Frequency and associated factors of structural progression of open-angle glaucoma in the Beijing Eye Study
- 1Beijing Institute of Ophthalmology, Beijing Tongren Hospital, Capital University of Medical Science, Beijing, China
- 2Department of Ophthalmology, 306th Hospital of PLA, Beijing, China
- 3Department of Ophthalmology, Medical Faculty Mannheim of the Ruprecht-Karls-University Heidelberg, Germany
- Correspondence to Dr Professor Liang Xu, Beijing Institute of Ophthalmology, 17 Hougou Lane, Chong Wen Men, 100005 Beijing, China;
Contributors All the authors were involved in: (1) Substantial contributions to conception and design, acquisition of data, or analysis and interpretation of data; (2) Drafting the article or revising it critically for important intellectual content; and (3) Final approval of the version to be published.
- Accepted 12 February 2012
- Published Online First 9 March 2012
Background To examine factors associated with progression of open-angle glaucoma in a population-based setting.
Methods The population-based Beijing Eye Study, which included 4439 subjects with an age of 40+ years in the year 2001, was repeated in 2006. Optic disc photographs of the baseline examination versus follow-up examination were compared.
Results Out of 111 open-angle glaucoma patients examined in 2001, 77 (69%) subjects participated in the follow-up examination and 16 (21%) eyes showed glaucoma progression. Glaucoma progression was associated with smaller rim area (p=0.001), larger β zone (p=0.037), higher frequency of β zone increase during follow-up (p=0.01), higher prevalence of disc haemorrhages (p=0.01) and higher single intraocular pressure (p=0.04). In multiple regression analysis, only smaller rim area remained significantly associated with glaucoma progression. Glaucoma progression was not associated with optic disc size (p=0.70), mean blood pressure (p=0.43), ocular perfusion pressure (p=0.96), retinal vessel diameter and retinal microvascular abnormalities (all p>0.10), prevalence of diabetes mellitus (p=0.75) and arterial hypertension (p=0.26), prevalence of dyslipidaemia (p=0.28), refractive error (p=0.69), and central corneal thickness (p=0.97).
Conclusions In a population-based setting in adult Chinese, factors showing an association with open-angle glaucoma progression were an advanced stage of the disease (ie, small rim), presence of disc haemorrhages, larger area of β zone and higher frequency of β zone increase, and higher intraocular pressure. Glaucoma progression was not significantly associated with optic disc size, central corneal thickness, retinal vessel diameter and retinal microvascular abnormalities, and systemic diseases such as diabetes mellitus and arterial hypertension.
- Glaucoma progression
- open-angle glaucoma
- neuroretinal rim
- intraocular pressure
- Beijing Eye Study
- public health
- optic nerve
Risk factors for the progression of glaucomatous optic neuropathy have been assessed in numerous large longitudinal investigations, such as Ocular Hypertension Treatment Study, the Collaborative Initial Glaucoma Treatment Study, the Advanced Glaucoma Intervention Study, the Early Manifest Glaucoma Trial, the Canadian Glaucoma Study and others.1–7 These studies revealed several risk factors for glaucoma progression, such as the level of intraocular pressure, presence of pseudoexfoliation of the lens and advanced stage of glaucoma. Most of these studies, however, were hospital-based investigations, and none of them was performed on Chinese patients. We, therefore, performed a longitudinal population-based study on Chinese with open-angle glaucoma.
The Beijing Eye Study is a population-based prospective cohort study in Northern China which has been described in detail recently.8 The Medical Ethics Committee of the Beijing Tongren Hospital approved the study protocol. The baseline examination carried out in the year 2001 consisted of 4439 individuals (2505 women) with a mean age of 56.2±10.6 years (range 40–101 years). The examinations included refractometry, frequency-doubling perimetry, non-contact pneumotonometry, slit lamp assisted biomicroscopy of the anterior segment, and photography of the cornea, lens, macula and optic disc after medical dilatation of the pupil. The optic disc photographs were digitalised and the optic disc structures were measured using planimetric software program. As described recently,9 the magnification by the optic media of the eye was corrected according to Littman's method taking into account the refractive error. In a second step of the examination, the width of the neuroretinal rim and the diameters of the optic cup and optic disc were measured in the vertical meridian. The vertical cup to disc diameter ratio and the optic cup area were calculated. The neuroretinal rim area resulted as difference of disc area minus cup area. Parapapillary chorioretinal atrophy was divided into a peripheral α zone and a central β zone at the optic disc border as described in detail previously.10 Additionally, retinal microvascular abnormalities (focal and generalised arteriolar narrowing, arteriolar sheathing, arteriovenous crossing abnormalities in four fundus regions) were assessed,11 and the diameters of the arteries and veins of the four main retinal vessel trunks were measured at the optic disc border.
