Glaucoma Progression in Eyes with a History of Refractive Corneal Surgery

Abstract

Purpose.: To evaluate the glaucoma progression in patients with a history of refractive corneal surgery (RCS).

Methods.: A total of 313 eyes of 313 glaucoma patients (RCS group, 34 eyes; non-RCS group, 279 eyes) with a mean follow-up period of 2.7 years were included. Glaucoma progression was determined either by experts' stereoscopic optic disc/retinal nerve fiber layer (RNFL) photographs or by serial visual field (VF) data. Uni- and multivariate Cox's proportional hazards models were used to determine associations between potential risk factors including a history of RCS and glaucoma progression.

Results.: Among 313 eyes, 87 (27.8%) eyes showed progression either by optic disc/RNFL photographs or VF assessment during the follow-up period. Ten (29.4%) of eyes in the RCS group and 77 (27.6%) of eyes in the non-RCS group showed glaucoma progression; however, prevalence of progression did not differ between the two groups (P = 0.482). Among all participants, baseline age and VF pattern standard deviation (PSD; hazard ratios of 1.013 and 1.119; P = 0.026 and P < 0.001, respectively) were significant risk factors for progression. In both the RCS and the non-RCS group, only VF PSD was a significant risk factor for progression according to multivariate Cox proportional hazards (1.193, 1.099, P < 0.001, P < 0.001, respectively).

Conclusions.: Baseline VF PSD and age substantially affected glaucoma progression in all study participants, although their RCS history was not a significant risk factor for glaucoma progression.

Introduction

Refractive corneal surgery (RCS) using a laser, such as photorefractive keratectomy (PRK), laser in situ keratomileusis (LASIK), or laser in situ epithelial keratomileusis (LASEK), has become popular for the treatment of refractive error. As most RCS candidates have myopia and as glaucoma is more prevalent in such patients, the safety of RCS in terms of glaucoma development or progression is frequently questioned by those who wish to undergo such surgery or who have a history of RCS. One concern is related to the reliability of intraocular pressure (IOP) measurements obtained after surgery. Because most current IOP measurement procedures are performed through the cornea, corneal ablation inevitably affects IOP readings. ^1–8 Inaccurate measurement of IOP and the subsequent possibility that IOP measurements cannot be used as a reliable guideline for the diagnosis and management of glaucoma have both become issues warranting attention. 

Another concern is whether the surgical process per se or postoperative management or other unknown causes may induce glaucomatous damage or aggravate glaucoma progression. Some frequently asked questions young glaucoma patients ask their physicians are, “As you said that I have glaucoma, if I underwent RCS, will the glaucoma progress faster than if I do not?” and “I underwent RCS a few years ago, and I didn't hear about glaucoma at that time. Now you are saying that I have glaucoma. Will it progress faster than in a glaucoma patient who did not undergo RCS?” 

Regarding the association between glaucoma and RCS, the possibility of optic nerve damage during acute IOP elevation when the suction ring of a microkeratome is placed has been suggested. ⁹ Postoperative IOP elevation caused by steroid use has been reviewed. ^10–13 However, there have been no reports which can provide answers to the above-stated patient questions. Therefore, we intended to evaluate glaucoma progression in patients with a history of RCS and to compare results with those of glaucoma patients without a history of RCS in order to answer those questions. 

Materials and Methods

Study Subjects

Subjects examined by a glaucoma expert (KRS) at the university-based glaucoma clinic (Asan Medical Center, Seoul, Korea) from March 2008 to November 2011 and who met the inclusion criteria described below were selected by medical record review and in a consecutive manner. 

