The data were collected from an ongoing Asan Glaucoma Progression Study. Between January 2011 and March 2017, the medical records of all patients who were evaluated by a single glaucoma specialist (KRS) at the glaucoma clinic of Asan Medical Center, Seoul, Korea were retrospectively reviewed in this longitudinal, observational study.

Initial testing included a comprehensive ophthalmologic examination, which included a medical history review, best-corrected visual acuity measurement, slit-lamp biomicroscopy, multiple intraocular pressure (IOP) measurements using Goldmann applanation tonometry, gonioscopy, dilated fundoscopic examination using a 90- or 78-diopter lens, stereoscopic optic-disc photography, retinal nerve fiber layer (RNFL) photography, visual field (VF) testing, central corneal thickness measurement (DGH-550 instrument, DGH Technology, Exton, PA, USA), and AXL measurement (IOL Master, Carl Zeiss Meditec, Dublin, CA, USA). RNFL thickness measurement was performed using spectral-domain optical coherence tomography (SD-OCT; Spectralis OCT, Heidelberg Engineering, Dossenheim, Germany). Young glaucoma patients aged 17 to 30 years old at baseline were selected. The inclusion criteria at the initial assessment included a best-corrected visual acuity of 20 / 40 or better, an AXL ≥24 mm, a normal anterior chamber, an open angle on slit-lamp and gonioscopic examinations, and all IOP measurements lower than 21 mmHg. Patients with glaucomatous optic disc changes, such as diffuse or focal neural rim thinning, disc hemorrhage, or RNFL defects, as confirmed by two glaucoma specialists (KRS and JWS), were included. Patients with any other ophthalmic or neurologic condition that could result in a VF defect, a history of diabetes mellitus, or severe myopic fundus changes precluding adequate examinations were excluded. Pseudophakic and aphakic eyes were also excluded. If both eyes in the same patient were found to be eligible, one eye was randomly selected for analysis.

All patients were followed up at 6-month intervals for at least 2 years using stereoscopic optic disc/RNFL photography, VF testing, and SD-OCT scanning. All tests were performed at the same visit or within 2 weeks thereof. Furthermore, at least five qualified SD-OCT images obtained at different visits were required for the patient to be included in the study. If the patient underwent intraocular surgery or laser therapy during the follow-up period, only data obtained before such operations, which should have been longer than 2 years prior to the operation, were included.

VF tests were performed using a Humphrey field analyzer (Swedish interactive threshold algorithm 24-2, Carl Zeiss Meditec). Only reliable VF test results (false-positive errors <15%, false-negative errors <15%, and fixation loss <20%) were included. The VF test was repeated within 2 weeks of the baseline measurement for confirmation. Patients were expected to return approximately 1 month after the baseline examination in order to assess their responses to medication and undergo a third VF test. Hence, all patients underwent three VF tests within the first 6 weeks. Data from the first VF test were excluded to obviate any learning effects; and the results of the second and third VF tests, which were performed within 1 month of each other, were considered the baseline examinations. Patients who underwent ≥5 reliable VF tests, excluding the first VF test, were included. Glaucomatous VF defects were defined by glaucoma hemifield test results outside the normal limits or a pattern standard deviation outside the 95% range of normal limits. In addition, these eyes had to have a cluster of 3 points with probabilities <5% on the pattern deviation map in ≥1 hemifield, including ≥1 point with a probability <1% or a cluster of 2 points with a probability <1%. Glaucomatous VF defects had to be confirmed on ≥2 VF examinations. All patients underwent medical therapy. The study was approved by the institutional review board of Asan Medical Center (2017-1311), and the study design followed the principles of the Declaration of Helsinki. Informed consent was waived due to the retrospective nature of this study.

Spectral-domain enhanced depth imaging was used to scan the optic disc. Brief ly, the entire optic disc was scanned using a 6-mm length line (512 A-scans) with an interval of 50 µm. In our study, an average of 35 horizontal B-scans was produced in enhanced depth imaging mode. From these B-scans, the frame that passed through the optic disc center was selected. The structure of the optic disc was analyzed with the intrinsic viewer. Images were imported to ImageJ software ver. 1.48 (National Institutes of Health, Bethesda, MD, USA; http://imagej.nih.gov/ij) and magnified by two. Bruch's membrane opening (BMO, points A and B in Fig. 1) was defined by the proximal tips of the Bruch's membrane. A line connecting the BMO on each side (line AB) was regarded as the BMO plane.

A line (line AC) connecting the nasal BMO and the proximal margin of the external oblique border tissue was drawn and set as the optic canal plane. The disc tilt angle (DTA) was defined as the angle between the BMO plane and the optic canal plane that passes through the optic disc (θ). The DTA was measured using the ‘angle tool’ in the ImageJ software. All accepted images exhibited a centered optic disc, were well-focused, and had even and adequate illumination. Eyes were excluded when these points could not be clearly identified. All measurements were performed by a single well-trained examiner (JYY). Before the main analysis, an intra-examiner intra-class correlation coefficient was calculated using 20 randomly selected images to test the reproducibility of the DTA measurements. Intra-examiner intraclass correlation coefficient value for DTA was 0.947 (95% confidence interval, 0.866 to 0.983). Inter-examiner intraclass correlation coefficient value was 0.909 (95% confidence interval, 0.817 to 0.975).

Glaucoma progression was determined by either structural or functional aspects. Structural progression was assessed by serial optic disc/RNFL photographs. Two glaucoma experts (KRS and JWS) independently assessed all photographs to estimate glaucoma progression. In each patient, the most recent photograph was compared with the baseline photograph. The two graders were unaware of the progression assessments made by the other, and each grader viewed all photographs of each eye before making an assessment. Both graders were asked to determine glaucomatous optic disc or RNFL progression as demonstrated by the increase in the extent of neuroretinal rim thinning, enhancement of disc excavation and/or any widening, deepening, or new RNFL defects. The graders classified each glaucomatous eye as either stable or progressing and determined the type of structural change. Subjects were classified with progression when an eye showed progression on either the optic disc or RNFL photograph. If the RNFL photographs were difficult to evaluate because of diffuse atrophy or invisible RNFL layers due to a lightly pigmented fundus, progression was determined using optic disc assessment. If the opinions of the two observers differed, a consensus was reached after discussion.

VF progression was determined by the Early Manifest Glaucoma Trial criteria or the linear regression analysis of mean deviation (MD) values. Definite Early Manifest Glaucoma Trial progression required that at least three test points were flagged as significantly deteriorated at the same test point locations in 3 consecutive fields [12]. These changes should also have been observed at the latest follow-up visit. In the linear regression, VF progression was defined as a significant negative slope between MD and age.

DTA measurements at baseline and last follow-up were compared using the Wilcoxon signed rank test. Demographics and clinical characteristics were compared between progressed eyes and stable eyes in terms of glaucomatous damage using the Mann-Whitney U test. A linear mixed effect model was used to assess the influence of parameters (age, sex, baseline IOP, mean follow-up IOP, RNFL thickness, VF MD, AXL, central corneal thickness, and presence of glaucomatous progression) upon the DTA change. Variables with a p-value ≤0.2 in univariate analyses were included in a multiple mixed effects model. The final models for outcomes were determined by backward elimination procedures. A p-value less than 0.05 was considered statistically significant. R ver. 3.2.2 (R Core Team, 2013) and IBM SPSS Statistics ver. 21.0 (IBM Corp., Armonk, NY, USA) were used for the statistical analyses.