Asymmetry was defined as an intraocular pressure (IOP) greater than 5 mm Hg.
eAppendix. Drugs According to Medication Class
eTable 1. Subgroup Analysis According to Systemic Medical Condition
eTable 2. Multivariable Linear Regression Models for Interactions Between Medication Classes
Ho H, Shi Y, Chua J, Tham Y, Lim SH, Aung T, Wong TY, Cheng C. Association of Systemic Medication Use With Intraocular Pressure in a Multiethnic Asian PopulationThe Singapore Epidemiology of Eye Diseases Study. JAMA Ophthalmol. Published online January 12, 2017. doi:10.1001/jamaophthalmol.2016.5318
What is the association of systemic medication with intraocular pressure in a multiethnic Asian population?
In a post hoc analysis of a population-based study of 8063 participants from 3 ethnic groups (Chinese, Malays, and Indians), lower intraocular pressure was more likely associated with participants using systemic β-blockers, whereas higher intraocular pressure was more likely associated with participants using angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, statins, or sulfonylureas.
The effect of commonly dispensed systemic medication on intraocular pressure may have implications for glaucoma risk in individuals taking these medications for coexisting comorbidities.
There is limited understanding of the associations between systemic medication use and intraocular pressure (IOP) in the general population.
To examine the association between systemic medication use and IOP in a multiethnic Asian population.
Design, Setting, and Participants
In this post hoc analysis of the Singapore Epidemiology of Eye Diseases study, a population-based study of 10 033 participants (78.7% response rate) from 3 racial/ethnic groups (Chinese [recruited from February 9, 2009, through December 19, 2011], Malays [recruited from August 16, 2004, though July 10, 2006], and Indians [recruited from May 21, 2007, through December 29, 2009]), participants with glaucoma, previous ocular surgery, or trauma and an IOP asymmetry greater than 5 mm Hg between eyes were excluded. Intraocular pressure was measured using Goldmann applanation tonometry. An interviewer-administered questionnaire was conducted to collect data on medication and other variables. Data analysis was performed from August 1 through October 31, 2015.
Main Outcomes and Measures
Associations between medication and IOP were assessed using linear regression models adjusted for age, sex, body mass index, ethnicity, and the medical condition for which the medication was taken (angiotensin-converting enzyme inhibitors [ACEIs], angiotensin receptor blockers [ARBs], and β-blockers adjusted for blood pressure, statins adjusted for lipids, and biguanides, sulfonylureas, α-glycosidase inhibitors [AGIs], and insulin adjusted for glycosylated hemoglobin). Medications associated with significant IOP differences were incorporated into regression models adjusted for concomitant use of multiple medications. Generalized estimating equation models were used to account for correlation between eyes.
Of the 10 033 participants, we analyzed 8063 (mean [SD] age, 57.0 [9.6] years; 4107 female [50.9%]; 2680 Chinese [33.2%], 2757 Malay [34.2%], and 2626 Indian [32.6%] individuals). Systemic β-blocker use was independently associated with an IOP of 0.45 mm Hg lower (95% CI, −0.65 to −0.25 mm Hg; P < .001). Conversely, higher mean IOP was associated with use of ACEIs (0.33 mm Hg higher; 95% CI, 0.08 to 0.57 mm Hg; P = .008), ARBs (0.40 mm Hg higher; 95% CI, 0.40-0.75 mm Hg; P = .02), statins (0.21 mm Hg higher; 95% CI, 0.02-0.4 mm Hg; P = .03), and sulfonylureas (0.34 mm Hg higher; 95% CI, 0.05-0.63 mm Hg; P = .02). An interaction between medication classes for additive, synergistic, or antagonistic effects on IOP was not identified.
Conclusions and Relevance
Although systemic β-blocker use was associated with lower IOP and systemic ACEI, ARB, statin, and sulfonylurea use was associated with higher IOP in this study, the associations were modest at best. Only the associations with systemic hypoglycemic agents were greater than 1 mm Hg, a threshold that has translated to a 14% greater risk of incident glaucoma across 5 years in other studies. At this point, the effect of systemic medication on IOP in eyes with glaucoma is not well elucidated but important. Our findings indicate that patients with glaucoma may potentially be at risk of higher or lower IOP, depending on medication class, and this would in turn affect management of IOP control.
