Association of Drugs With Acute Angle Closure

Key Points

Question  Which drugs are associated with acute angle closure (AAC), and what are the risks of AAC associated with each drug?

Findings  In this case-crossover study of data from Korea’s Health Insurance Review and Assessment Service database, the prescription records of 13 531 patients with AAC were analyzed, and 61 drugs were found to be associated with AAC. Among these, sumatriptan, topiramate, and duloxetine were found to have the highest odds of AAC development, and several drugs not previously reported to be associated with AAC, such as lactulose and metoclopramide, were identified.

Meaning  These findings suggest a need to consider these associations in patients taking any of the 61 drugs found to be associated with AAC.

Importance  Acute angle-closure (AAC) glaucoma is a sight-threatening disease and can reportedly occur in association with various drugs.

Objective  To identify drugs that are associated with AAC glaucoma occurrence and evaluate the risk of AAC associated with each drug.

Design, Setting, and Participants  A case-crossover study was conducted using the Health Insurance Review and Assessment Service database, which contains medical information of the entire Korean population. Patients who were first diagnosed with AAC and treated between 2013 and 2019 were identified using diagnostic and procedure codes. All drugs that the study participants were prescribed as well as prescription dates during the period of 1 to 180 days before the onset of AAC were extracted from the database. For each patient, 1 to 30 days before onset was considered the hazard period, and 91 to 180 days before AAC onset was considered the control period.

Main Outcomes and Measures  Drugs associated with AAC and odds (calculated as odds ratios [ORs] with 95% CIs) of AAC development associated with each identified drug.

Results  A total of 949 drugs that were prescribed to 13 531 patients with AAC (mean [SD] age, 66.8 [10.6] years; 9585 [70.8%] female) during the period of 1 to 180 days before the onset of AAC were analyzed. A total of 61 drugs were found to be associated with AAC, among which sumatriptan (OR, 12.60 [95% CI, 4.13-38.44]) was associated with the highest odds of AAC development, followed by topiramate (OR, 5.10 [95% CI, 2.22-11.70]) and duloxetine (OR, 4.04 [95% CI, 2.95-5.54]). The median (IQR) period from prescription of the drug to the onset of AAC for the 61 drugs was 11.9 days (10.9-12.8). A number of drugs not previously considered to be associated with AAC, including lactulose (OR, 2.81 [95% CI, 1.72-4.61]) and metoclopramide (OR, 2.52 [95% CI, 1.95-3.25]), were identified.

Conclusions and Relevance  Results of this case-crossover study suggest a need to consider AAC risk in patients taking any of the 61 drugs found to be associated with AAC.

Globally, glaucoma is a common cause of visual impairment and blindness.^1,2 The number of patients with primary angle-closure glaucoma was reported to be 20.17 million worldwide in 2013 and is expected to increase to 32.04 million by 2040.³

Acute angle-closure (AAC) glaucoma is a sight-threatening disease that causes irreversible damage to the optic nerves. It occurs when intraocular pressure rises rapidly as a result of abrupt occlusion of the anterior chamber angle, which constitutes a pathway for the aqueous humor to exit the eyeball. The pathophysiology of AAC can be classified into 2 types: pupillary block mechanism due to lens-iris apposition at the pupil, which can result from several conditions, such as mydriasis and posterior synechiae, and nonpupillary block due to the forward pushing of the peripheral iris because of the anterior position of the lens or ciliary body edema.

The demographic factors associated with angle closure are older age, female sex, and Asian ancestry.⁴ The major ocular risk factor is a shallow anterior chamber.⁵ In addition, several drugs have also been reported to be associated with AAC.^6-24

Understanding the association of drugs with AAC, a disease that can lead to sudden vision loss, is important in preventing the condition. If a patient needs to take a drug that is highly associated with AAC and has demographic and ocular risk factors for this disease, appropriate preventive measures, such as prophylactic laser peripheral iridotomy or cataract surgery, may reduce the risk of AAC occurrence.²⁵

However, to our knowledge, no risk assessment has been accurately performed for most drugs associated with AAC. Furthermore, there may be other drugs that have not yet been reported that may also be associated with AAC. Thus, in this study, we identified drugs associated with AAC occurrence and evaluated the risk of this disease using a national population-based database.

