Design

This research was designed as a prospective cross-sectional study that was conducted at a tertiary referral center. The protocol for this study was granted approval by the Adiyaman University Clinical Research Ethics Committee (2021-3-5) and registered as an International Standard Randomized Controlled Trial (International Standard Randomized Controlled Trial Number 04891211). All procedures performed were conducted in accordance with the ethical principles stated within the Declaration of Helsinki for research conducted involving human subjects. A signed written informed consent form was obtained from the families of the children who participated in the study.

Study Subjects

Study subjects underwent a detailed ophthalmologic evaluation. Caucasian pediatric subjects who met the following criteria were accepted for inclusion in the study group (group 1): 1) age <18 years; 2) serum vitamin D (25[OH]D, calcidiol) level <20 ngr/mL in previous 2 weeks ; 3) visual acuity ≥20/20; 4) manifest refraction spherical equivalent ≤3 diopters (D); and 5) normal biomicroscopic and fundus examinations. The exclusion criteria were as follows: 1) history of any chronic ocular diseases, except for refractive error (eg, uveitis, glaucoma, retinopathy of prematurity); 2) history of ocular surgery (eg, cataract, strabismus, open globe injury, laser photocoagulation); 3) having an ocular abnormality (eg, persistent fetal vasculature, optic disc hypoplasia, fovea plana); 4) history of systemic diseases that have the potential to affect ocular tissues (eg, diabetes mellitus, Graves’ disease, albinism, Down syndrome, Fabry disease, or Wilson syndrome); 5) history of any systemic diseases that can affect serum vitamin D levels (eg, parathormone or calcium metabolism disorders); and 6) already receiving vitamin D treatment. Group 1 was also separated into subgroups, as group 1a (≤5 ngr/mL vitamin D, severe deficiency), group 1b (≤15 ngr/mL vitamin D, deficiency), and group 1c (≤20 ngr/mL vitamin D, insufficiency). In the control group (group 2), the same inclusion and exclusion criteria were determined for white pediatric subjects who had normal vitamin D levels (>20 ngr/mL vitamin D <100 ngr/mL). Subjects were confirmed to meet the inclusion criteria by an experienced pediatric endocrinologist (S.B.) and a retinal specialist (E.A.).

Quantitative Assessments

To assess serum vitamin D levels, samples were analyzed within 20 minutes at a single laboratory with the electrochemiluminescence immunoassay method using a Cobas 8000 analyzer (Roche Diagnostics, Mannheim, Germany).

Peripapillary RNFL, central macula, and retinal layer thickness parameters were evaluated using a spectral-domain optical coherence tomography (OCT) device (Spectralis, Heidelberg, Germany) and Heidelberg Eye Explorer software (Heidelberg, Germany). The device takes 40,000 a-scans per second, simultaneously, and also provides quantitative and qualitative data, in detail, about the macular thickness and configuration. The measurements were performed under the same dim light conditions and a quality score ≥20 was considered acceptable. The peripapillary RNFL thickness was obtained from a 3.4-mm peripapillary circular area that was located within the center of the optic disc. To evaluate the detailed central macula thickness parameters, the macular thickness map was formed by a 25-line horizontal raster scan, covering 20° × 20°, centered on the fovea. The system also gave the average thicknesses of the retinal layers in macular area.

Using the same OCT system with the enhanced depth imaging function, choroidal thickness was manually measured in the subfoveal and subfoveal 1500-µm– and 3000-µm–diameter distance nasal and temporal areas. Realizing that small differences in the positioning could affect the measured thicknesses, a reference point was determined as the thinnest point of the macula to obtain the best reliable measurements. The OCT evaluations were performed by a highly trained ophthalmic technician, and the compatibility of the results were confirmed by 2 experienced retinal specialists (E.A., G.A.A).

A VISUCAM 500 fundus camera system (Carl Zeiss Meditec, Jena, Germany) was used to take 50° colored fundus photographs. The central retinal artery and vein calibers were analyzed with the Interactive Vessel Analyzer (with the permission of Dr. Nicola Ferrier at the University of Wisconsin-Madison, Madison, Wisconsin, USA), which is a semiautomated system that is used to conduct measurements of the retinal vessel widths via digital retinal images. The fundus photographs were taken by the same trained technician and were sent to masked experienced researchers (C.I., A.H.B) for measurement of the retinal vasculature caliber. For determination of the vascular measurement field, 3 concentric rings were placed on the fundus images and 2 zones were described, comprising the zone extending from the disc margin to the half-disc diameter, as well as the zone extending from the half-disc to 1-disc diameter. The measurements were performed by the first researchers and confirmed by the second researchers. When the difference between 2 measurements was >2%, remeasurements were performed. The CRAE and CRVE values were calculated using the formula that was established by Hubbard and associates, which was later revised in a study by Knudtson and associates.

Statistical Analysis

The data obtained from the right eyes of the participants were included in the statistical analyses. The analyses were conducted using IBM SPSS Statistics for Windows (version 22.0; IBM Corp., Armonk, New York, USA). The descriptive statistics were presented as the mean ± standard deviation (SD) and minimum–maximum values. Comparison of the categorical values was conducted using the chi-square test. Testing of the normal distribution of the variables was conducted using the Kolmogorov–Smirnov test. Comparison of the independent samples was conducted using the 2-sample t test. It was confirmed by analysis of covariance that some variables including age and refractive status are not significant factors for variables found as statistically significant. To investigate any relationship between the quantitative results, Pearson correlation analysis was used. Analysis of the subgroups was performed using 1-way analysis of variance. The 2-sample t test was also used in the post hoc analysis. P < .05 was considered statistically significant.