Chronic systemic diseases are increasingly prevalent1 and bring with them an approximately commensurate increase in individual use of prescription medication.2 Despite widespread medication use in light of the increasing burden of chronic systemic diseases, there is limited understanding of their effects on the eye.3,4
Glaucoma is the leading cause of irreversible blindness in older people, with an estimated burden of approximately 112 million people affected by glaucoma by 2040.5 Intraocular pressure (IOP) is the only modifiable risk factor for glaucoma, and lowering IOP prevents the development and progression of the disease.6,7 Typically, patients with glaucoma have an individualized target IOP, and surgery may be needed to control persistently high IOP above this threshold. Concomitantly, in people without evidence of glaucoma, a higher IOP (eg, >21 mm Hg) may trigger further tests and investigations and often necessitates long-term follow-up to monitor for progression, increasing the consumption of health care resources.
With an increasing burden of glaucoma and chronic diseases in elderly people, understanding how systemic medication use can influence or modify IOP is important. Although certain medications have been found to confer increased glaucoma risk,8,9 the influence of systemic medication on IOP is controversial.10- 12 Systemic β-blockers have yielded the most consistent findings, demonstrating an IOP-lowering effect in several clinic-based studies.11,13 The only population-based study14 to review systemic medication use and IOP, the European Prospective Investigation of Cancer–Norfolk Eye Study, found significantly lower IOP among a largely white population of participants taking nitrates in addition to β-blockers. No such study has been performed in an Asian population, which possesses different epidemiologic patterns in chronic diseases, medication use, and glaucoma compared with a white population. In this study, we examine the association of systemic medication use on IOP in a multiethnic Asian population in Singapore.
In a post hoc analysis, we analyzed data from the Singapore Epidemiology of Eye Diseases (SEED) study,15 a population-based study of Singaporeans aged 40 to 80 years from 3 major ethnic groups: Malays (recruitment conducted from August 16, 2004, through July 10, 2006), Indians (May 21, 2007, through December 29, 2009), and Chinese (February 9, 2009, through December 19, 2011). Data analysis was performed from August 1 through October 31, 2015. The SEED populations and methods have been reported previously.16,17 The study participants were selected by an age-stratified random sampling method and were invited for a standardized interview and comprehensive ocular examination at the Singapore Eye Research Institute. Ethnicity information was generated from questionnaires, as defined by the participants. Ethical approval was obtained from the institutional review board at the Singapore Eye Research Institute. Written informed consent was obtained from each participant, and the study adhered to the Declaration of Helsinki. All data were deidentified.
Briefly, each participant underwent standard ophthalmologic examination, including visual acuity and subjective refraction. Ocular biometry, including axial length, was measured using noncontact partial coherence interferometry (IOLMaster V3.01, Carl Zeiss Meditec AG). Slitlamp biomicroscopy (model BQ-900, Haag-Streit) was performed by study ophthalmologists and included optic disc evaluation using 78-diopter lenses. The IOP measurements were obtained using a Goldmann applanation tonometer (Haag-Streit) after administration of a single drop of topical anesthesia (0.5% amethocaine hydrochloride) into the inferior conjunctival sac and staining of cornea with a dye strip of fluorescein. Care was taken to ensure that just enough fluorescein was used to make the tonometer head visible. One reading was taken from each eye and recorded. If the reading was greater than 21 mm Hg, another reading was taken and this reading was used.