Glaucoma was defined according to the criteria of the International Society of Geographic and Epidemiological Ophthalmology ISGEO.12 As described in detail previously,8 the whole glaucoma group was then differentiated into subjects with open-angle glaucoma and subjects with angle-closure glaucoma. In the year 2006, a follow-up examination was carried out, reinviting all participants of the 2001 survey. In the year 2006, additional examinations were carried out including an optical coherence tomography of the anterior segment, measurement of fasting blood concentrations of glucose, low-density lipoproteins, high-density lipoproteins and cholesterol, and measurement of body height and weight and blood pressure. We defined diabetes mellitus as any plasma glucose concentration ≥7.0 mmol/l or by a self-reported history of a physician diagnosis of diabetes mellitus or by a history of drug treatment for diabetes (insulin or oral hypoglycemic agents). Arterial hypertension was defined as a systolic blood pressure ≥140 mm Hg and/or a diastolic blood pressure ≥90 mm Hg, and/or self-reported current treatment for arterial hypertension with antihypertensive medication. Ocular perfusion pressure was calculated as 2/3×(mean arterial blood pressure−intraocular pressure) with the mean arterial blood pressure=diastolic blood pressure+1/3×(systolic blood pressure−diastolic blood pressure). Dyslipidaemia was defined as any of hypercholesterolaemia (total cholesterol concentration ≥5.72 mmol/l (220 mg/dl)) or hypertriglyceridemia (triglyceride concentration ≥1.70 mmol/l (150 mg/dl)) or low high-density lipoprotein-cholesterol (concentration ≤0.91 mmol/l (35 mg/dl)) and/or a positive history for dyslipidaemia.
For the assessment of a progression of glaucomatous optic neuropathy, the photographs of the same eye taken at baseline and taken at the follow-examination were mixed and qualitatively examined in a masked fashion by a panel of experienced examiners (including LX, YXW, JBJ). A progression of glaucoma was defined as a smaller neuroretinal rim at the follow-examination compared with the baseline examination. Appearance of the parapapillary atrophy and presence of a disc haemorrhage were not taken into account. The rationale for not taking visual field progression as the main criterion to define glaucoma progression was that more than two psychophysical examinations may be necessary to reliably detect a progression of glaucoma.
Selection criteria for inclusion into the present study was the diagnosis of open-angle glaucoma in 2001 and the availability of optic disc photographs from 2001 to 2006.
Statistical analysis was performed using SPSS (SPSS for Windows, V.19.0, SPSS). Continuous data were presented as mean±SD. χ2 Tests and modified χ2 tests were used to compare proportions. Multivariate regression models were used to investigate the associations with the progression of glaucoma. For continuous variables, β coefficients (β) were reported; for categorical variables, ORs were presented and 95% CIs were described. All p values were two-sided and were considered statistically significant when the values were <0.05.
In the baseline examination in 2001, optic disc photographs were available for 4315 (97.2%) subjects (8458 eyes), with a mean age of 55.8±10.3 years (median, 56 years; range, 40–101 years) and a mean refractive error of −0.38±2.23 dioptres (median: 0 dioptres; range: −20.13 dioptres to +7.50 dioptres). The mean intraocular pressure was 16.1±3.4 mm Hg, with the 97.5th percentile at 26 mm Hg. Using the criteria of the International Society of Geographical and Epidemiological Ophthalmology, glaucoma was detected in 158 subjects (prevalence: 3.7%±0.3% (95% CI 3.1% to 4.2%)).12 Out of the 158 glaucoma subjects, 111 subjects (2.6%, 95% CI 2.1% to 3.0%) fulfilled the criteria of open-angle glaucoma, and 44 subjects (1.0%, 95% CI 0.7% to 1.3%) fulfilled the criteria of primary angle-closure glaucoma. In three subjects (0.07%, 95% CI 0% to 0.1%) with secondary glaucoma, glaucoma was considered to be due to uveitis (n=1 subject), due to a preceding pars plana vitrectomy which was performed because of proliferative diabetic retinopathy (n=1), or it was lens-related (n=1).
