Drugs, Inflammation, and the Eye

Although uncommon, drug-induced uveitis has been associated strongly with a number of medications, including systemic cidofovir, rifabutin, sulfonamides, and bisphosphonates; intraocular cidofovir and vascular endothelial growth factor (VEGF) inhibitors; and topical intraocular pressure (IOP) lower agents, such as brimonidine, metipranolol, and the prostaglandin analogs.1–4 In most instances, such drug-induced ocular inflammation is believed to be a relatively nonspecific immune response to local tissue toxicity. More specific immune reactions appear to be less common, but have been suggested for several agents, including the tumor necrosis factor-alpha (TNFα) inhibitors,5 paradoxically used to treat inflammatory disease, and, the recently introduced, immune checkpoint inhibitors. Currently approved biologic agents in this latter class target either cytotoxic T-lymphocyte antigen-4 (CTLA-4; ipilimumab) or programmed cell death protein 1 (PD-1; pembrolizumab and nivolumab) and are used for the treatment of metastatic melanoma and non-small cell lung cancer.6–8 Three articles in this issue of Ocular Immunology and Inflammation (OII) present important findings related to drugs, inflammation, and the eye,9–11 including two that describe the less well-appreciated occurrence of orbital inflammation.10^,11

Beltz and Zamir9 studied 19 eyes in 12 patients who developed evidence of anterior uveitis while on brimonidine, a topical alpha₂ adrenoreceptor agonist that lowers IOP by suppressing aqueous production. Two patients were taking brimonidine co-formulated with timolol. All patients were seen at a referral hospital in Melbourne, Australia. The duration of brimonidine use prior to the onset of uveitis varied from 1 week to 5 years, with a mean of 19.7 months and a median of 1 year. Concurrent conjunctivitis occurred in 10 eyes (52.6%). Most eyes (16/19; 84.2%) presented with stellate keratic precipitates, whereas three eyes (15.8%) had small mutton-fat precipitates. The precipitates were distributed diffusely across the corneal endothelium in 17 eyes (89.5%) and at a site of a prior surgical incision in the remaining two eyes (10.5%). Overt anterior chamber inflammation was either absent or mild. No patients had synechiae or posterior segment inflammation. Intraocular pressure was elevated by ≥5 mmHg above baseline in eight eyes (42.1%) of five patients. Treatment in all patients involved discontinuation of brimonidine and, in eight eyes (42.1%), included topical 1% prednisolone acetate, 3–4 times daily. Time to resolution following cessation of brimonidine ranged from 4 days to 8 weeks, with a mean of 3.3 weeks and a median of 4 weeks. The authors highlighted the fact that the occurrence of both diffuse, stellate keratic precipitates and elevated IOP in eyes with brimonidine-induced inflammation can lead to misdiagnosis of either Fuchs Uveitis Syndrome or herpetic anterior uveitis.12^,13 Neither of these conditions would be expected to improve with discontinuation of brimonidine alone, however. The concurrence of conjunctivitis, periocular dermatitis, or both, can also provide strong clues to causality and occurs in up to 15% of all patients on brimonidine.14 The authors’ comprehensive review of the literature at the time of acceptance identified an additional 43 eyes of 24 patients, or a total of 62 eyes in 32 patients, with uveitis linked to brimonidine use.

