Prevalence of glaucoma by number of metabolic abnormalities.
Tan GS, Wong TY, Fong C, Aung T. Diabetes, Metabolic Abnormalities, and GlaucomaThe Singapore Malay Eye Study. Arch Ophthalmol. 2009;127(10):1354-1361. doi:10.1001/archophthalmol.2009.268
To examine the relationship of diabetes mellitus and metabolic abnormalities with intraocular pressure and glaucoma.
A population-based study was conducted in 3280 (78.7% response) Malay adults aged 40 to 80 years. Diabetes was defined as a random serum glucose level of 200 mg/dL or greater or physician diagnosis of diabetes mellitus. Metabolic abnormalities including body mass index, lipid levels, and blood pressure were measured. Glaucoma was defined from a standardized examination by means of the International Society for Geographical and Epidemiological Ophthalmology criteria.
There were 764 persons (23.3%) who had diabetes. After controlling for age, sex, education, smoking, central corneal thickness, and diabetes treatment, intraocular pressure was higher in persons with than without diabetes (16.7 vs 15.0 mm Hg, P < .001) and in those with higher serum glucose levels (P < .001), glycosylated hemoglobin concentrations (P < .001), total cholesterol levels (P = .001), triglyceride levels (P = .002), and body mass index (P = .001). However, the prevalence of glaucoma was similar between persons with and without diabetes (4.7% vs 4.5%). In multivariate logistic regression models adjusting for age, sex, education, smoking, central corneal thickness, and diabetes treatment, diabetes was not associated with glaucoma (odds ratio, 1.00; 95% confidence interval, 0.63-1.61).
These data suggest that, although diabetes and metabolic abnormalities may be associated with a small increase in intraocular pressure, they are not significant risk factors for glaucomatous optic neuropathy.
Glaucoma is a leading cause of irreversible blindness worldwide.1,2 Diabetes mellitus is known to cause microvascular damage and may affect vascular autoregulation of the retina and optic nerve. Diabetes has been found to be associated with elevated intraocular pressure (IOP)3- 6 and has therefore been suggested as a possible risk factor for glaucoma, particularly primary open-angle glaucoma (POAG). However, the current evidence to support this relationship remains conflicting. Four population-based studies reported a statistically significant association between diabetes and POAG,7- 10 but 7 others have not found a significant relationship.5,11- 16 Furthermore, in many studies there was no measurement of central corneal thickness (CCT), which has been shown to be different in people with diabetes.17 Thus, it is uncertain whether the relationship of diabetes and IOP is influenced by variations in CCT.
Other metabolic abnormalities, often occurring in a cluster referred to as the metabolic syndrome, may also have microvascular effects, but their relationship with glaucoma has not been well studied. Oh et al18 found insulin resistance to be associated with elevated IOP, but no other study has explored its relationship with glaucoma. Metabolic syndrome components include hypertriglyceridemia, low levels of high-density lipoprotein cholesterol, obesity, hypertension, and hyperglycemia. Of these, hypertension is a well-recognized risk factor for elevated IOP,6,19,20 although it does not have the same consistent association with glaucoma.3,19,21 In fact, low diastolic blood pressure has been suggested to be a risk factor for glaucoma in some studies.19 Obesity and elevated body mass index have been associated with higher IOP,22- 24 but again, there are no convincing data to support a direct association with glaucoma.25 The relationship between lipid abnormalities, IOP, and glaucoma has not been well characterized.
The purpose of this study was to examine the relationship between diabetes and metabolic syndrome abnormalities with IOP and glaucoma in a population-based study of Asian Malays. A particular aspect of this racial/ethnic group is that a high prevalence of diabetes has been previously reported.26,27
The Singapore Malay Eye Study was a population-based cross-sectional study of 3280 Malay individuals (78.7% response rate) aged 40 to 80 years in Singapore. The study method has been described previously28,29 and is summarized as follows. The sampling frame consisted of all Malays aged 40 to 80 years living in 15 residential districts across the southwestern part of Singapore. From an initial list of 16 069 Malay names provided by the Ministry of Home Affairs, an age-stratified random sampling procedure was used to select 5600 names (1400 people from each decade of 40-49, 50-59, 60-69, and 70-79 years). Of the 5600 names initially identified, 4168 participants (74.4%) were determined to be eligible to participate on the basis of the specified inclusion criteria.28 Of these, 3280 (78.7%) were examined in the clinic and the remaining 888 (21.3%) were classified as nonparticipants. Nonparticipants were more likely than participants to be in the oldest age group (70-79 years of age), but there was little difference in sex, sampling location, and telephone ownership between the 2 groups (data not shown).
Approval for the study protocol was granted by the hospital's institutional review board, and the study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants before enrollment.
All participants underwent a standardized interview, examination, and ocular imaging at a centralized study clinic.28,30 A structured slitlamp examination (model BQ-900; Haag-Streit AG, Koeniz, Switzerland) was performed by study ophthalmologists before and after pupil dilation. A Goldmann applanation tonometer (Haag-Streit AG) was used to obtain 1 reading of IOP from each eye before dilation; the reading of the right eye was used for analysis because there was no statistical difference between the mean readings of the left and right eyes. If the IOP reading was greater than 21 mm Hg, a repeat reading was taken and the second reading was used for analysis. The optic disc was examined after dilation through a 78-diopter lens at 10× magnification on the basis of the protocol used in the Tanjong Pagar Survey.31,32
Gonioscopy was performed at the first slitlamp examination for all participants with suspected glaucoma (see the next paragraph for definition) and for participants with a van Herick limbal chamber depth grade of less than 25% of corneal thickness. For participants meeting the glaucoma suspect criteria, a 24-2 Swedish Interactive Thresholding Algorithm static, threshold-related visual field examination was performed with near refractive correction before dilation by means of a perimetry device (Humphrey Visual Field Analyzer II, model 750; Carl Zeiss AG, Oberkochen, Germany). Test reliability was determined by the instrument's algorithm (fixation losses <20%, false-positive results <33%, or false-negative results <33%). Visual field testing was repeated once if deemed unreliable.
Glaucoma suspects were defined according to a set of prespecified criteria: IOP greater than 21 mm Hg, gonioscopic findings of closed or occludable angles, presence of peripheral anterior synechiae, cup-disc ratio (CDR) greater than 0.6, disc asymmetry with CDR greater than 0.2, abnormal deposits on pupil margin consistent with pseudoexfoliation syndrome, pigment deposition on the cornea consistent with pigment dispersion syndrome, and known glaucoma.31,32 Five CCT measurements were obtained from each eye with an ultrasound pachymeter (Advent; Mentor O & O, Norwell, Massachusetts), and the median reading was taken.
The diagnosis and classification of glaucoma cases have been reported previously33 and will be described briefly. “Glaucoma cases” were defined according to the International Society of Geographical and Epidemiological Ophthalmology criteria on the basis of 3 categories.31 Category 1 cases were defined as having optic disc abnormality (vertical CDR [VCDR]/VCDR asymmetry ≥97.5th percentile or neuroretinal rim (NRR) width between 11 and 1 o’clock or 5 and 7 o’clock <0.1 VCDR) and glaucomatous visual field defect. Category 2 cases were defined as having a severely damaged optic disc (VCDR or VCDR asymmetry ≥99.5th percentile) in the absence of an adequate visual field test. Assignment of category 1 or 2 glaucoma required that there be no other explanation for the VCDR finding (dysplastic disc or marked anisometropia) or visual field defect (retinal vascular disease, macular degeneration, or cerebrovascular diseases). Category 3 cases were defined as participants without visual field or optic disc data who were blind (corrected visual acuity, <3/60) and who had had previous glaucoma surgery or had IOP greater than the 99.5th percentile.31 A narrow anterior chamber angle was diagnosed if the posterior trabecular meshwork was seen for 180° or less of the angle circumference during static gonioscopy. Primary angle-closure glaucoma was defined as a narrow anterior chamber angle, features of trabecular obstruction by peripheral iris (peripheral anterior synechiae, elevated IOP, iris whirling, “glaukomflecken” lens opacities, or excessive pigment deposition on the trabecular surface), and evidence of glaucoma as defined earlier. Cases of POAG included those meeting the definition of glaucoma without any evidence of narrow angles, primary angle-closure glaucoma, or a secondary cause (eg, abnormal anterior segment deposits or iris neovascularization). Final definition, adjudication, and classification of glaucoma cases were reviewed by one of us (T.A.). Ocular hypertension was defined as the presence of IOP greater than 21 mm Hg in participants who did not meet the criteria for glaucoma.
Each participant underwent height and weight measurements, which were used to determine the body mass index (BMI; calculated as weight in kilograms divided by height in meters squared). Systolic and diastolic blood pressures were also taken with an automated sphygmomanometer.
A 40-mL sample of venous blood was collected to determine levels of serum lipids (total cholesterol, high-density lipoprotein cholesterol, direct low-density lipoprotein cholesterol), glycosylated hemoglobin, creatinine, and random glucose. Participants were not fasting. All serum biochemistry tests were sent to the National University Hospital Reference Laboratory for measurement on the same day.
Diabetes mellitus was defined in this study as nonfasting glucose levels of 200 mg/dL or greater (to convert to millimoles per liter, multiply by 0.0555) or physician diagnosis of diabetes mellitus and use of diabetes medications.
Metabolic syndrome components were defined as follows: for obesity, BMI of 25 or greater; for hypertriglyceridemia, triglyceride level of 1504 mg/dL or greater (to convert to millimoles per liter, multiply by 0.0113); for low level of high-density lipoprotein cholesterol, a value less than 38.7 mg/dL in men and less than 50.3 mg/dL in women (to convert to millimoles per liter, multiply by 0.0259); for high blood pressure, 130/85 mm Hg or greater or use of blood pressure medication; and for diabetes mellitus, as defined earlier.
Statistical analysis was performed with SPSS, version 11.5 (SPSS Inc, Chicago, Illinois). Proportions were compared by use of the χ2 test and means by the 2-tailed t test. Analysis of covariance models were used to estimate mean IOP adjusted for covariates. Multivariate logistic regression models were used to assess the odds ratio (OR) and its 95% confidence interval (CI) for glaucoma and POAG. Models were adjusted for age, sex, education, and smoking status.
A total of 3280 participants were recruited, giving a response rate of 78.7% for the study. Data on diabetes status were available for 3278 individuals. There were 764 persons with diabetes (23.3%); 658 represented known cases with a previous diagnosis of diabetes, whereas 106 participants were newly diagnosed on the basis of their random glucose level of 200 mg/dL or greater. Table 1 shows the characteristics of the study population by diabetes status.
Readings for IOP were available on 3263 right eyes. After controlling for age, sex, education, smoking status, CCT, and diabetes treatment, IOP was significantly higher in persons with than without diabetes (16.7 vs 15.0 mm Hg, P < .001). The IOP was also higher in the total population with higher serum glucose (P < .001), higher glycosylated hemoglobin (P < .001), higher total cholesterol (P = .001), and higher triglyceride (P = .002) levels, as well as in those with a higher BMI (P = .001) and a greater number of metabolic syndrome components (P < .001) (Table 2). The association between these risk factors and higher IOP was also statistically significant in the population of participants without glaucoma. The correlation coefficients between IOP and metabolic syndrome components are included in Table 3.
There were 150 persons (4.6%) with glaucoma, of whom 104 (3.2% of all 3280 participants) had POAG.33 The prevalence of glaucoma had a U-shaped relationship with the number of metabolic abnormalities, with the lowest prevalence being found in those with 3 metabolic abnormalities (Figure).
Table 3 shows the OR for glaucoma and POAG among the population by presence of diabetes and various metabolic abnormalities. The prevalence of glaucoma among diabetic participants was 4.7% compared with 4.5% in nondiabetic participants. For POAG, the prevalence among diabetic participants was 3.3% compared with 3.1% in nondiabetic participants. In multivariate logistic regression models adjusting for age, sex, education, smoking status, CCT, and diabetes treatment, this difference was not statistically significant (OR, 1.00; 95% CI, 0.63-1.61). There was no statistically significant association between glaucoma and serum glucose level, glycosylated hemoglobin concentration, hypertension, systolic blood pressure, or serum total cholesterol level. There was, however, a statistically significant negative association between increasing BMI and glaucoma (OR, 0.46; 95% CI, 0.28-0.76; comparing second vs first quartile of BMI) and between triglyceride level and glaucoma (OR, 0.59; 95% CI, 0.35-0.98; comparing fourth vs first quartile). Increasing number of metabolic syndrome components also had a negative association with glaucoma. The findings were largely similar for POAG (Table 3). There was no statistically significant association between diabetes and POAG (OR, 1.02; 95% CI, 0.58-1.79). We ran 2 additional multivariate models, which showed no statistically significant association between diabetes and POAG; the first adjusted for age, sex, education, smoking status, CCT, and BMI (OR, 0.98; 95% CI, 0.61-1.58) and the second for systolic blood pressure, serum glucose level, glycosylated hemoglobin concentration, total cholesterol level, and triglyceride level in addition to the foregoing factors (OR, 0.58; 95% CI, 0.29-1.18) (data not shown).
In this population-based study, we show that persons with diabetes mellitus and various metabolic syndrome abnormalities were more likely to have higher IOP, while controlling for CCT and other factors. However, persons with diabetes were not more likely to have glaucomatous optic neuropathy or POAG.
Our study adds to the literature on the association of diabetes with IOP and glaucoma (Table 44,5,7- 16,21,34). There have been several hypotheses for the association between diabetes and elevated IOP. Diabetes-related autonomic dysfunction may contribute to increased IOP.35 Epidemiologic data on the association between ocular hypertension and family history of type 2 diabetes mellitus36 suggest that genetic factors may also play some role. Glycation-induced corneal collagen cross-links in diabetes can cause corneal stiffening,37 which has also been shown to increase the level of measured IOP over the true IOP.38,39
The impact of hyperglycemia on the cornea may play a role in the discrepancy in findings between diabetes with IOP and glaucoma. Persons with diabetes have been shown to have greater CCT,17,40 which may artifactually increase IOP readings as measured by Goldmann applanation tonometry.41 In a previous study by our group, we reported a statistically significant increase in CCT in persons with diabetes compared with nondiabetic persons (547.2 vs 539.3 μm, P < .001).17 Increased CCT in fact has also been shown by the Ocular Hypertension Treatment Study42 and the European Glaucoma Prevention Study40,43,44 to reduce risk for progression from ocular hypertension to glaucoma. Diabetes is also associated with increased corneal stiffness, and cornea hysteresis has also been shown to have an effect on glaucoma risk independent of corneal thickness.45 Beyond overestimation of IOP, increased corneal thickness has also been associated with a smaller and more robust optic nerve head, which may be less susceptible to developing glaucomatous optic nerve damage in response to increased IOP.46 Another study reported increased lamina cribrosa movement with fluctuations in IOP in patients with thin corneas,47 suggesting that thick corneas may have a protective effect against damage induced by increased IOP. Other authors have also speculated that increased blood pressure in diabetic individuals may compensate for increased IOP, with respect to IOP-induced damage to the optic nerve,16 and that increased blood pressure may also increase the cerebrospinal fluid pressure, which may act as trans–lamina cribrosa counterpressure against elevated IOP.48
The association between increased BMI and IOP is consistent with previous studies that showed obesity and increasing BMI to be risk factors for elevated IOP.4,6,22- 24 The mechanisms explained in these previous reports include excess intraorbital fat tissue, increased episcleral venous pressure, and increased blood viscosity with increased outflow resistance of episcleral veins. These factors could result in a consequent decrease in outflow facility. Another study has suggested that the increased IOP may be due to transitory elevations in IOP resulting from breath holding and thorax compression while Goldmann tonometry is performed at the slitlamp on obese patients.49
