Rituximab for treatment of non-infectious and non-malignant orbital inflammatory disease

Purpose To provide a comprehensive review of rituximab use for the treatment of non-infectious/non-malignant orbital inflammation. Methods Review of literature through January 2021. Results Individual data was available for 167 patients with refractory non-infectious/non-malignant orbital inflammation who received treatment with rituximab (RTX). Rituximab was generally utilized as third-line or later treatment (108/149, 72.5%) at a mean of 44.6 months following the diagnosis of orbital inflammation (range = 0 to 360 months; median = 13.7 months). Patients with non-infectious/non-malignant orbital inflammation either received prior treatment with corticosteroids only (27/122, 22.1%), or with one (31/122, 25.4%), two (25/122, 20.5%), or three or more (25/122, 20.5%) corticosteroid-sparing immunosuppressive agents with or without corticosteroids before initiation of RTX treatment. The rheumatologic protocol (two infusions of 1 gram of RTX separated by 14 days) was utilized most frequently (80/144, 55.6%), followed by the oncologic protocol (four weekly infusions of 375 mg/m2 RTX; 51/144, 35.4%). Various other off-label regimens were used infrequently (13/144, 9.0%). Rituximab treatments resulted in a positive therapeutic response for the majority of patients with orbital inflammation (146/166, 88.0%). Commonly treated diagnoses included granulomatosis with polyangiitis (99/167, 59.3%), IgG-4 related disease (36/167, 21.6%), and orbital inflammation of indeterminate cause (25/167, 15.0%). No side effects were reported in 83.3% (55/66) of cases. The most common RTX-induced adverse event was an infusion-related temporary exacerbation of orbital disease (4/66, 6.1%), which occurred prior to the routine use of systemic corticosteroids as pre-conditioning. Conclusions Overall, RTX appears to be both efficacious and well-tolerated as second- or third-line therapy for patients with non-infectious/non-malignant orbital inflammation.


Introduction
Non-infectious/non-malignant orbital inflammation accounts for 6-27% of orbital diseases [1,2]. Possible etiologies include thyroid disease, granulomatosis with polyangiitis (GPA), Churg-Strauss syndrome, polyarteritis nodosa, atypical Cogan syndrome, temporal arteritis, Kawasaki syndrome, Behçet disease, sarcoidosis, Sjögren syndrome, systemic lupus erythematosus (SLE), rheumatoid arthritis, dermatomyositis, and IgG4-related disease [3]. A sizable proportion of cases will remain idiopathic or of indeterminant etiology [2]. High-dose corticosteroids have been utilized as first-line agents due to their availability and efficacy in inducing disease remission [4,5], but prolonged use is limited by side effects [6]. Patients with disease non-responsive to corticosteroids or requiring long-term immunosuppression have been treated with methotrexate, infliximab, cyclosporine-A, radiotherapy, mycophenolate mofetil, interferon-A, tacrolimus, rituximab (RTX), cyclophosphamide, chlorambucil, leflunomide, and azathioprine, but there is currently no consensus on treatment regimen [7]. The use of RTX has been reported increasingly and with generally good efficacy in 167 cases of refractory non-infectious/non-malignant orbital inflammation, as summarized below.