This study was approved by the institutional review board of Kangdong Sacred Heart Hospital and complied with the tenets of the Declaration of Helsinki²⁶; and informed consent was not required. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observational studies.

In Korea, compulsory insurance policies are in place for the entire population of approximately 53 million people. Overall, 97.1% of the population is covered by the National Health Insurance Service, and the remaining 2.9% is covered by Medical Aid Program benefits and health care benefits for veterans.

For these insurance policies, information on all medical practices is collected and managed by the Health Insurance Review and Assessment (HIRA) Service. The HIRA database contains all of the medical codes for diagnoses, procedures, and prescriptions for the entire population. We searched the HIRA database for patients with AAC and investigated the potential association between this disease and drugs prescribed prior to the onset of AAC. All data we received from the HIRA database were deidentified.

For all patients, diagnoses were coded with the Korean Standard Classification of Disease Version 5 (KCD) (modified from the International Statistical Classification of Diseases and Related Health Problems, Tenth Revision [ICD-10]). Diagnoses coded as H40.20 (a diagnosis of AAC) from 2011 to 2019 were extracted from the HIRA database. From the records with an H40.20 code, we included those in which the claim date of the H40.20 code was earlier than that for the following KCD codes: H40.1 (primary open-angle glaucoma), H40.2 (primary angle-closure glaucoma), H40.3 (glaucoma secondary to eye trauma), H40.4 (glaucoma secondary to eye inflammation), H40.5 (glaucoma secondary to other eye disorders), H40.6 (glaucoma secondary to drugs), H40.8 (other glaucoma), H40.9 (glaucoma, unspecified), H42.0 (glaucoma in diseases classified elsewhere), and Q15.0 (congenital glaucoma). Meanwhile, cases in which the earliest claim date of code H40.20 was 2011 or 2012 were excluded to ensure that only patients with a first diagnosis of AAC between 2013 and 2019 were included.

Among the cases selected using diagnostic codes, those with at least 1 of the following medical procedure codes²⁷ from 2 weeks before and 4 weeks after the earliest claim date containing code H40.20 were identified: S5041 (surgery for glaucoma-iridectomy), S5030 (laser iridectomy), S5110 (surgery for cataract or lens [pars plana lensectomy]), S5111 (surgery for cataract or lens extracapsular or intracapsular extraction), S5113 or S5119 (surgery for cataract or lens phacoemulsification), and S5117 (intraocular lens implantation—primary). Patients who satisfied both the diagnostic and procedure codes were included in the present study.

The case-crossover design was first introduced in 1991 by Maclure et al²⁸ as a research method suitable for the study of factors that trigger acute diseases and has been used in various fields, including clinical epidemiology.^29-32 In patients with AAC, the earliest claim date for H40.20 was set as the index date. For each patient, the period between 1 to 30 days before the index date was considered the hazard period, and 91 to 120 days, 121 to 150 days, and 151 to 180 days were considered the control periods. All drugs prescribed to the study participants and the prescription dates from 1 to 180 days before the index date were extracted from the HIRA database. Drugs prescribed by ophthalmology departments nationwide were excluded.

Using the Mantel-Haenszel estimator, we conducted a case-crossover study²⁸ to assess which of these drugs had an association with AAC. Odds of developing AAC, calculated as odds ratios (ORs) with 95% CIs, were assessed for each AAC-associated drug. φ Coefficients were analyzed to evaluate the correlations between different AAC-associated prescribed drugs. To reduce the probability of false-positive results arising from multiple comparisons, a 2-tailed P < .001 was considered statistically significant and indicative of an association between a drug and AAC. The World Health Organization Anatomical Therapeutic Chemical classification system³³ was used to divide AAC-associated drugs into groups based on the organ or system on which they act and their chemical properties.

The numbers of patients exposed to each AAC-associated drug over the entire period (1-180 days before onset), hazard period (1-30 days before onset), and total control period (91-180 days before onset) were analyzed. The number, age, and sex of the patients exposed to each AAC-associated drug only during the hazard period were obtained. The onset time was defined as the duration from the prescription of the drug to the onset of AAC.

