A Pharmacovigilance Study of Drug-Induced Glaucoma Utilizing the Japanese Adverse Event Reporting System

Introduction

Glaucoma is one of the leading causes of irreversible blindness worldwide,¹ and is a disease of great social importance. The proportion of glaucoma is known to increase with age and the global prevalence of those was approximately 3–5% for people aged 40–80 years.² Glaucoma is broadly divided into 2 categories: open-angle glaucoma and angle-closure glaucoma. The common feature for all forms of glaucoma is loss of retinal ganglion cells, thinning of the retinal nerve fiber layer, and cupping of the optic disc.³ Glaucoma can be mostly caused by elevated intraocular pressure,⁴ and also secondarily by trauma or medications.⁵ Drug-induced glaucoma occurs when the drugs interfere with the drainage of the aqueous humor that fills the eye, causing an abnormal increase in intraocular pressure. This elevated pressure without optic nerve damage is referred to as “ocular hypertension”. The most common subjective symptoms of glaucoma are the appearance of areas that cannot be seen or a narrowing of the range of vision. However, we see with both eyes in daily life and the disease progresses slowly in most cases. Most people are almost unaware of visual field disturbances in the early stages of the disease of glaucoma, leading to irreversible blindness. Therefore, it is important to know which drugs more often cause glaucoma in clinical settings.

There is a growing consensus that a detailed evaluation of the information through pharmacovigilance activities is important for all drugs to ensure their safe use.^6,7 Of note, pharmacovigilance practices can improve information feedback to medical staff and their patients in a timely manner, thereby reducing the overall risk to patients. Certainly, drugs are approved for clinical use on the basis of indicating a satisfactory balance between benefits and risks. However, the safety profile of drugs can change over time as their use expands with patient characteristics and an increase in the number of patients exposed. Spontaneous reporting systems represent a primary source of information to detect safety signals. The Japanese Adverse Drug Event Report (JADER) database is a published large database managed by the Pharmaceuticals and Medical Devices Agency (PMDA) for the pharmacovigilance approach. The objective of this study was to identify the drugs reported most frequently reported to be associated with glaucoma using a spontaneous reporting system database.

Methods

The JADER reports were downloaded from the PMDA website (http://www.pmda.go.jp/safety/info-services/drugs/adr-info/suspected-adr/0003.html), which contains data on the adverse effects of medications and patient information in Japan since April 1, 2004. We used 378,533 cases of data from JADER between April 2004 and January 2017. JADER consists of 4 datasets: patient demographic information (DEMO), drug information (DRUG), adverse events (REAC), and medical history (HISTO). Details of this database have been described previously.^8–16 For the current analyses, we only extracted cases that were classified as “suspected medicine” since a “suspected medicine” is defined as a pharmaceutical product with which an adverse event is suspected to be associated and analyzed the reports of suspected drugs and adverse drug reactions (ADRs).

The classification and standardization of adverse drug reactions in JADER data is referred to the Medical Dictionary for Regulatory Activities (MedDRA). In the JADER database, each report is coded using Preferred Terms (PTs) from MedDRA. In MedDRA, a given PT can be assigned to a specific High-level Term (HLT), High-level Group Term (HLGT), and System Organ Class (SOC) level, but each HLT, HLGT, and SOC often contains multiple PT. Based on JADER database, information concerning glaucoma in PT Name coded in MedDRA (version 20.1) was collected. Subsequently, we excluded reports without information of sex (n = 29,976) and age (n = 49,713) from the data (n = 1,984,122). After exclusion, data were available 1,904,433 reports.

We compiled a cross-tabulation table based on two classifications: the presence or absence of glaucoma and the presence or absence of the suspected medicine. Then, we calculated the reporting odds ratio (ROR) by the following formula.

a: the number of patients with a target event when they received a target drug.

b: the number of patients with non-target adverse events when they received a target drug.

c: the number of patients with a target event when they received non-target drugs.

d: the number of patients with non-target adverse events when they received non-target drugs.

Generally, ROR is used with the spontaneous reporting database as an index of the relative risk of drug-associated adverse events. A signal is considered to be present when the lower limit of the 95% confidence interval (CI) of the ROR is greater than one. In this database, age, height, and weight information are indicated as follows: age in decades, height in centimeters, and weight in kilograms. These data are not given as continuous variables because of privacy considerations. Because the prevalence of glaucoma is high at older ages and among men,¹⁷ ROR was adjusted for age (< 60s and ≥ 60s) and sex. Statistical significance was set at p < 0.05. All analyses were performed with SPSS for Windows software (ver. 19.0; SPSS Inc., Tokyo, Japan).