At their initial diagnosis, each patient received a comprehensive ophthalmologic examination including a review of their medical history and measurement of their best-corrected visual acuity in order to confirm that their VA was adequate for the performance of automated perimetry, slit lamp biomicroscopy, autorefractometry, autokeratometry, axial length measurement by intraocular lens master (Carl Zeiss; Meditec, Dublin, CA), IOP measurement by Goldmann applanation tonometry (GAT), gonioscopy, a dilated fundoscopic examination using a 90- or 78-diopter (D) lens, stereoscopic optic disc photography, repeated visual field (VF) examination by standard automated perimetry (Swedish interactive threshold algorithm standard strategy 24-2; Carl Zeiss; Meditec), and central corneal thickness (CCT; DGH-550 instrument; DGH Technology Inc., Exton, PA) assessment. 

For inclusion in the study, all patients had to meet the following criteria at baseline examination: a best-corrected visual acuity of 20/30 or better and the presence of a normal anterior chamber and open-angle on slit-lamp and gonioscopic examinations. Any patient with any other ophthalmic or neurologic condition that could have resulted in a VF defect or those with a history of diabetes mellitus were excluded. If surgical or laser treatment was performed during follow-up, only data obtained during the period before such treatment were analyzed. 

Diagnosis of glaucoma was based on the presence of typical glaucomatous optic disc changes, that is, disc excavation, diffuse or focal neural rim thinning, disc hemorrhage or retinal nerve fiber layer (RNFL) defects, and glaucomatous VF defects as agreed upon by two glaucoma experts (JHN, KRS). Eyes with glaucomatous VF defects were defined as those that met at least two of the following criteria, as confirmed by more than two reliable consecutive test results, in addition to compatibility with their optic nerve appearance: (1) a cluster of three points with a probability of less than 5% on a pattern deviation map in at least one hemifield and 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 (GHT) result outside the normal limits; and/or (3) a pattern standard deviation (PSD) outside of 95% of the normal limits. Reliable VF assessment was defined as a VF test with a false-positive error rate of <15%, a false-negative error of <15%, and a fixation loss of <20%. To minimize any learning effect, the first VF test was excluded from analysis. All participants required at least 2 years of follow-up from their baseline examination. One eye was randomly chosen for analysis if both eyes met inclusion criteria. If the patient clinic identification number was an odd one, the right eye was selected, and a left eye was selected for an even number. 

Among the eligible study subjects, those with a history of uneventful RCS for the treatment of myopia and with a diagnosis of glaucoma in at least one eye, diagnosed for the first time at their baseline examination at our glaucoma clinic, were categorized as the RCS group. Study subjects with no history of RCS were categorized as the non-RCS group. 

All procedures conformed to the principles of the Declaration of Helsinki, and the study was approved by the Institutional Review Board of the Asan Medical Center at the University of Ulsan, Seoul, Korea. 

Glaucoma Progression Assessment

Because glaucomatous damage involves both structural and functional aspects and these two aspects may not appear at the same period, glaucoma progression was determined by either a structural or functional measure. Structural progression was assessed using stereoscopic optic disc and red-free RNFL photographs. Serial stereoscopic photographs were displayed on an LCD monitor. Two glaucoma experts (KRS and JHN) independently assessed all photographs in order to estimate glaucoma progression between patients' first and last visits. Either grader was masked to the progression assessments rendered by the other 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 each 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 an RNFL defect, and/or appearance of new disc hemorrhage. Each grader classified each glaucomatous eye as either stable or progressing. If the opinions of the two observers differed, a third examiner (YJK) made the final decision. 

Functional progression was determined using two VF methods, that is, event-based analysis (EA) and trend-based analysis (TA). To conduct EA, commercial software (guided progression analysis; Carl Zeiss Meditec) was used. VF EA progression was defined as a significant deterioration from the baseline pattern deviation at three or more of the same test points and was evaluated on three consecutive examinations. ¹⁰ The other VF progression criterion used linear TA regression, using visual field index data. A significantly negative slope (P < 0.05) indicated VF TA progression. 

