This retrospective study included 343 glaucoma subjects who were examined at the glaucoma clinic of the Asan medical center. Patients who met our inclusion criteria of at least three years of follow-up and seven or more reliable VF measurements were consecutively enrolled by medical record review. Eyes that underwent intraocular surgery (including cataract and glaucoma filtering procedures) during the follow-up period were excluded. At initial diagnosis, each patient received a comprehensive ophthalmologic examination that included a review of medical history for the presence of systemic diseases like systemic hypertension, diabetes, cerebrovascular accident, dyslipidemia, and migraine, measurement of best-corrected visual acuity to confirm that visual acuity was adequate for automated perimetry performance, slit-lamp biomicroscopy, Goldmann applanation tonometry, gonioscopy, dilated fundoscopic examination using a 90- or 78-diopter lens, stereoscopic optic disc and retinal nerve fiber layer (RNFL) photography, VF examination by standard automated perimetry (Humphrey Swedish interactive threshold algorithm standard strategy 24-2; Carl Zeiss Meditec, Dublin, CA, USA), and measurement of central corneal thickness (CCT; DGH-550 instrument, DGH Technology Inc., Exton, PA, USA).

All included patients met the following criteria at the baseline exam: best-corrected visual acuity of 20 / 30 or better, with a spherical equivalent (SE) within ±5 diopters and a cylinder correction within +3 diopters; VF mean deviation (MD) higher than -20 dB; and the presence of a normal anterior chamber and open-angle on slit-lamp and gonioscopic examinations. Patients with concomitant diseases known to affect the VF were excluded. Diagnosis of glaucoma was based on both the presence of typical glaucomatous changes in the optic disc and glaucomatous VF defect. Thus, eyes with a glaucomatous optic disc and a normal VF were not defined as having glaucoma. Glaucomatous optic disc change included increased cupping (vertical cup-disc ratio more than 0.7), a difference in the vertical cup-disc ratio (more than 0.2 between the eyes), diffuse or focal neural rim thinning, disc hemorrhage, or RNFL defects. Eyes defined as having glaucomatous VF defects met two of the following three criteria (as confirmed by more than two reliable consecutive tests): 1) a cluster of three points with a probability of less than 5% on a pattern deviation map in at least one hemifield, including at least one point with a probability of less than 1% or a cluster of two points with a probability of less than 1%; 2) a glaucoma hemifield test result outside the normal limit; and 3) a pattern standard deviation result of less than 5%. Reliable VF assessment was defined as a VF test with a false-positive error of less than 15%, a false-negative error of less than 15%, and a fixation loss of less than 20%. The first VF test was excluded from analysis to reduce the learning effect. If at least one eye showed glaucomatous changes, that patient was included for analysis. IOP-lowering medication was prescribed if the eyes showed glaucomatous optic disc changes or had both a glaucomatous optic disc and VF changes. Glaucomatous eyes with secondary causes were excluded in the current study.

All procedures conformed to the Declaration of Helsinki, and the study was approved by the institutional review board of the Asan Medical Center.

Since glaucomatous damage involves both structural and functional changes that may not appear at the same time, glaucoma progression was determined by either structural or functional measures. Structural progression was assessed using a stereoscopic optic disc and red-free RNFL photographs. Serial stereoscopic photographs were displayed on an liquid crystal display monitor. Two glaucoma experts (KRS and JHN) independently assessed all of the photographs in order to estimate the glaucoma progression between the patients' first and last visits. Both graders were blind to each other's progression assessments and to all clinical and VF information. Photographs were presented in chronological order and with masking of patient identification, age, and test date. Each grader viewed all photographs of each eye before making an assessment and was asked to determine the possible presence of a glaucomatous optic disc or of RNFL progression, as revealed by an increase in the extent of neuroretinal rim thinning, enhancement of disc excavation, any widening or deepening of a RNFL defect, and/or the appearance of new disc hemorrhage [14,15]. Each grader classified each glaucomatous eye as either stable or progressing. If the opinions of the two observers differed, a third examiner (DJ) made the final decision.

Functional progression was determined using a VF series. Commercial software (Guided Progression Analysis, Carl Zeiss Meditec) was used to determine progressive functional loss of the VF. VF defect progression was defined as a significant deterioration from the baseline pattern of deviation at three or more of the same test points and was evaluated on three consecutive examinations [16].

If only one eye showed progression by either structural or functional assessment during the follow-up period, the patient was included in the UPG group. In contrast, if both eyes showed progression, the patient was included in the BPG group. Age, gender, and the prevalence of systemic diseases (systemic hypertension, diabetes, cerebrovascular accident, migraine, and dyslipidema) were compared between the UPG and BPG groups. In the UPG group, pre-treatment IOP, mean post-treatment IOP, CCT, SE, baseline VF MD, VF pattern standard deviation, and progression rates were compared between the PE and the NPE within the same patient. The progression rates (MD slope) were determined by linear regression analysis of MD values in serial VF analyses. In the BPG group, the same parameters were compared between the FPE and the SPE within the same patient. The Wilk-Shapiro test was used to determine the distribution of numerical data. Normally distributed data were compared between the UPG and BPG groups using unpaired t-tests. To compare categorical data, the chisquare test was employed. The paired t-test was used to compare eyes within the same patient. Non-normally distributed data were compared using either the Mann-Whitney or Wilcoxon signed rank test. Statistical analyses were performed using IBM SPSS ver. 21.0 (IBM Corp., Armonk, NY, USA). A p-value less than 0.05 was considered significant.