The drug-associated AAC group was defined as patients with AAC who were exposed to at least 1 of the AAC-associated drugs only during the hazard period. Patients with AAC not associated with a drug were referred to as the non–drug-associated AAC group. The ages of the drug-associated AAC group and non–drug-associated AAC group were compared using the t test, and sex was compared using the χ² test. Statistical analyses were performed using RStudio, version 1.1.463 (RStudio, PBC).

A total of 41 070 patients with the H40.20 code were identified in the HIRA database from 2011 to 2019, and their 19 102 275 claims data during the same period were analyzed. Among the 41 070 patients with the H40.20 code, 22 942 were included for whom the claim date of the H40.20 code was earlier than the claim date of the following codes: H40.1, H40.2, H40.3, H40.4, H40.5, H40.6, H40.8, H40.9, H42.0, and Q15.0. Among the 22 942 patients, we excluded patients for whom the earliest claim dates of H40.20 were in 2011 (n = 2652) and 2012 (n = 2523). Among the 17 767 patients selected using diagnostic codes, 13 531 patients with at least 1 of the following codes from 2 weeks before to 4 weeks after the date of the earliest claim with H40.20 were identified: S5041, S5030, S5110, S5111, S5113, S5119, and S5117. The mean (SD) age of the 13 531 patients was 66.8 (10.6) years. There were 9585 female patients (70.8%) and 3946 male patients (29.2%). The earliest claim date of H40.20 was set as the index date, and the association between drugs and AAC occurrence was analyzed in these 13 531 patients with AAC. The numbers of patients with AAC by year were 1962 in 2013, 2208 in 2014, 2198 in 2015, 1853 in 2016, 1762 in 2017, 1802 in 2018, and 1746 in 2019.

In all, 949 drugs that were prescribed for 13 531 patients with AAC from 1 to 180 days before the index date were analyzed. A total of 61 drugs were significantly associated with the occurrence of AAC (according to the threshold of P < .001) in the Mantel-Haenszel estimate. The φ coefficient was estimated to evaluate the correlation between each of the 61 drugs. The φ coefficient of acetylcysteine and chlorpheniramine was 0.23, and the φ coefficient of aceclofenac and eperisone was 0.21. The φ coefficients for the other drugs were less than 0.20.

The Table shows the names and results of the analysis of each of the 61 drugs: the number of patients exposed to each AAC-associated drug over the entire period, hazard period, and overall control period; results of the Mantel-Haenszel estimate; and number, age, sex, and onset time of patients exposed to each AAC-associated drug only during the hazard period.

Based on the Anatomical Therapeutic Chemical classification system, the 61 drugs were classified into eight groups: nervous system drugs, respiratory system drugs, drugs for the alimentary tract and metabolism, genitourinary system drugs and sex hormones, drugs for the musculoskeletal system, anti-infectives for systemic use, corticosteroids for systemic use, and other drugs.

In the nervous system group, 12 drugs were associated with AAC. Among them, sumatriptan (OR, 12.60 [95% CI, 4.13-38.44]) had the highest OR, followed by topiramate (OR, 5.10 [95% CI, 2.22-11.70]) and duloxetine (OR, 4.04 [95% CI, 2.95-5.54]). Other antidepressants that showed associations were amitriptyline (OR, 1.63 [95% CI, 1.33-1.99]), escitalopram (OR, 2.93 [95% CI, 2.13-4.04]), and nortriptyline (OR, 2.61 [95% CI, 1.82-3.74]). The anxiolytics alprazolam (OR, 1.59 [95% CI, 1.40-1.82]), buspirone (OR, 1.98 [95% CI, 1.47-2.66]), and diazepam (OR, 1.49 [95% CI, 1.34-1.66]) also showed associations with AAC.