Out of the 111 subjects with open-angle glaucoma in 2001, 9 (8%) patients had died and 24 (22%) patients were alive but did not reparticipate in the study. The reason not to return to the examination in 2006 for the majority of the 24 patients was that they had moved away. Some patients were not available at the days of examinations in 2006 and some simply refused to reparticipate in the survey. For one (1%) subject, the quality of the optic disc photographs taken in 2006 was not sufficient for an examination, and so 77 (69%) subjects returned for the follow-up in the year 2006 and had examinable optic disc photographs. Out of the 77 patients, 13 (17%) patients were on treatment in the year 2006, with four patients under medical treatment and nine patients had undergone surgical or laser surgery. Mean age was 59.4±10.6 years (median 61 years; range 40–82 years). Mean refractive error was −1.24±3.66 dioptres (median: 0.00 dioptres; range: −16.25 to +5.88 dioptres). The subjects who had the diagnosis of glaucoma in 2001 and who did not participate in the follow-up examination 2006 and the study participants who had the glaucoma diagnosis in 2001 and who participated in the survey of 2006 did not differ significantly in age (61.7±11.5 years vs 59.4±10.6 years; p=0.31), refractive error (−1.21±2.91 dioptres vs −1.24±3.66 dioptres; p=0.97), gender (p=0.84), intraocular pressure (16.3±3.0 mm Hg vs 16.7±3.5 mm Hg; p=0.48) and neuroretinal rim area (1.56±0.80 mm2 vs 1.58±0.62 mm2; p=0.87).
Out of the 77 subjects involved in the present study, 16 (21%) subjects (10 men) showed a progression of glaucoma. In univariate analysis, progression of glaucoma was associated with smaller neuroretinal rim area (p=0.001), larger optic cup area (p=0.004), higher cup to disc diameter ratios (p=0.003), higher frequency of an increase of β zone of parapapillary atrophy (p=0.01), larger area of β zone (p=0.037), higher prevalence of disc haemorrhages (p=0.01), higher concentration of low-density lipoproteins (p=0.04) and higher single intraocular pressure measurement (p=0.04). The association between glaucoma progression and higher age (p=0.06) and higher intraocular pressure measurement (p=0.056) did not reach the level of statistical significance. The glaucoma progressive group and the glaucoma stable group did not vary in optic disc size (p=0.70), mean blood pressure (p=0.43), ocular perfusion pressure (p=0.96), retinal vessel diameter and retinal microvascular abnormalities (all p>0.10), prevalence of diabetes mellitus (p=0.75) and arterial hypertension (p=0.26), prevalence of dyslipidaemia (p=0.28), refractive error (p=0.69), central corneal thickness (p=0.97), and anterior chamber depth (p=0.26) (table 1).
A multiple binary regression analysis was performed, which included the progression of glaucoma as the dependent parameter, and which included as independent parameters all variables, for which the p value of the association with glaucoma progression was ≤0.10. It revealed that glaucoma progression remained to be significantly associated with smaller neuroretinal rim area (p=0.016; regression coefficient: −2.08; exponent B: 0.13 (95% CI 0.02 to 0.68)) and marginally associated with presence of disc haemorrhages (p=0.08; regression coefficient: 2.41; exponent B: 11.1 (95% CI 0.76 to 163)).
Using the criteria of the International Society of Geographic and Epidemiological Ophthalmology ISGEO for the definition of glaucoma,12 we found in our population-based study that glaucoma progression was significantly more likely in patients with a smaller neuroretinal rim (or larger optic cup) as surrogate for a more advanced stage of glaucoma, in patients with disc haemorrhages and in patients with a large β zone of parapapillary atrophy at baseline. Among other parameters, optic disc size, central corneal thickness, and presence of major systemic diseases such as diabetes mellitus, dyslipidaemia and arterial hypertension, and presence of major ocular disorders such as age-related macular degeneration were not significantly associated with glaucoma progression (table 1).
The results of the present study confirm a previous report of the Ocular Hypertension Treatment Study in which the occurrence of a disc haemorrhage increased the risk of developing primary open-angle glaucoma 3.7-fold in a multivariate analysis.2 Confirming also other hospital-based studies,13 14 it shows the importance of disc haemorrhages to indicate the risk of glaucoma progression. Since disc haemorrhages can usually not be detected on confocal laser scanning images, it also shows the importance of a clinical ophthalmoscopic examination in addition to imaging techniques for the assessment of glaucomatous optic neuropathy.