Glaucoma was diagnosed and classified using the International Society Geographical and Epidemiological Ophthalmology scheme. Glaucoma was defined as the presence of glaucomatous optic neuropathy, which was a loss of neuroretinal rim with a vertical cup to disc ratio of 0.7 or higher or vertical cup to disc ratio asymmetry greater than 0.2 between eyes and/or notching attributable to glaucoma, with compatible visual field loss.18
A detailed interviewer-administered questionnaire was used to collect relevant sociodemographic and medical information. All interviewers were bilingual, and participants were given a choice to be interviewed in English, Chinese, Malay, or Tamil. Self-reported systemic diseases and medication use were elicited from the interview. History of glaucoma medication use was also ascertained during the personal interview. Those who answered yes to the question were asked to show the interviewer the medication containers of all the products used for crosschecking to reduce recall errors. The interviewer then entered the product names into a computer.
Systemic examination was performed, including collection of blood samples.16,19 Participants’ height was measured in centimeters using a wall-mounted measuring tape, and weight was measured in kilograms using a digital scale (SECA model 782 2321009, Vogel & Halke). Height and weight were measured without shoes and with the participant standing. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared.20 Blood pressure was measured with the participant seated and after 5 minutes of rest. Systolic and diastolic blood pressures and pulse rate were measured using a digital automatic blood pressure monitor (Dinamap model Pro Series DP110X-RW, 100V2, GE Medical Systems Information Technologies).21 Blood pressure was measured on 2 occasions 5 minutes apart. A third measurement was taken if the blood pressure differed by more than 10 mm Hg systolic and 5 mm Hg diastolic. The mean between the 2 closest blood pressure readings was recorded.22 Hypertension was defined as systolic blood pressure of 140 mm Hg or higher, diastolic blood pressure of 90 mm Hg or higher, or use of antihypertensive medication. Diabetes was defined as a random glucose level of 200 mg/dL or higher (to convert to millimoles per liter, multiply by 0.0555), use of diabetic medication, or a physician diagnosis of diabetes. Nonfasting venous blood samples were drawn and sent for analysis of serum lipid levels (total cholesterol, high-density lipoprotein cholesterol, and low density lipoprotein cholesterol), glycosylated hemoglobin, and glucose at the National University Hospital Reference Laboratory on the same day. Hyperlipidemia was defined as a total cholesterol level of 239 mg/dL (to convert to millimoles per liter, multiply by 0.0259) or more or use of lipid-lowering drugs.19
For this post hoc analysis, 22 classes of medications that were used by at least 50 participants were identified. The classes of medication examined included statins, hypertensive medication (β-blockers, calcium channel blockers, angiotensin-converting enzyme inhibitors [ACEIs], angiotensin receptor blockers [ARBs]), diabetic medication (biguanides, sulfonylureas, α-glucosidase inhibitors, insulin), diuretics, proton pump inhibitors, histamine2-receptor antagonists, anticoagulants and antiplatelets, nonsteroidal anti-inflammatory medication, aspirin, antidepressants, antihistamines, inhaled steroids, thyroid hormone replacement, analgesics, fibrates, and nitrates (eAppendix in the Supplement). Vitamins and other oral supplements, such as iron, folate, and glucosamine, were excluded from analysis.
For each medication class, the IOP of the participants who were taking the medication was compared with those of the participants who were not taking medications in that particular class, irrespective of the use of other medication classes, using the t test. Statistical analysis was performed using R software, version 3.22 (R Development Core Team), and was conducted from August 1 through October 31, 2015.23 The association between medication and IOP was examined by multiple linear regression models with generalized estimating equations to account for intereye correlation. Association between single drug and IOP was adjusted for age, sex, BMI,24 ethnicity, and corresponding medical conditions according to the medication taken. Specifically, ACEIs, β-blockers, and ARBs were adjusted for blood pressure, statins were adjusted for lipid levels, and biguanides, sulfonylureas, AGIs, and insulin were adjusted for glycosylated hemoglobin levels. Medication classes significant on univariate analysis were further included in multiple regression models with multiple drugs. To take into account the mutual effects of each of the 6 medication classes significantly associated with IOP, participants taking this medication were included in the analysis, regardless of how many medications from other classes they were taking. Multiple linear regression analysis was then performed with the 6 medication classes together as independent variables, adjusting for concurrent use of all 6 classes. P < .05 was considered statistically significant (2-sided t test).
Of the 10 033 participants, we analyzed 8063 (mean [SD] age, 57.0 [9.6] years; 4107 female [50.9%]; and 2680 Chinese [33.2%], 2757 Malay [34.2%], and 2626 Indian [32.6%] individuals) (Figure). We excluded 1970 participants because of a history of glaucoma, previous ocular surgery or trauma, and missing records and 153 participants with a difference in IOP of more than 5 mm Hg between 2 eyes to account for possible measuring errors or undiagnosed ocular disease.14 The demographic characteristics of the included participants are given in Table 1.
Of the 22 medication classes included for analysis, participants taking ACEIs, ARBs, sulfonylureas, biguanides, and statins had significantly higher IOPs compared with participants who were not taking those medications (Table 2). After adjusting for age, sex, BMI, and ethnicity, lower IOP was found in participants taking β-blockers (0.24 mm Hg lower; 95% CI, 0.05-0.44 mm Hg; P = .01) (Table 3). Conversely, higher mean IOP was observed in participants taking ACEIs (0.56 mm Hg higher; 95% CI, 0.33-0.79 mm Hg; P < .001), statins (0.23 mm Hg higher; 95% CI, 0.06-0.40 mm Hg; P = .009), biguanides (0.70 mm Hg higher; 95% CI, 0.50-0.90 mm Hg; P < .001), sulfonylureas (0.83 mm Hg higher; 95% CI, 0.61-1.05 mm Hg; P < .001), α-glucosidase inhibitors (0.63 mm Hg higher; 95% CI, 0.02-1.24 mm Hg; P = .04), and insulin (0.81 mm Hg higher; 95% CI, 0.61-1.05 mm Hg; P = .02) compared with participants not taking each respective medication.
The association was then analyzed by further accounting for disease-specific confounders: systolic blood pressure for antihypertensive medication, lipid levels for statins, and glycosylated hemoglobin for hypoglycemic medication. The use of β-blockers remained significantly associated with lower IOP, whereas the use of ACEIs (0.53 mm Hg higher; 95% CI, 0.31-0.76 mm Hg; P < .001), ARBs (0.47 mm Hg higher; 95% CI, 0.13-0.81 mm Hg; P = .007), statins (0.32 mm Hg higher; 95% CI, 0.14-0.51 mm Hg; P = .001), biguanides (0.32 mm Hg higher; 95% CI, 0.10-0.55 mm Hg; P = .004), and sulfonylureas (0.45 mm Hg higher; 95% CI, 0.21-0.69 mm Hg; P < .001) remained associated with higher IOP (Table 3). The associations of ACEIs and ARBs with higher IOP and β-blockers with lower IOP remained significant when adjustments were made for diastolic blood pressure only or both diastolic and systolic blood pressures.
The mutual effects of each of the 6 medication classes were reviewed using multiple linear regression analysis, adjusting for concurrent use of all 6 classes. The results remained similar (Table 4): ACEIs resulted in an increase in IOP of 0.33 mm Hg (95% CI, 0.08-0.57 mm Hg; P = .008), ARBs increased IOP by 0.40 mm Hg (95% CI, 0.05-0.75; P = .02), statins by 0.21 mm Hg (95% CI, 0.02-0.4; P = .03), and sulfonylureas by 0.34 mm Hg (95% CI, 0.05-0.63; P = .02). β-Blockers remained the only medication class associated with a significant reduction in IOP of 0.45 mm Hg (95% CI, −0.65 to −0.25 mm Hg; P < .001).
Subgroup analysis was further performed for the 6 medication classes associated with significant effects on IOP for participants treated for the specific medical condition: ACEIs, ARBs, and β-blockers in participants with hypertension, biguanides and sulfonylureas in participants with diabetes, and statins in participants with hyperlipidemia (eTable 1 in the Supplement). The direction and magnitude of the effect estimate remained similar across all classes of medication. Further analyses were performed to examine interactions among medication classes for additive, synergistic, or antagonistic effects on IOP. In all classes, an interaction was not identified (eTable 2 in the Supplement).