Analysis

Baseline characteristics were compared between the RCS group and the non-RCS group. Among non-RCS group, age-matched non-RCS group subjects were selected and also compared with the RCS group subjects. The Wilk-Shapiro test was used to explore the distribution of numerical data. Normally distributed data are presented as means with standard deviations (SDs), whereas non-normally distributed data are shown as medians with interquartile ranges. Normally distributed data were compared between the two groups by using the unpaired t-test. Non-normally distributed data were compared using the Mann-Whitney U test. To compare categorical data, the chi-squared test was used. Hazard ratios (HRs) for associations among potential risk factors, including a history of RCS and glaucoma progression based on optic disc/RNFL/VF examinations, were obtained using Cox's proportional hazards model. Univariate analyses were performed separately for each variable. Variables with a probability value of ≤0.20 in univariate analysis were considered significant and were included in a multivariate Cox proportional hazards model. A backward elimination process was used to develop the final multivariate model, and adjusted HRs with 95% confidence intervals (CIs) were calculated. Schoenfeld residuals and the log [-log (survival rate)] test were used to verify proportional hazards assumptions were not violated. Model fit was assessed using residual analyses. In each RCS and non-RCS group subject, uni- and multivariate Cox proportional hazards analysis was perform to evaluate risk factors. All statistical analyses were performed using SAS version 9.1 software (SAS Institute Inc., Cary, NC) and SPSS version 15.0 software (SPSS Inc., Chicago, IL). 

Results

A total of 313 eyes of 313 study subjects were included in our study. All patients were Korean, the mean (±SD) follow-up period was 2.7 (±0.4) years, and the VF deviation (MD) was −3.2 (±5.5) dB. Mean number of analyzed VF examinations was 4.7 (±1.2). For the optic disc/RNFL photographs assessment, two experts showed the same opinion in 274 eyes (87.5%) in terms of progression determination. Among those eyes studied, 47 eyes (15.0%) showed progression by optic disc or RNFL photographic assessment, 61 eyes (19.5%) by VF analysis, and 21 eyes (6.7%) by both methods. Thus, 87 (27.8%) eyes showed progression by either optic disc/RNFL photographic or VF assessment during the follow-up period. Among those 313 eyes, 34 (10.9%) eyes had a history of RCS and were thus classified as the RCS group. The remaining 279 eyes were classified as the non-RCS group. The time interval between RCS and glaucoma diagnosis ranged from 1 to 15 years (median, 5 years). All study subjects indicated that the RCS treatment had fully corrected their myopia and that they had not experienced IOP elevation after RCS for any other reason. 

Among the 34 eyes of the RCS group, 10 (29.4%) eyes progressed by either experts' optic disc and RNFL assessment or by VF. Among the 279 eyes of the non-RCS group, 77 (27.6%) eyes progressed during the follow-up period. The prevalence of progression did not differ between the two groups (P = 0.482). The mean patient age was significantly higher in the non-RCS group (P < 0.001). The mean VF MD, PSD, sex, spherical equivalent (SE), and percentage of treatments performed during the follow-up period also did not differ between the two groups (P = 0.595, 0.416, 0.467, 0.682, and 0.146, respectively). However, baseline and follow up IOP, axial length (AL), and CCT were significantly different between the two groups (all, P < 0.001). Comparative data for the RCS and non-RCS groups are shown in Table 1. 

Table 1. 

 