Results

During the study period, a total number of 1,904,433 ADR reports including both age and sex information were obtained. Among cases represented by glaucoma, we obtained glaucoma (609 reports), angle-closure glaucoma (207 reports), normal-tension glaucoma (17 reports), open-angle glaucoma (9 reports), and exfoliation glaucoma (1 report). Reports of glaucoma and angle-closure glaucoma accounted for the majority of cases, so further analysis was conducted focused on them. The patients’ characteristics are shown in Table 1. Glaucoma was almost equally reported in men (54%) and women (46%); whereas angle-closure glaucoma was reported more frequently in women (86%). According to the age distribution of the study population, glaucoma occurred frequently in their 60s (19.7%) and angle-closure glaucoma did in their 70s (38.6%). Of note, the rate of remission from drug-induced glaucoma was high (20.5%); whereas the rate of recovery from angle-closure glaucoma was high (34.3%).

Table 1 Characteristics of Study Patients

In our analysis, 186 and 106 different drugs were identified as “suspected medicine” which were involved in glaucoma and angle-closure glaucoma, respectively. Of these, the 10 most frequently reported drugs inducing glaucoma were listed in Tables 2 and 3. As shown in Table 2, the most frequently reported drugs were prednisolone (61 reports), pregabalin (37 reports), and ranibizumab (31 reports). Of those, we analyzed ROR of each drug with most frequent route of administration. As shown in Table 4, 10 medications yielded positive signals, with lower CI of ROR of greater than 1. Of note, the association with glaucoma was more noteworthy for fluorometholone eye-drop (ROR, 227.8; 95% CI, 126.1–411.4), triamcinolone acetonide eye-drop (ROR, 201.6; 95% CI, 114.4–355.1), and intravenous foscarnet sodium hydrate (ROR, 57.5; 95% CI, 30.6–107.9).

Table 2 Most Frequently Reported Drugs That Induce Glaucoma

Table 3 Most Frequently Reported Drugs That Induce Angle-Closure Glaucoma

Table 4 Most Frequently Reported Drugs That Induce Glaucoma, focusing on the administration route

As for angle-closure glaucoma, the most frequently reported drugs were atropine sulfate hydrate, followed by scopolamine butylbromide, fluvoxamine maleate, and bromazepam (Table 3). Of those, we analyzed ROR of each drug with most frequent route of administration (Table 5). Angle-closure glaucoma was strongly associated with intravenous atropine sulfate hydrate (ROR, 225.1; 95% CI, 105.1–482.5), inhalational tiotropium bromide hydrate (ROR, 97.4; 95% CI, 36.0–263.5), and oral bromazepam (ROR, 69.7; 95% CI, 30.9–157.5). Even the same reports, there were variety of strength of association of angle-closure glaucoma with inhalational tiotropium bromide hydrate (ROR, 97.4; 95% CI, 36.0–263.5) and oral aripiprazole (ROR, 4.12; 95 CI, 1.53–11.1).

Table 5 Most Frequently Reported Drugs That Induce Angle-Closure Glaucoma, focusing on the administration route

Discussion

This is the first pharmacovigilance study to provide a comprehensive overview of the occurrence and characteristics of patients with drug-induced glaucoma, based on information from the JADER database. In our results, glaucoma and angle-closure glaucoma tended to show a peak age of onset in patients in their 60s and 70s, respectively. Glaucoma was almost equally reported in men and women; whereas, angle-closure glaucoma occurred more often in women. The present study first demonstrated that oral crizotinib, intravenous darbepoetin alfa, and intravenous foscarnet sodium hydrate were associated with glaucoma, and that oral bromazepam, oral brotizolam, and oral milnacipran hydrochloride were associated with angle-closure glaucoma.

In our analysis, it is noteworthy that many of the patients with glaucoma and angle-closure glaucoma were elderly (most frequently in their 60s). These results are consistent with self-reported glaucoma prevalence in elderly patients.¹⁸ Elderly patients often have a significant increase in visual impairment with age. Especially in the elderly, the aqueous humour flow is often reduced, resulting in increased intraocular pressure. Age may also play a role in neuroinflammation balance and in retinal environment.¹⁹