In this multiethnic, population-based study in Asians, we found that systemic β-blocker use had a modest association with a lower IOP, consistent with its pharmacologic effects and previous reports. Conversely, higher IOP had a modest association in participants who were taking ACEIs, ARBs, sulfonylureas, biguanides, and statins. Only the association with systemic hypoglycemic agents was greater than 1 mm Hg, a threshold that has translated to a 14% greater risk of incident glaucoma across 5 years in another study.25 The differences in IOP compared with participants who were not taking medication were independent of concurrent use of other medication classes and systemic diseases.
To our knowledge, our study is the first to report the association of greater IOP with common systemic hypoglycemic medication of sulfonylureas and biguanides. Diabetes is a major risk factor for the development of glaucoma.26,27 Both diseases present an increasing burden in the face of an aging population,28,29 yet knowledge about how the 2 influence each other and the effect of management of one on the other is limited. Mapstone and Clark30 and Brazier31 reported a greater likelihood of abnormal response to oral glucose tests in patients with shallow anterior chambers, including patients with acute glaucoma, and hypothesized that this result could be explained by autonomic dysfunction. Other hypotheses for the association between high glucose level and IOP include genetic predisposition32 and possibly fluid movement via osmosis into the intraocular space.33 Although diabetes itself may account for increased IOP for patients taking sulfonylureas, this is unlikely based on the negative results of other hypoglycemic agents, such as biguanides, α-glucosidase inhibitors, and insulin.
The association between oral hypoglycemic agents and IOP is important, yet not comprehensively investigated in the general population. Most work has been performed in participants with known glaucoma, which may induce a different effect on aqueous humor dynamics. Despite a previous report34 that metformin appears to confer a dose-dependent risk reduction of open-angle glaucoma believed to involve mechanisms apart from improved glycemic control, our study revealed an associated increase in IOP with metformin use. A more plausible explanation may account for the increased IOP and sulfonylurea use found in our study. Sulfonamides, also known as sulfa drugs, can induce angle closure via nonpupillary block mechanisms, including choroidal effusion, lenticular swelling, and secondary shallowing of the anterior chamber with an increase in IOP.35- 37 Sulfonylureas are derivatives of sulfonamide and share a 10% risk of cross-reactivity with sulfa-based antibiotics,38 suggesting a potential for similar adverse effects attributable to structural similarities. Our finding of associated increase in IOP contrasts with the Norfolk Eye Study, in which IOP change was not significant in users of oral hypoglycemic agents. Our findings are especially noteworthy because the difference in IOP between participants using the systemic hypoglycemic agents biguanides and sulfonylureas compared with those who were not using these agents was greater than 1 mm Hg (Table 2), which could potentially translate to a 14% greater risk of incident glaucoma in 5 years.25 Although we corrected for glycemic control in our study, the influence of other factors that may affect our results, such as duration of diabetes, was not reviewed.
The IOP-lowering effect of β-blockers has been extensively investigated and documented among individuals with glaucoma39,40 and without glaucoma.13,41,42 A review of data found that a 1-mm Hg increase in IOP at baseline has been consistently associated with the development of glaucoma in healthy eyes collectively across various studies7,14; therefore, the clinical consequences of taking such a commonly prescribed systemic medication on IOP begs to be determined. Although the Rotterdam Study found a nonsignificant risk reduction (P = .06) of developing incident glaucoma associated with systemic β-blocker use,9 results of data obtained from a UK primary care database suggest that β-blockers may be protective against the development of glaucoma (odds ratio, 0.87; 95% CI, 0.8-0.94).43 In the Norfolk Eye Study, the authors found that the use of systemic β-blockers was sufficient to effect a significant lowering of IOP by approximately 1 mm Hg in participants compared with those not receiving treatment,14 whereas the magnitude of IOP lowering was less pronounced in our multiethnic population, with a difference of 0.13 mm Hg between treated and untreated participants. The difference in drug response could be partly explained by pharmacogenetics,44- 46 although such studies in Southeast Asian populations compared with whites are lacking.