Comparison between the RCS Group and the Non-RCS Group

Parameter	RCS Group (n = 34)	Non-RCS Group (n = 279)	Age-Matched Non-RCS Group Subjects (n = 51)	*P Value	†P Value
Age (y)	41.5 ± 11.3	56.7 ± 11.5	42.2 ± 7.3	<0.001	0.733
Sex (male/female)	16/18	138/141	30/21	0.467	0.199
VF MD (dB)	−3.6 ± 4.9	−3.1 ± 5.5	−3.1 ± 5.1	0.595	0.646
VF PSD (dB)	4.3 ± 4.7	3.8 ± 3.8	4.0 ± 3.7	0.416	0.676
SE (D, range)	−0.9 ± 0.7 (−0.31 to −3.25)	−0.8 ± 0.7 (0.75 to −4.25)	−1.7 ± 2.0 (−0.50 to −5.25)	0.682	0.028
Baseline IOP (mm Hg)	12.4 ± 2.8	17.2 ± 4.7	16.6 ± 4.3	<0.001	<0.001
CCT (μm)	483.7 ± 35.6	540.3 ± 30.5	539.8 ± 36.0	<0.001	<0.001
AL (mm)	26.4 ± 2.8	23.4 ± 1.2	24.0 ± 1.2	<0.001	<0.001
Prevalence of progression (%)	10 eyes (29.4)	77 eyes (27.6)	15 eyes (29.4)	0.482	0.598
Percentage of treatment (%)	90.0	82.4	82.0	0.146	0.527
Percentage of trabeculectomy (%)	2.94	1.79	3.92	0.501	0.650
No. of IOP lowering medication used in each eye	0.94 ± 0.60	0.93 ± 0.47	0.90 ± 0.56	0.852	0.772
Mean follow-up IOP (mm Hg)	10.1 ± 2.6	14.4 ± 2.4	13.9 ± 2.9	<0.001	<0.001

 dB, decibel.

*  Comparison between the RCS group and non-RCS group.

†  Comparison between RCS group subjects and age-matched non-RCS group subjects.

Among the 279 eyes of the non-RCS group, 51 eyes formed age-matched non-RCS group. Except for baseline and follow-up IOP, AL, and CCT, all parameters including the prevalence of progression were not significantly different between the two groups (Table 1). 

Among the 34 eyes in the RCS group, 5 had been treated with PRK, 21 eyes with LASIK, and 8 eyes with LASEK. Among the 21 eyes treated with LASIK, 6 eyes showed progression, while 4 eyes deteriorated among the 13 eyes treated with non-LASIK (P = 0.594). 

Table 2 shows univariate and multivariate HRs for each putative risk factor, including the history of RCS in all study participants according to the Cox proportional hazard model. As baseline age, sex, VF MD, PSD, SE, AL, and any history of RCS were found to be predictive of glaucoma progression according to univariate analysis (variables with a probability of <0.2), they were included in the multivariate analysis. In multivariate analysis using a backward elimination method, baseline age and PSD (HRs, 1.013, and 1.119; and P = 0.026, and P < 0.001, respectively) were significant. 

Table 2. 

 

Univariate and Multivariate Cox Proportional Hazards Model with Backward Elimination for Predicting Glaucoma Progression in All Participants

Analysis	HR	95% CI	P Value
Univariate
Age (y)	1.008	0.998–1.019	0.127
Sex (male/female)	1.267	0.963–1.666	0.091
VF MD (dB)	0.934	0.919–0.949	<0.001
VF PSD (dB)	1.119	1.091–1.147	<0.001
SE (D)	0.936	0.873–1.003	0.062
Baseline IOP (mm Hg)	1.000	0.997–1.003	0.840
CCT (μm)	0.997	0.980–1.013	0.696
AL (mm)	0.942	0.887–1.001	0.053
History of RCS	1.818	1.213–2.726	0.004
Multivariate
Age (y)	1.013	1.002–1.025	0.026
Sex (male/female)	1.245	0.944–1.642	0.121
VF MD (dB)	0.925	0.837–1.024	0.134
VF PSD (dB)	1.119	1.090–1.149	<0.001
SE (D)	0.917	0.799–1.051	0.213
AL (mm)	0.985	0.954–1.016	0.343
History of RCS	1.245	0.944–1.642	0.121

 CI, confidence interval; dB, decibel.

Cox proportional hazard analysis was performed in each RCS and non-RCS group. In the RCS group, baseline VF MD, PSD, and AL were significant in univariate analysis; however, only VF PSD was significant in multivariate analysis (HR, 1.193; P < 0.001 (Table 3). In the non-RCS group, only VF PSD was significant in multivariate analysis (HR, 1.099; P < 0.001) (Table 4). 