Lefebvre et al.10 described four men and two women who developed orbital inflammation shortly after treatment with a bisphosphonate, a class of agents that inhibits osteoclast-mediated bone absorption. All patients were seen over an 18-month period at a tertiary referral center in Boston, MA, USA. Five of the six patients (83.3%) received intravenous zoledronate, whereas one (16.4%) was treated with oral alendronate. Time from treatment to onset ranged from 1 to 11 days (median 4 days) following zoledronate therapy and occurred at 28 days after alendronate. Inflammation was unilateral in all patients. Imaging revealed isolated rectus muscle involvement in three patients (50%), diffuse orbital involvement in two patients (33.3%), and orbital apex inflammation in one patient (16.7%). Treatment involved discontinuation of the bisphosphonate in all patients and included a short course of oral corticosteroids in two patients (40%), in whom the severity of inflammation and/or symptoms warranted such treatment. Resolution occurred within 4 weeks (median 2 weeks) in the patients treated with zoledronate, but took over 2 months in the patient who received alendronate. A comprehensive review of the literature performed by the authors at the time of acceptance identified an additional 15 patients, or a total of 21 patients, with orbital inflammation linked to bisphosphonate treatment. Bisphosphonate-associated orbital inflammation appears to occur less frequently than uveitis, which has been reported in up to 1% of all patients treated with these agents.15–18 Whereas most patients with bisphosphonate-associated orbital inflammation received intravenous zoledronate and have unilateral findings, intravenous pamidronate has been associated most often with uveitis, which tends to be bilateral. The somewhat unexpected unilaterality of orbital inflammation following systemic bisphosphonate administration, together with the use of bisphosphonates in many patients with comorbidities, such as cancer, led the authors to recommend routine orbital imaging for patients who develop orbital inflammation while on a bisphosphonate. Bisphosphonate-associated scleritis/episcleritis has also been reported, but appears to be less common than uveitis and, as with orbital inflammation, is more often unilateral.

Papavasileiou et al.11 described seven patients with metastatic melanoma who developed ocular or orbital inflammation following treatment with the immune checkpoint inhibitor ipilimumab—a monoclonal antibody directed against CTLA-4. All patients were seen at a tertiary referral center in Boston, MA, USA. Four patients had orbital inflammation (two unilateral and two bilateral); two patients had bilateral anterior uveitis; and one patient had bilateral peripheral ulcerative keratitis. The number of ipilimumab infusions prior to the onset of inflammation ranged from 1 (14.3%) to three (28.6%), with most episodes occurring after two treatments (57.1%). Treatment involved discontinuation of ipilimumab in all patients and adjunctive systemic and/or topical corticosteroids in six patients. One patient with orbital inflammation was left with residual diplopia. The authors were able to successfully retreat three of their seven patients (42.9%) with ipilimumab, but emphasized that these patients were all pre-treated with high-dose systemic corticosteroids and that such retreatment should be considered only when prior immune-related adverse events (IRAE) were mild. The authors observed colitis in patients in their series and, in light of previous reports, suggested that the concurrence of colitis is patients with uveitis may be common. A comprehensive review of the literature performed by the authors at the time of acceptance identified an additional 16 patients, for a total of 23 patients, with ocular or orbital inflammation linked to ipilimumab. In addition to keratitis, uveitis, and orbital inflammation, reported ocular complications have included conjunctivitis, amaurosis fugax in the setting of giant cell arteritis, and a Vogt–Koyanagi–Harada disease-like syndrome. In a recent meta-analysis of CTLA-4 inhibitors used in 1265 patients from 22 clinical trials, Bertrand et al. found the overall incidence of all-grade IRAEs to be 72% (95% CI, 65–79%) and the incidence of high-grade IRAEs to be 24% (95% CI, 18–30%). Dermatologic IRAEs were most common and included rash, pruritus, and vitiligo. Other IRAEs included colitis and, less frequently, hepatitis, hypophysitis, and thyroiditis. Cases of sarcoidosis, Guillain–Barré syndrome, and uveitis were also identified, but were described in the meta-analysis as rare.8 Uveitis has also been reported following treatment with immune checkpoint inhibitors directed against PD-1, including both pembrolizumab19^,20 and nivolumab.21 Tremelimumab, like ipilimumab, inhibits CTLA-4, whereas pidilizumab, like pembrolizumab and nivolumab, inhibits PD-1, but neither of these agents has yet to complete clinical trials.

While the reports published in this issue of OII fulfilled many of the criteria required to establish these medications as probable causes of ocular or orbital inflammation, none documented recurrence of inflammation following drug re-challenge – the most definitive means of establishing causality.1–4 Other reports have provided this critical piece of evidence, however, and it now appears clear that use of brimonidine, the bisphosphonates, and immune checkpoint inhibitors can be associated with ocular inflammation in a small minority of patients. Together, these articles highlight both the expanding spectrum of drug-induced ocular inflammation and the importance of a thorough drug history in the evaluation of all patients with ocular or orbital inflammation.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Funding

Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC).