifier: NCT04825769). This trial evaluates the efficacy and safety of this novel intervention for preventing myopia onset and controlling myopic shift among pre-myopic children aged 6–10 years.
We conducted screenings of students in grades 1 through 4 across 10 primary schools, enrolling those identified as pre-myopia (-0.50 D < cycloplegic SER ≤ 0.75 D for at least one eye) and with one parent at least having a SER ≤ -3.00 D in either eye. Given the −1.00 D decrease in SER during the one year before myopia onset, the inclusion criteria were established as −0.50 D < SER ≤ 0.50 D for at least one eye. This was intended to select participants who were at a higher risk of developing myopia within the subsequent year. Children were excluded from the study if they had astigmatism ≥ 1.50 D, anisometropia ≥ 1.50 D, strabismus, or other ocular abnormalities. Additionally, those with systemic diseases or prior myopia interventions were also not eligible for inclusion.
After enrollment, children were assigned to the RLRL group and control groups randomly and evenly. Children in the RLRL group were exposed to the red light twice daily, five days per week, with each session having a duration of three minutes. Each session was conducted with a minimum interval of four hours between them. If a child was eligible for enrollment with only one eye, we sealed the laser beam hole on the opposite side of the instrument ( Supplementary Figure 1 ). Initial treatment was supervised on-site by the research team, while subsequent treatments were conducted under the supervision of school teachers.
During the last three months of the trial, due to the COVID-19 pandemic, a subset of subjects in the RLRL group continued to receive the intervention at home, while others did not take the machine home and had their treatment interrupted. Therefore, we further subdivided the participants into the continued intervention group, the interrupted intervention group, and the control group (group A, B, C, respectively). In addition, a dedicated information system recorded the actual number of interventions received by the children in real-time to evaluate compliance. During the trial period, no interventions were applied to the participants in the control group.
This trial received approval from the Shanghai General Hospital Ethics Committee ([2021]022) and conducted in accordance with the ethical guidelines proposed by the Declaration of Helsinki.
Ophthalmic examinations
Ophthalmic data collection during the five visits were obtained by the same examiner, utilizing consistent criteria and protocols throughout the study. Before the trial began, researchers and project implementation staff underwent and completed relevant training and assessments.
Uncorrected visual acuity (UCVA) was assessed using the Early Treatment Diabetic Retinopathy Study (ETDRS) logMAR chart (Wehen 06, Guangzhou Xieyi Weishikang). AL was measured using the IOLMaster (Carl Zeiss 500, Meditec) before cycloplegia. The final recorded results were the averages of five readings, with the maximum difference between readings being less than 0.05 mm. 1 % cyclopentolate (Alcon) was administered as two drops, separated by a five-minute interval, to induce cycloplegia. Pupillary light reflex and pupil size were checked 30 minutes after the administration of the eye drops. SER after cycloplegic were measured by optometrist using an autorefractor (KR8900, Topcon). The subsequent optometrist verified that the maximum difference among the three readings of spherical or cylindrical ≤ 0.25 D, of axis ≤ 5 degrees. Anterior segment indicators were captured using a Scheimpflug camera (Pentacam HR, Oculus Optikgeräte GmbH).
OCT imaging and analysis
Fundus imaging was obtained with a swept-source optical coherence tomography (OCT) instrument (DRI-OCT Triton, Topcon). Before performing OCT scanning, it is essential to reconfirm cycloplegia completed to optimize image quality. Additionally, to minimize the impact of diurnal rhythm on ChT measurements, OCT scans were taken between 10 AM and 3 PM.
The fundus images, collected under a 12-line radial 9 mm scan pattern, were analyzed using the OCT’s built-in software to segment the fundus layers. All images were reviewed by a specialized and experienced technician (B.Z.). Image quality was scored by the built-in software, while images with a score lower than 90 were excluded from the analysis. Choroidal thickness was defined as the vertical distance between the choroid-sclera interface and Bruch’s membrane.
An ETDRS grid was applied to all OCT images, defining the regions of subfovea, parafovea, and fovea with diameters of 1 mm, 3 mm, and 6 mm, respectively. These circles were further divided into superior, inferior, temporal, and nasal quadrants. The average ChT of all nine sectors and ChT in each grid sector were calculated using the built-in software.
Statistical analysis
Data from both eyes of the participants were analyzed, except in cases where only one eye met the inclusion criteria. Missing values were not imputed. This analysis included only SER data that confirmed full cycloplegia to ensure the accuracy of the results. All P values were derived from two-sided tests, with statistical significance set at P < 0.05. Statistical analyses were conducted using SAS, version 9.4 (SAS Institute, Inc.).
The distribution of the data was examined by the Kolmogorov–Smirnov test. Baseline characteristics were described as means with standard deviations or counts with percentages. Comparison of baseline data was performed using chi-square tests for gender and analysis of variance (ANOVA) tests for other characteristics. Univariate regression models were conducted to explore the correlation between each possible indicator and SER at 12 months or 12-month changes in AL. Mixed-effects models were employed to further evaluate the association while adjusting for intraclass correlation between bilateral data, allowing for a comparison with the results from the univariate analysis. To facilitate visualization of the results, baseline central corneal thickness (CCT) and changes in ChT were rescaled and divided by 10 to avoid overly small β values. Factors with P values < 0.10, including age, gender, UCVA, SER, and AL at baseline and 3-month changes in ChT, were included and examined in the consequent analysis. Stepwise regression models were then established using these potential associated covariates to predict myopia prevention and control effect over 12 months in group A. To compare and evaluate the predictive performance of various models, receiver operating characteristic (ROC) curves were utilized, and the area under the curve (AUC) was calculated. According to previous reports, a satisfactory myopic shift control threshold was defined as SER at 12-month not less than −0.50 D and annual progression rates of AL no more than 0.15 mm.
Result
Baseline characteristics
This study enrolled 287 subjects. Among these, 111 of the 574 eyes did not meet the criteria for pre-myopic status, leaving 463 eyes for inclusion in this analysis. There were 55 (11.88 %) eyes in group A, 176 (38.01 %) in group B and 232 (50.11 %) in group C. The baseline characteristics are described in Table 1 . The distribution of most indicators was well balanced among the three groups (P > 0.05). Only baseline CCT among groups were distributed statistically different ( P = 0.036) but without clinical significance.
| Included | ||||
|---|---|---|---|---|
| Group A | Group B | Group C | P value | |
| Age, years | 8.27 (1.10) | 8.30 (1.11) | 8.32 (1.09) | 0.730 |
| Male, % | 26 (47.27 %) | 99 (56.25 %) | 113 (48.71 %) | 0.259 |
| Height, cm | 131.54 (7.95) | 133.43 (9.41) | 132.41 (8.02) | 0.979 |
| Weight, kg | 29.23 (6.16) | 30.68 (8.11) | 29.97 (7.55) | 0.938 |
| UCVA, logMAR | 0.08 (0.09) | 0.09 (0.08) | 0.08 (0.08) | 0.408 |
| SER, D | 0.20 (0.28) | 0.16 (0.28) | 0.21 (0.27) | 0.427 |
| Corneal Radius, mm | 7.87 (0.24) | 7.88 (0.24) | 7.83 (0.25) | 0.099 |
| AL, mm | 23.30 (0.56) | 23.42 (0.68) | 23.29 (0.67) | 0.341 |
| Pupil diameter, mm | 6.27 (0.83) | 6.50 (0.85) | 6.31 (0.96) | 0.453 |
| Central corneal thickness, μm | 553.31 (34.59) | 552.09 (36.53) | 560.79 (29.39) | 0.036* |
| Anterior chamber depth, mm | 3.69 (0.21) | 3.72 (0.22) | 3.71 (0.20) | 0.766 |
| Lens thickness, mm | 3.61 (0.20) | 3.67 (0.19) | 3.64 (0.22) | 0.965 |
| Subfoveal ChT, μm | 265.80 (64.71) | 278.06 (57.73) | 278.63 (55.98) | 0.338 |
| Para temporal ChT, μm | 279.00 (62.01) | 288.42 (54.50) | 294.18 (53.04) | 0.107 |
| Para superior ChT, μm | 266.62 (59.72) | 276.21 (55.30) | 274.40 (53.93) | 0.664 |
| Para nosal ChT, μm | 233.06 (62.84) | 242.33 (55.74) | 240.53 (55.45) | 0.687 |
| Para inferior ChT, μm | 263.19 (66.34) | 273.55 (59.30) | 279.71 (55.93) | 0.103 |
| Peri temporal ChT, μm | 280.20 (55.55) | 285.27 (46.99) | 294.14 (50.18) | 0.052 |
| Peri superior ChT, μm | 260.95 (54.31) | 269.11 (51.24) | 266.15 (51.96) | 0.875 |
| Peri nosal ChT, μm | 186.15 (56.26) | 189.58 (49.24) | 189.42 (49.33) | 0.799 |
| Peri inferior ChT, μm | 249.01 (56.18) | 257.21 (51.78) | 264.23 (51.31) | 0.072 |
| Average ChT, μm | 248.27 (54.35) | 255.43 (48.07) | 258.33 (48.08) | 0.266 |
Topographic longitudinal changes in ChT
To intuitively visualize the changes in the choroid, we divided the macular-centered choroid into nine sectors based on the ETDRS grid ( Fig. 1 ). In group A and B, ChT changes increased horizontally from the temporal to the central sector, then decreased toward the nasal quadrant. A similar trend was observed vertically from the superior to the inferior quadrant. Of note, in group B, the 3-month changes in ChT across the nine sectors were all significantly different from their adjacent sectors (all P < 0.05), with the most evidence increase in the subfoveal sector (21.36 ± 21.56 μm). Over time, fewer sectors showed significant differences in ChT changes from neighboring sectors. In group C, the trend was the opposite; the three-month ChT changes remained consistent across most sectors. However, by the 12th month, significant differences emerged between most sectors and their adjacent sectors, with the subfoveal sector showing the most significant changes in ChT.