The effect of systemic ACEI and ARB use on IOP is not as clearly defined because of the paucity of data. Angiotensin-converting enzyme inhibitors are believed to exert their effect on IOP via involvement of prostaglandins,47 which increases uveoscleral outflow.48 Most studies49- 52 that evaluated the ocular hypotensive effects of topical ACEIs were conducted in experimental animal models and found significant IOP reduction after administration of eye drops. In patients with glaucoma taking topical prostaglandin analogues, systemic ACEI use was associated with significantly less need for adjunctive glaucoma medication,53 which could indicate either an independent IOP-lowering effect or an enhancing synergistic effect with prostaglandin analogues.
Costagliola et al10,54 also reported significant IOP lowering among a small group of hypertensive and normotensive individuals with and without glaucoma given oral captopril and oral losartan. Conversely, no significant effect on IOP was seen after administration of oral captopril experimentally in rabbits55 or for ACEI and ARB use in the population-based Norfolk Eye Study.14 The significantly greater IOP for our participants using ACEIs and ARBs, to our knowledge, is unprecedented in the literature. Although one would expect greater ocular adverse effects with direct topical application of medication compared with systemic use, intraocular drug concentrations may provide further insight to our findings.
Strengths of our study include a large population-based sample. In all medication classes with a significant effect on IOP, a large number of participants were taking the medication. As a prospective study, uniformed data collection was performed across all ethnic groups at the onset, which also allowed for verification of medication classes in all participants.
However, our study also has several limitations. This analysis was post hoc, that is, the protocol for this study during the time of data collection and initial analyses did not explicitly plan for this analysis. Intraocular pressure was obtained at 1 time point, and the effects of IOP fluctuation must be considered. Participant adherence to medication use was also not reviewed, and the duration of use and dosage of medication are unknown, which would be important in interpreting results. Because our findings were derived from a population without glaucoma, we are unable to conclude that similar associations will be found in those with glaucoma. The associations between systemic medication use and change in IOP are not cause and effect, are modest at best, and, apart from the associations with hypoglycemic agents, may have limited clinical relevance.
In this post hoc analysis of a large multiethnic Asian population, the change in IOP associated with systemic medication is modest. Participants taking systemic β-blockers had lower IOPs. Conversely, the use of systemic ACEIs, ARBs, statins, and sulfonylureas was associated with higher IOP. Only the associations with systemic hypoglycemic agents were greater than 1 mm Hg, a threshold that has translated to a 14% greater risk of incident glaucoma across 5 years in other studies. At this point, the effect of systemic medication on IOP in eyes with glaucoma is not well elucidated but important. Our findings indicate that patients with glaucoma may potentially be at risk of higher or lower IOP, depending on medication class, and this would in turn affect management of IOP control.
Corresponding Author: Ching-Yu Cheng, MD, PhD, Singapore Eye Research Institute, Singapore National Eye Center, 20 College Rd, The Academia, Level 6, Singapore 169856 (firstname.lastname@example.org).
Accepted for Publication: November 15, 2016.
Published Online: January 12, 2017. doi:10.1001/jamaophthalmol.2016.5318
Author Contributions: Dr Cheng had full access to all data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Ho, Tham, Wong, Cheng.
Acquisition, analysis, or interpretation of data: Ho, Shi, Chua, Tham, Lim, Aung, Wong.
Drafting of the manuscript: Ho.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Ho, Shi, Aung.
Obtained funding: Aung, Cheng.
Administrative, technical, or material support: Aung, Cheng.
Study supervision: Wong.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Funding/Support: The SEED study is supported by grants 0796/2003 (Dr Wong), STaR/0003/2008 (Dr Wong), CIRG/1371/2013 (Dr Cheng), and CIRG/1417/2015 (Dr Cheng) from the National Medical Research Council (NMRC), Singapore, and grants 08/1/35/19/550 and 09/1/35/19/616 from the Biomedical Research Council (BMRC), Singapore. Dr Cheng is supported by award CSA/033/2012 from the NMRC. The NMRC and BMRC provided funding to allow for data collection from participants of the SEED study.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.