The study cohort included 223 full-term infants and children (104 females and 119 males), ages 2 days to 15 years (mean, 5.7 ± 4.1 years). Newborns were prospectively recruited from the Maternity Unit at the University Hospitals of Leicester, Leicester, UK. Older children were recruited from Leicester City nurseries and schools. Adult values were established from an additional 59 adults (41 females and 18 males), ages 16 to 47 years (mean, 33.5 ± 8.7 years). Adult participants were students and staff of the University of Leicester. This study followed the tenets of the Declaration of Helsinki and was approved by the local ethics committee. Informed consent was obtained from all participant parents or guardians and older children gave their assent before examination.
All subjects underwent an examination of best-corrected visual acuity (BCVA), ocular motility, stereopsis, refraction, slit-lamp biomicroscopy, fundoscopy, and posterior and AS-OCT. Refraction was tested in most children using a portable Plusoptix A12C autorefractor (Plusoptix, Nuremberg, Germany), without cycloplegic eye drop administration. The BCVA of infants and young children unable to cooperate was tested using preferential looking (Teller Acuity Cards; Washington Research Foundation, Seattle, WA). If cooperation did not allow us to obtain BCVA or refraction, children were invited back at an older age for re-examination; logMAR Kay acuity cards were used for cooperative children. All subjects were free of ocular and neurological pathology and had good general health.
All subject eyes were scanned with an Envisu C-Class handheld OCT system (Leica Microsystems, Wetzlar, Germany) without sedation. The protocol used had a volumetric horizontal scan of 18 mm width and 6 mm height and contained 11 B-scans, with 3000 A-scans per B-scan. The rapid acquisition sequence used (0.96-second duration for the full scan and 87 ms for each B-scan) helped reduce motion artifacts. This was also aided by the child's head being kept still by the parent/carer and with the child's attention being distracted by watching a cartoon video. The scans were acquired and reviewed by a trained examiner. The criteria defined for the quality of usable images included (1) both temporal and nasal angles could clearly be seen on the horizontal B-scan; (2) a clear peripheral cornea was visible; and (3) both irises and corneoscleral junctions could be distinguished. The examination room light was controlled at ∼200 lux. Limited scanning depth of the handheld OCT (only 2.5 mm in tissue) does not permit imaging of the whole anterior segment; therefore, we obtained two separate scans: one scan showing the cornea (
Fig. 1A) and another scan showing the nasal and temporal anterior chamber angles (
Fig. 1B).
The OCT images were imported into ImageJ software (National Institutes of Health, Bethesda, MD)
17 for analysis. The best individual B-scans were selected to further assess the quality of images and perform quantitative measurements. At least one good quality scan of the anterior chamber and one of the cornea was successfully obtained per eye, per participant.
The central corneal thickness (CCT) was measured from a separate corneal image (
Fig. 1A). The anterior chamber measurements were obtained from the B-scan showing both nasal and temporal angles (
Fig. 1B). An ImageJ script was developed to perform semiautomated measurements after manual identification of the angle landmarks: the scleral spur (SS) and Schwalbe's line (SL) in both nasal and temporal angles as described by Sakata et al.
18 The agreement in our identification of SS and SL was tested by two assessors. The intra-assessor and inter-assessor agreement yielded high repeatability and reproducibility of the localization of SS and SL.
19
The anterior chamber width (ACW) was measured as the distance between (1) the nasal scleral spur to temporal scleral spur (SS-SS-D), and (2) the nasal Schwalbe's line to temporal Schwalbe's line (SL-SL-D). The nasal and temporal angle parameters were derived from the Schwalbe's line in each angle. These measurements correspond to Schwalbe's line–angle opening distance (SL-AOD) (
Fig. 2) and Schwalbe's line–trabecular iris surface area (SL-TISA), described by Cheung et al.
20 The trabecular meshwork length (TML) was measured as the distance between Schwalbe's line and the scleral spur. We investigated the test–retest reproducibility of these measurements and detected high reproducibility and repeatability, with intraclass correlation coefficient > 0.9 and small measurement errors.
21
Multivariable fractional polynomial regression models were used to model the effect of age on anterior segment measurements. The modeling automatically transformed age into a set of powers (–2, –1, –0.5, 2, or 3 or logarithmic transformations) in order to achieve a more normal distribution of the skewed data. Linear mixed models (adjusted for the transformed age, eye laterality, and gender variation) were then used to establish normative anterior chamber measurements by calculating the mean fit, 95% confidence intervals, and 95% upper and lower prediction intervals of the data. The nasal and temporal angle measurements were analyzed separately. We also calculated the rate of change of each measurement, per year, to determine the development of the anterior chamber and to identify the age when the maximum value, or a plateau of adult values, is reached. Gender differences in anterior segment measurements and difference between nasal and temporal angle measurements were assessed at different ages, including at birth and at 1, 5, 18, and 35 years of age.
The correlation between spherical equivalent and the anterior chamber measurements was investigated in a subgroup of 154 children (average spherical equivalent power = 0.5 ± 1.1 diopters). First-order partial correlations were used to adjust for the effect of age, gender, eye, and angle variations. The analysis was performed using Stata 15 (StataCorp, College Station, TX).