In the present study, sensitive and quantitative method based on the disproportional reporting rate, such as ROR, has been utilized to capture the drug-related risk for a particular ADR. Our findings that steroids such as prednisolone, betamethasone sodium phosphate, and triamcinolone acetonide, and fluorometholone were associated with glaucoma are consistent with previous studies.^20–22 Steroids induce eye pressure with an open angle systematically, topically, or intravitreally.^5,23 In our database, it is possible that reports as “glaucoma” may include “open-angle glaucoma”. Our results revealed that pregabalin induced glaucoma. This is consistent with a nested case-control study that pregabalin use is associated with incidence of glaucoma.²⁴ Animal experiments demonstrated that pregabalin reduce intraocular pressure, and underlying mechanisms are thought to be that pregabalin binds to CACNA2D1, which causes the α1 pore to close, leading to a decrease in Ca²⁺ influx into cells and a resultant decrease in free cytosolic Ca²⁺. Consequently, the cells relax and aqueous humor inflow may be reduced and/or outflow may be increased, leading to a reduction in intraocular pressure.²⁵ As for anti-VEGF agents (bevacizumab, ranibizumab, and aflibercept), our study revealed that ranibizumab was associated with glaucoma, which is consistent with the results of several studies.^26–29 For example, a retrospective cohort study involving nondiabetic patients without pre-existing glaucoma revealed that the incidence of intraocular pressure increase was higher among bevacizumab and ranibizumab users compared with aflibercept users.²⁷ On the other hand, one retrospective and nationwide cohort study showed that no significant differences in the risk of major arterial thromboembolic events and glaucoma were found between ranibizumab and aflibercept.²⁸ Another retrospective chart review of patients receiving intravitreal ranibizumab (0.5 mg) and/or bevacizumab (1.25 mg) injection showed that incidence of delayed and sustained ocular hypertension is low after their single or multiple intravitreal injections.²⁹ Several studies reported that aflibercept is less risk of glaucoma, and our results did not include aflibercept in the list of most frequently reported drugs inducing glaucoma. Among VEGF inhibitors, aflibercept binds to VEGF-B and PlGF in addition to VEGF-A (the target of bevacizumab and ranibizumab), which differentiate its pharmacodynamic effects from bevacizumab and ranibizumab. It is known that VEGF inhibitors can facilitate to up-regulate other growth factors such as PlGF, which acts to sustain pathological angiogenesis and inflammation and is not involved in physiological angiogenic processes.³⁰ Repeated intravitreal injections of ranibizumab and bevacizumab may promote an inflammatory response by increase in PlGF levels in the eye, which could lead to treated eyes being at a higher risk of inflammatory-related intraocular pressure increase than aflibercept. As for tacrolimus hydrate, our findings that it was associated with glaucoma; however, it is not agreement with a retrospective longitudinal study that there was no significant difference in incidence of primary open-angle glaucoma between FK506-treated patients (15%) and non-FK506 group (22%).³¹ Interestingly, our results revealed that oral crizotinib, intravenous darbepoetin alfa, and intravenous foscarnet sodium hydrate were significantly associated with glaucoma. There have been few studies of them involving with glaucoma (Table 6).

Table 6 Comparison of Our Study with Previous Studies Regarding Glaucoma

As for drug-induced angle-closure glaucoma, we found frequently in women (86.0%) and over 60 years, which is line with previous studies showing high prevalence of women and older age in drug-induced angle-closure glaucoma.^32,33 In our results, anticholinergic drugs (atropine and tiotropium), antidepressants (fluvoxamine, bromazepam, aripiprazole, quetiapine, paroxetine, brotizolam, and milnacipran), VEGF inhibitors (aflibercept), and fentanyl ranked in top-10 drugs which is associated with angle-closure glaucoma. This observation is in accordance with previous studies that the main causes of drug-induced angle-closure glaucoma were anticholinergic drug,^34,35 antidepressants,^36–39 VEGF inhibitors,⁴⁰ and fentanyl.⁴¹ Interestingly, there were few studies showing the association of oral bromazepam, brotizolam, milnacipran hydrochloride with angle-closure glaucoma (Table 7). We found that signals were detected in oral bromazepam (ROR, 69.7; 95% CI, 30.9–157.5), oral brotizolam (ROR, 16.6; 95% CI, 6.18–44.8), and oral milnacipran hydrochloride (ROR, 22.8; 95% CI, 8.46–61.4) for angle-closure glaucoma. To the best of our knowledge, this is the first study to show the association of oral bromazepam, oral bromazepam, and oral milnacipran hydrochloride with angle-closure glaucoma.

Table 7 Comparison of Our Study with Previous Studies Regarding Angle-Closure Glaucoma

The JADER database is considered a useful tool to screen potential associations in drug-induced glaucoma; however, several limitations inherent to spontaneous reporting are included. First, the JADER database has various biases, such as the lack of a denominator that indicates the total number of patients who received the drugs of interest, as well as missing data and confounding factors. Second, the ROR does not provide a robust indication of the signal strength. In this kind of study, the ROR corresponds to the risk of spontaneous notification of an ADR and not the risk of glaucoma occurrence per se. Finally, the present method did not provide us with detailed clinical information on the patients’ clinical status such as on unknown etiology of underlying disease and the possible role of disease in the reported adverse event.

Conclusion

In conclusion, the rank-orders of the suspected drugs associated with glaucoma and angle-closure glaucoma were determined using a nationwide pharmacovigilance database. Especially, oral crizotinib, intravenous darbepoetin alfa, and intravenous foscarnet sodium hydrate were associated with glaucoma, and that oral bromazepam, oral brotizolam, and oral milnacipran hydrochloride were associated with angle-closure glaucoma. These data strongly suggest that physicians can be alerted to take precautions against drugs inducing glaucoma and angle-closure glaucoma, select appropriate therapeutic medicine, and potentially avoid glaucoma.