In the respiratory system group, 14 drugs were associated with AAC. Among them, aminophylline (OR, 2.23 [95% CI, 1.62-3.06]) had the highest OR, followed by mequitazine (OR, 2.02 [95% CI, 1.62-2.52]) and piprinhydrinate (OR, 1.87 [95% CI, 1.60-2.19]). Pseudoephedrine was also associated with AAC (OR, 1.50 [95% CI, 1.36-1.65]).

In the alimentary tract and metabolism group, 13 drugs showed associations with AAC. Among them, lactulose (OR, 2.81 [95% CI, 1.72-4.61]) had the highest OR, followed by metoclopramide (OR, 2.52 [95% CI, 1.95-3.25]) and dimenhydrinate (OR, 1.76 [95% CI, 1.53-2.03]). In the genitourinary system and sex hormones group, propiverine (OR, 3.22 [95% CI, 2.53-4.10]) and solifenacin (OR, 2.50 [95% CI, 1.93-3.24]) were associated with AAC.

In the musculoskeletal system group, 11 drugs were associated with AAC. Among them, naproxen (OR, 3.49 [95% CI, 2.28-5.34]) had the highest OR, followed by orphenadrine (OR, 1.83 [95% CI, 1.42-2.34]) and ibuprofen (OR, 1.57 [95% CI, 1.33-1.86]).

In the anti-infectives for systemic use group, 4 drugs showed associations with AAC. Among them, levofloxacin (OR, 1.57 [95% CI, 1.28-1.92]) had the highest OR, followed by ofloxacin (OR, 1.56 [95% CI, 1.31-1.85]) and clarithromycin (OR, 1.50 [95% CI, 1.24-1.80]).

In the corticosteroids for systemic use group, prednisolone (OR, 1.48 [95% CI, 1.32-1.64]) and methylprednisolone (OR, 1.37 [95% CI, 1.24-1.51]) were associated with AAC.

In the other drugs group, ammonium chloride (OR, 1.46 [95% CI, 1.32-1.61]) and caffeine anhydrous (OR, 1.74 [95% CI, 1.50-2.02]) both showed associations with AAC in this study, as did streptokinase-streptodornase, although the OR was lower at 1.18 (95% CI, 1.11-1.25).

Among the 13 531 patients with AAC, 4951 (36.6%) were exposed to at least 1 of the 61 drugs only during the hazard period. Compared with the 8580 patients with non–drug-associated AAC, the patients with drug-associated AAC were older (mean [SD] age, 67.7 [9.8] years vs 66.2 [11.0] years, P < .001), and there was a greater proportion of female patients (3786 of 4951 patients [76.5%] vs 5799 of 8580 patients [67.6%]; P < .001).

The median (IQR) onset time for AAC-associated drugs was 11.9 (10.9-12.8) days . Sumatriptan showed the shortest onset time (mean [SD], 6.0 [7.8] days), followed by duloxetine (6.5 [5.5] days) and metoclopramide (7.7 [8.5] days).

In patients with drug-associated AAC, the mean (SD) number of drugs to which the patients were exposed only during the hazard period was 2.83 (1.89), with 1430 exposed to 1 AAC-associated drug, 1196 exposed to 2 drugs, 935 exposed to 3 drugs, 578 exposed to 4 drugs, 349 exposed to 5 drugs, 207 exposed to 6 drugs, 123 exposed to 7 drugs, 72 exposed to 8 drugs, 24 exposed to 9 drugs, 19 exposed to 10 drugs, 6 exposed to 11 drugs, 6 exposed to 12 drugs, 4 exposed to 13 drugs, 1 exposed to 14 drugs, and 1 exposed to 17 drugs.

This case-crossover study found that AAC was associated with a variety of prescription drugs, some of which are well known for their association with AAC; however, to the best of our knowledge, some have not been reported previously.

Various drugs that act on the nervous system are reportedly associated with AAC. Indeed, duloxetine,¹⁸ escitalopram,²³ alprazolam,¹⁰ topiramate,^7,11,13,16 and sumatriptan^9,24 have been reported to induce AAC in previous case reports.