Table 3. 

 

Univariate and Multivariate Cox Proportional Hazards Model with Backward Elimination for Predicting Glaucoma Progression in the RCS Group

Analysis	HR	95% CI	P Value
Univariate
Age (y)	0.985	0.954–1.017	0.340
Sex (male/female)	0.645	0.332–1.291	0.215
VF MD (dB)	0.866	0.829–0.906	<0.001
VF PSD (dB)	1.193	1.119–1.272	<0.001
SE (D)	1.161	0.628–2.149	0.634
Baseline IOP (mm Hg)	0.824	0.858–1.021	0.112
CCT (μm)	1.002	0.990–1.013	0.794
AL (mm)	0.803	0.659–0.977	0.028
Multivariate
VF MD (dB)	0.946	0.871–1.004	0.089
VF PSD (dB)	1.193	1.119–1.272	<0.001
Baseline IOP (mm Hg)	0.998	0.977–1.012	0.483
AL (mm)	0.941	0.883–1.002	0.125

 CI, confidence interval; dB, decibel.

Table 4. 

 

Univariate and Multivariate Cox Proportional Hazards Models with Backward Elimination for Predicting Glaucoma Progression in the Non-RCS Group

Analysis	HR	95% CI	P Value
Univariate
Age (y)	1.005	0.993–1.017	0.431
Sex (male/female)	1.477	1.093–1.994	0.011
VF MD (dB)	0.947	0.930–0.964	<0.001
VF PSD (dB)	1.102	1.071–1.134	<0.001
SE (D)	0.942	0.881–1.008	0.085
Baseline IOP (mm Hg)	0.981	0.954–1.010	0.199
CCT (μm)	0.996	0.989–1.011	0.196
AL (mm)	1.077	0.964–1.203	0.188
Multivariate
Sex (male/female)	1.112	1.004–1.232	0.061
VF MD (dB)	0.972	0.804–1.175	0.765
VF PSD (dB)	1.099	1.068–1.130	<0.001
SE (D)	0.953	0.864–1.05	0.329
Baseline IOP (mm Hg)	0.984	0.954–1.015	0.314
CCT (μm)	0.925	0.837–1.024	0.133
AL (mm)	1.275	0.962–1.689	0.091

 CI, confidence interval; dB, decibel.

Discussion

RCS has become popular worldwide for the treatment of myopia. An increasing number of patients with myopia are expected to receive RCS in the future. Therefore, glaucoma is frequently diagnosed in some RCS candidates upon preoperative examination, and it is also seen in some subjects with a history of RCS. Although concerns regarding possible inconsistencies in IOP measurements conducted before and after RCS or IOP elevation upon steroid use after RCS have been raised, ^{1–8,11–14} the progression of glaucoma has not been studied in glaucoma patients who had previously undergone RCS. 

As expected, mean CCT, AL, and baseline IOP values differed significantly between the RCS and non-RCS group. As all IOP measurements were performed using GAT and GAT measurement significantly depends on CCT, a lower IOP reading can be attributed to the presence of a thinner and flatter cornea after RCS. Lower IOP GAT readings after RCS have been reported in previous studies. ^1–8 Patient age was significantly lower in the RCS group, which may have been due to the fact that RCS is usually performed in younger myopic patients. In the mean time, baseline VF severity did not differ between the two patient groups in our study. 

For all study participants, we performed Cox proportional hazard analysis to identify the risk factors related to glaucoma progression. In univariate analysis, age, sex, VF MD, PSD, history of RCS, SE, and AL were suggested as possible risk factors, showing a P value less than 0.2. Those variables were thus integrated in the multivariate analysis. Interestingly, only age and VF PSD were related to glaucoma progression, as seen in multivariate analysis. A history of RCS, SE, and AL may be correlated with each other, and those variables were found to be significant in univariate analysis but not in multivariate analysis. Myopia is also known as a risk factor for the development of glaucoma. ^15–20 However, it is not certain whether myopia is a risk factor for the progression of glaucoma. In all of our study participants enrolled in the university-based glaucoma clinic, myopia was not a risk factor for progression of glaucoma. 