Our population-based study confirms several previous hospital-based studies in that a large β zone of parapapillary atrophy at baseline indicates an increased risk for further glaucoma progression.6 14–17 The pathogenic reasons for the association between β zone and glaucomatous optic neuropathy have remained, however, unclear. As in the Advanced Glaucoma Intervention Study and other investigations,1 3–6 17 our study confirms that an advanced stage of glaucoma (ie, a small neuroretinal rim) is a risk factor for further progression of the disease, after adjustment for intraocular pressure. Although the reason for this finding has remained unclear so far, histological studies have suggested that morphologic changes in the optic nerve head including a thinning of the lamina cribrosa and a shortening of the distance between the intraocular compartment and the compartment of the orbital cerebrospinal fluid pressure may play a role.18 The Beijing Eye Study agrees with the previous studies also in that a high intraocular pressure was a risk factor, and that optic disc size was not a risk factor for glaucoma progression.6 19
In the Beijing Eye Study, central corneal thickness was not associated with glaucoma progression. It disagrees with previous hospital-based studies in which central corneal thickness was a predictor for the development of primary open-angle glaucoma and the progression of glaucomatous visual field defects.4 14 20–22 In the Ocular Hypertension Treatment Study, central corneal thickness was a powerful predictor for the development of primary open-angle glaucoma.20 In a study by Medeiros and colleagues, a low central corneal thickness was a risk factor for development of visual field loss among patients with preperimetric glaucoma.21 In a recent study by De Moraes and coworkers, a thinner central corneal thickness was a risk factor for progression of treated glaucoma.14 The results of our study are, however, in agreement with results of hospital-based studies, in which corneal thickness was unrelated to the development of visual field defects.22 23 A peculiarity of our population-based study is that the majority of the glaucoma patients were untreated so that a potential bias in the amount of therapy, introduced by falsely low intraocular pressure readings, was less likely than in hospital-based studies. It has been discussed that the potential influence of corneal thickness on treatment and referral decisions may have influenced the findings obtained in hospital-based studies on the association between central corneal thickness and progression of glaucomatous optic neuropathy.23
Potential limitations of our study should be mentioned. First, the number of glaucoma patients included into the study was relatively small, so that the statistical power to detect difference was limited. It may have been the reason why, in contrast to previous studies,2 4 14 higher age in our study was not statistically significantly associated with glaucoma progression. The relatively small number of glaucoma patients in our study may also have been the reason why in multiple regression analysis, only smaller rim area remained significantly associated with glaucoma progression. Second, as in any population-based study, non-participation can be of major concern. Third, only two intraocular pressure measurements were available (one from 2001 and one from 2006). It may have been the reason, why, in contrast to previous studies,1–4 7 14 higher intraocular pressure in our study was not statistically significantly associated with glaucoma progression. Fourth, pseudoexfoliation was not assessed, so that we could not differentiate between primary open-angle glaucoma and secondary open-angle glaucoma. We could also not assess whether pseudoexfoliation is an additional risk factor for glaucoma progression.4 Fifth, as in other population-based studies that have used the ISGEO scheme to diagnose glaucoma, a small proportion of subjects with primary macrocups may have been misclassified as glaucomatous (without change during the follow-up). Sixth, progressive cataract may have partially obscured the optic disc image at the follow-up examination and may have confounded the comparison of the optic nerve appearance. The strengths of our study include that it was a population-based investigation avoiding the risks of a potential bias due to the referral of the patients; that it is the first study on glaucoma progression for the Chinese on a population base; and that most of our patients were untreated, so that the dependence of intraocular pressure readings on corneal thickness did not introduce a bias by influencing the intensity of antiglaucomatous therapy. It may be one of the reasons why central corneal thickness was not related with glaucoma progression in our study.
In conclusion, in a population-based setting in adult Chinese, factors showing an association with open-angle glaucoma progression were an advanced stage of the disease (ie, small neuroretinal rim), presence of disc haemorrhages, larger area of β zone of parapapillary atrophy and higher frequency of an increase of β zone and higher intraocular pressure. Glaucoma progression was not significantly associated with optic disc size, central corneal thickness, retinal vessel diameter and retinal microvascular abnormalities, and systemic diseases such as diabetes mellitus and arterial hypertension.
Funding This research received a grant from the National Key Laboratory Fund, Beijing, China.
Competing interests None.
Ethics approval Approval provided by theMedical Ethics Committee of the Beijing Tongren Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.