In a case-crossover study, Seitz et al³⁴ found an association between AAC and antidepressant use within 30 days before AAC onset (OR, 1.62 [95% CI, 1.16-2.26]). Symes et al³⁵ analyzed patients less than 50 years of age who had AAC and reported associations between AAC and bupropion (adjusted rate ratio, 1.98 [95% CI, 1.02-3.84]) and topiramate (adjusted rate ratio, 5.30 [95% CI, 2.54-11.04]) . In a study by Chen et al,³⁶ an association was found between AAC and selective serotonin reuptake inhibitor use within 7 days before AAC onset (OR, 5.80 [95% CI, 1.89-17.9]). Park et al³⁷ reported an association between AAC and new benzodiazepine use within 7 days before AAC (OR, 3.09 [95% CI, 1.58-5.88]).

In our study, amitriptyline, nortriptyline (a nonselective monoamine reuptake inhibitor), duloxetine (a selective serotonin-norepinephrine reuptake inhibitor), and escitalopram (a selective serotonin reuptake inhibitor) were associated with AAC. Alprazolam, buspirone, and diazepam (anxiolytics) also showed associations with AAC. In addition, the antiepileptic topiramate (OR, 5.10 [95% CI, 2.22-11.70]) and the antimigraine preparation sumatriptan (OR, 12.60 [95% CI, 4.13-38.44]) had the highest ORs among all of the drugs analyzed in this study.

The mechanism of action in the association of antidepressant use with AAC is thought to be mydriasis induced by anticholinergic and adrenergic effects, as well as an increase in serotonin level.³⁸ Uveal effusion is also a possible mechanism of antidepressant-associated AAC.^8,20,23 Anxiolytics, such as alprazolam, may induce mydriasis through their anticholinergic effects.^10,37 Topiramate is well known to be associated with AAC through uveal effusion.³⁹ For sumatriptan as well, uveal effusion is considered a mechanism of AAC.⁹

In this study, various antihistamine drugs were associated with AAC. Classic first-generation oral antihistamines and some second-generation oral antihistamines have an anticholinergic effect that induces mydriasis.⁴⁰ Pseudoephedrine (a sympathomimetic), which was previously reported as a possible cause of bilateral AAC,^6,19 was also associated with AAC in our study. Additionally, various mucolytic agents and cough suppressants were associated with AAC.

Nonprescription, flu medication–associated bilateral AAC cases have been reported.^6,12,15,17 Nonprescription flu medications generally consist of sympathomimetics that act as nasal decongestants,⁴¹ antihistamines,⁴² and cough suppressants⁴³ with anticholinergic effects. Acetaminophen, ammonium chloride, and caffeine anhydrous, which showed associations with AAC in this study, are also widely used as nonprescription flu medications. Drugs that are commonly prescribed together with drugs that induce AAC may show associations with AAC despite the fact that they do not directly induce AAC. Thus, the φ coefficient was evaluated between the drugs that were associated with AAC in this study. However, none of the drugs showed a mutual association among the drugs included in our study, which may have occurred because there are various combinations of similar drugs being prescribed to these patients.

The histamine 2 receptor antagonists cimetidine and ranitidine, as well as the antiemetic dimenhydrinate, have anticholinergic effects and may induce AAC.⁴⁴ Metoclopramide is also known to have anticholinergic effects.⁴⁴

Lactulose, a treatment for constipation, was associated with AAC. Qu et al⁴⁵ reported that constipation is associated with angle-closure glaucoma, with the Valsalva maneuver posited as a possible pathophysiological mechanism. The Valsalva maneuver is performed by attempting strong exhalation against an obstructed airway, and it can be induced during coughing, vomiting, and constipation. The Valsalva maneuver may lead to an increase in choroidal volume⁴⁶ and narrowing of the anterior chamber angle,⁴⁷ which can cause AAC. In this study, the association between AAC and various drugs used for constipation, cough, and vomitinwas considered to be associated with the Valsalva maneuver. Among such drugs, even lactulose and metoclopramide, which show relatively high odds of AAC development, have not previously been reported to be associated with AAC.

Antimuscarinic drugs for an overactive bladder may cause mydriasis as an adverse effect,⁴⁸ resulting in AAC. In this study, propiverine and solifenacin showed associations with AAC, with relatively high ORs of 3.22 (95% CI, 2.53-4.10) and 2.50 (95% CI, 1.93-3.24), respectively.