In previous studies, aging has also been reported as a risk factor for glaucoma. ^21–26 Our results also showed that age was a significant risk factor for glaucoma progression when assessed in all of our study participants. When we assessed the risk factors for progression in the RCS and non-RCS groups, interestingly, baseline VF PSD was the only risk factor seen in both groups. Baseline VF severity has been reported as a significant risk factor in previous studies. ^27,28 Therefore, our results coincided with such previous outcomes. 

In fact, to know exactly the effect of RCS on glaucoma progression, prospective evaluation should include performing RCS in patients with glaucoma; however, such a clinical trial would be difficult to perform in actual clinical practice because conducting RCS on eyes with a proven pathology such as glaucoma may not be acceptable under many circumstances. Therefore, it may be difficult to study the influence of RCS on the progression of glaucoma by using prospective long-term follow-up. Therefore, we consecutively enrolled all eligible glaucoma patients and determined whether RCS history could be a risk factor for progression while incorporating each putative risk factor. As a result, RCS history was not seen as a risk factor in that analysis. When we divided our total number of participants into the RCS and non-RCS groups, the groups did not show a significant difference in terms of the prevalence of progression. Because age was a significant risk factor for progression and non-RCS group subjects were older than RCS subjects, we performed subgroup analysis. The prevalence of progression was not significantly different between age-matched the non-RCS group and RCS group. Therefore, according to the retrospective analysis of our average 2.7-year follow-up, RCS history was not a significant risk factor for glaucoma progression. One factor that should be considered here is that as most of our study participants had been treated with antiglaucoma medication (RCS group, 90.0%; non-RCS group, 82.4%; P = 0.146), we propose that patients with a history of glaucoma treatment and an RCS did not show a significant difference in terms of glaucoma progression compared with that for patients without history of an RCS. As far as we could determine, as this is the first report which evaluated glaucoma progression in patients with a history of RCS, we could not compare our result with those of other studies. 

Among several RCS types, the LASIK procedure poses the possibility of optic nerve damage during acute IOP elevation, when the suction ring of a microkeratome is placed. ⁹ Hence, we divided our RCS patients into two groups, those who underwent LASIK and those who underwent non-LASIK. However, those two groups did not show significant differences in terms of prevalence of glaucoma progression. 

Glaucoma can progress structurally and functionally. Glaucomatous damage may be generalized or localized. Accordingly, detection of progression can be variable, depending on the detection strategy used, and this problem is well-recognized. ^29–35 We used several relevant clinical criteria to define progression. Upon expert examination of photographs of the optic disc and RNFL or use of VF-guided progression analysis or by VF index TA, 87 eyes (27.8%) in our study were determined to have progressed. 

The present study had several limitations including a short follow-up period and a small sample size. We used disc and RNFL photography or VF analysis as the reference standards, and these two tests may not be perfect in terms of detecting progression, although both methods currently serve as reference standards. Also, as all glaucomatous eyes were Asian (Korean), our data may therefore not be automatically generalizable to other races. All patients previously underwent RCS at other clinics at some time in the past, and they could not provide pre-RCS medical records, although we tried to include them. Thus, this should be admitted as another limitation of current study. 

In summary, regarding the clinical course of glaucoma patients with a history of RCS, we may conclude that those eyes did not show a significant difference in terms of glaucoma progression compared with eyes with no history of RCS. Baseline VF PSD and age substantially affected glaucoma progression in all of our study participants. Therefore, baseline VF PSD was a risk factor for progression in both the RCS and the non-RCS groups. 

Footnotes

 Disclosure: Y.J. Kim, None; S.-C. Yun, None; J.H. Na, None; H.W. Tchah, None; J.J. Jung, None; K.R. Sung, None