We found an increasing number of metabolic abnormalities to be associated with elevated IOP, which has previously been suggested to be attributable to insulin resistance.18 Our study found a small positive association between total cholesterol and triglyceride levels and IOP, which concurs with a previous report.50 Another population-based study found an association between cholesterol and IOP,4 whereas a study on patients with suspected glaucoma found them to have hypertriglyceridemia.51 However, another study found no relationship between cholesterol and IOP in patients with glaucoma or ocular hypertension.52
Interestingly, BMI was found in our study to have a protective association with glaucoma, with those with higher BMI being at lower risk of glaucoma. This has also been noted in the Barbados Eye Study.13 There was a U-shaped association between BMI and glaucoma, with those in the first quartile of BMI being at highest risk and those in the second quartile at lowest risk. The U shape could be related to the fact that persons in the normal BMI range may be at lower risk of glaucoma than persons at either extreme. However, BMI may not be the best surrogate for body fat and vascular risk because it does not discriminate those with higher lean body mass. Furthermore, underweight individuals may be more likely to have chronic systemic disease, which may affect their susceptibility to glaucoma. A similar protective association against glaucoma was found with an increasing number of metabolic abnormalities. For example, hypercholesterolemia was not found to be a risk factor for glaucoma despite the association with increased IOP. The use of statins to control dyslipidemia in this group of patients may contribute to this finding because statins have been found to be associated with reduced risk of glaucoma53 as well as with slowed progression of optic nerve damage.54 Statins have also been shown to increase retinal blood flow and decrease IOP in a small clinical trial55 and have been found to exert an ocular hypotensive response in an organ-culture perfusion model.56 The discordance between risk profiles for IOP and glaucoma can also be explained by a study that showed that, in patients with glaucoma, family history and genetic factors may be stronger predictors of IOP than in those without glaucoma, where lifestyle and physiologic factors are more strongly related to IOP.57
The strengths of this study are that it was a large population-based study with a high participation rate (78.7%). Diagnosis of glaucoma was based on optic nerve changes and perimetric findings, and the classification of diabetes was based on history as well as serum glucose levels.
Limitations include selection bias, which may exist even with high participation rates. Serum glucose and lipid levels were ascertained from nonfasting random samples. This may result in underdiagnosis of diabetes compared with fasting serum glucose levels or a formal oral glucose tolerance test, and cholesterol and triglyceride levels will likely be higher compared with fasting samples. A meta-analysis of 12 studies by Bonovas et al58 suggested that diabetic individuals are at a significantly high risk of developing POAG, with an OR of 1.50 (95% CI, 1.16-1.93). Our study did not find a similar conclusion. Reasons for this difference could include a different study population (only 1 of the studies in the meta-analysis involved an Asian population [Korean]) and that none of the studies corrected for CCT. However, for a study with 103 participants with POAG and 3177 controls, and with a frequency of diabetes of 23.3%, we have 80% power (based on type I error of .05) to detect a true OR for POAG of 1.8 or higher. Therefore, our study may not be statistically powered to detect a more modest association (eg, OR of 1.5) between diabetes and POAG.
In conclusion, our study suggests that diabetes and metabolic abnormalities are associated with a small increase in IOP, but not with glaucomatous optic neuropathy. Further research may aid understanding of the complex interactions between diabetes, metabolic abnormalities, IOP, and the risk and pathogenesis of glaucoma.
Correspondence: Tin Aung, PhD, FRCSE, Singapore National Eye Centre, 11 Third Hospital Ave, Singapore 168751 (firstname.lastname@example.org).
Submitted for Publication: October 29, 2008; final revision received April 10, 2009; accepted May 31, 2009.
Financial Disclosure: None reported.
Funding/Support: This study was supported by grant 0796/2003 from the National Medical Research Council and grant 501/1/25-5 from the Biomedical Research Council, with additional support from the Singapore Prospective Study Program and the Singapore Tissue Network, A*STAR.