In this study, various nonsteroidal anti-inflammatory drugs were associated with AAC. Previously, mefenamic acid²¹ and aspirin⁴⁹ were reported to induce transient myopia and angle closure. Orphenadrine is a high-potency anticholinergic⁴⁴ and was associated with AAC in this study.

The following considerations should be taken into account regarding the drugs found to be associated with AAC in this study. First, the drugs themselves can induce AAC through various mechanisms, such as mydriasis and uveal effusion. Also, the drugs may be prescribed along with other drugs that cause AAC. Additionally, the diseases for which the drugs are prescribed may be associated with AAC-inducing mechanisms, such as the Valsalva maneuver in constipation, coughing, and vomiting. In addition, drugs may be prescribed for AAC symptoms, such as a unilateral headache (symptom similar to migraine), prior to AAC diagnosis.

In this study, 36.6% of patients with AAC had drug-associated AAC. They were prescribed an average of 2.83 AAC-associated drugs for the first time in the period of 1 to 30 days prior to AAC diagnosis during the 180-day study period. This outcome may be attributable to several factors. First, multiple drugs that can induce AAC may be prescribed and taken simultaneously, thereby increasing the possibility of AAC occurrence. Second, there may be some drugs that do not induce AAC that are often prescribed along with other drugs that can cause AAC. Third, several medications may be prescribed for the treatment of diseases associated with AAC. Fourth, several medications may be prescribed for the treatment of AAC symptoms prior to AAC diagnosis.

This study has several limitations. First, it was not possible to know the actual context in which AAC occurred for each patient since the patients were identified using claims code data. To enhance the accuracy of these results, procedure codes were also used to select patients who experienced AAC for the first time between 2013 and 2019. Second, we were unable to confirm whether patients were taking the prescribed drugs or how they were being taken. Because of the limitations of this study design and to reduce the probability of false-positive results due to multiple comparisons, we set the threshold of statistical significance at P < .001, which is lower than the generally used value of P < .05. Therefore, we decreased the possibility that drugs that were not actually associated with AAC were included in the AAC-associated drug group; this method likely yielded more precise results, even if it resulted in the discovery of fewer drugs that may be associated with AAC.

Acute angle closure is associated with various prescription drugs. Some of these drugs are well known for this outcome, although this study found associations with drugs that have not previously been reported. The results of this study suggest that patients who are prescribed these drugs may be at risk of developing AAC. Clinicians should be aware of the increased risk, provide appropriate explanations to the patient, conduct an ophthalmic examination, and take precautionary measures when other risk factors for AAC are present.

Accepted for Publication: July 31, 2022.

Published Online: September 22, 2022. doi:10.1001/jamaophthalmol.2022.3723

Corresponding Author: Sung Pyo Park, MD, PhD, Department of Ophthalmology, Kangdong Sacred Heart Hospital, Hallym University College of Medicine, 150, Seongan-ro, Gangdong-gu, Seoul 05355, South Korea (eyepyo@gmail.com).

Author Contributions: Drs Na and Park had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Both authors.

Acquisition, analysis, or interpretation of data: Both authors.

Drafting of the manuscript: Na.

Critical revision of the manuscript for important intellectual content: Both authors.

Statistical analysis: Na.

Obtained funding: Na.

Administrative, technical, or material support: Both authors.

Supervision: Park.

Conflict of Interest Disclosures: Dr Park reported receiving grants from the Korean Association of Retinal Degeneration and Basic Science Research Program through the National Research Foundation of Korea, funded by the Ministry of Science and Information and Communication Technology, outside the submitted work. No other disclosures were reported.

Funding/Support: This study was supported by grant 2020-02 from the Kangdong Sacred Heart Hospital Fund (Dr Na).

Role of the Funder/Sponsor: The funder of this study 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.

Additional Contributions: We thank Jeong Jin Young, PhD, Research Institute of Clinical Epidemiology, Hallym University, Chuncheon, Korea, for assistance with statistical analyses in this study. He received no compensation for this contribution.