Methods
The authors conducted a literature search using the National Library of Medicine's PubMed database for all English language articles published through January 2021 with the following search terms: "rituximab AND eye", "rituximab AND orbital inflammation," "rituximab AND pseudotumor," and "rituximab AND orbital granuloma." Relevant references within these articles were also reviewed. Included here were all cases of non-infectious/ non-malignant orbital inflammation for which individual case data was available. Articles describing large series of patients in which individual case data was not provided were excluded from the current analysis, but were read for content and citations, and were referenced when appropriate. Individual information on patient age, sex, anatomical localization and cause of disease, prior treatments, time from diagnosis to initiation of RTX, pre and post treatment visual acuities, RTX treatment regimen (dosage and cycles), therapeutic response, additional treatments, duration of follow-up and whether disease recurrence occurred, and adverse events attributed to RTX were collected when available. In this review, a patient was considered to have had a positive therapeutic response to RTX if they achieved disease quiescence, or if the authors subjectively documented improvement in disease severity. Line of therapy was tallied according to the following criteria: first-line -RTX initiated before or at same time as corticosteroids; second-line -RTX initiated after corticosteroidseither alone or with a second more traditional immunosuppressive agent; third-line or greater -RTX initiated after a non-corticosteroid immunosuppressive agent, such as nonbiologic or biologic disease modifying antirheumatic drug, alkylating agents, or intravenous immunoglobulins, with or without corticosteroids. Treatment regimens were classified into either the rheumatologic protocol (two doses of 1000 mg separated by 14 days) [8], the oncologic protocol (four doses of 375 mg/m 2 weekly) [9], or "other" category for the less commonly utilized dosing protocols. Univariate comparisons were made with two-tailed T-test and nomimal, uncorrected p-values were reported.
Rituximab use in thyroid orbitopathy was not included in this review as it has been reviewed previously [10][11][12]. Use of RTX for non-infectious uveitis and scleritis was summarized in a separate companion review [13].

Granulomatosis with Polyangiitis (GPA)
Patients with GPA accounted for 59.3% (99/167) of the overall cohort with non-infectious/non-malignant orbital inflammation who received treatment with RTX. There was a 0.89 to 1 male to female ratio, and mean age at time of treatment was 42.7 years (range = 4 to 72 years; median = 42.0 years). The mean interval from diagnosis of GPA to initiation of RTX treatment was 59 [47,48,51]. One case each (1/38, 2.6%) of de novo hepatitis [24], nausea [18], and adenovirus pneumonitis that resulted in death [15] was also described.

Discussion
Rituximab, a fully humanized monoclonal anti-CD20 antibody, has increasingly been utilized as a safe and efficacious treatment of refractory, non-infectious/nonmalignant orbital inflammation arising from GPA, IgG4related disease, and indeterminate etiologies. Rituximab was used as a third-line or later treatment in nearly three-quarters of patients with non-infectious orbital inflammation who had been unresponsive to corticosteroid or other corticosteroid-sparing immunosuppressants, with treatment generally initiated nearly 4 years following initial diagnosis. Half of these patients received the rheumatologic protocol, two fifths the oncologic protocol, and nearly one fifth various off-label dosing regimens. More than three fifths of treated patients received only 1 cycle of RTX, while one fifth received 2 cycles, and one sixth three or more cycles. Patients with orbital inflammation had a positive therapeutic response in nearly nine out of 10 cases, with more than three fifths experiencing disease remission or regression, and one sixth reporting generally mild adverse events, the most common of which were infusion reactions.
Use of RTX for treatment of non-infectious/non-malignant orbital inflammation was efficacious for a variety of underlying systemic causes. All of the patients with orbital inflammation secondary to both IgG4-related disease and AOXGD (n = 4, 100%) had a positive therapeutic response, while somewhat less efficacy was observed for patients with orbital inflammation attributed to IgG4-related disease (n = 35, 94.3%), for those with GPA (n = 99, 84.8%) and those with an indeterminate etiology (n = 25, 88.0%). Pair-wise differences in rates of positive therapeutic response for the most common underlying causes of non-infectious/non-malignant orbital inflammation did not reach a nominal p-value of less than 0.05. The overall efficacy of RTX for treatment of orbital inflammation (88.0%) was comparable to what has been reported for non-infectious uveitis (83.5%) and scleritis (93.3%) [13].
While a variety of RTX dosing regimens and cycles were utilized to treat non-infectious/non-malignant orbital inflammation, the rheumatologic protocol was both most commonly utilized and numerically most efficacious in reducing disease severity, while the overall number of treatment cycles had no discernable impact on outcomes. Compared to patients who received the oncologic protocol, which was the second most common treatment regimen, patients given the rheumatologic protocol had a significantly higher chance of a positive therapeutic response (nominal p = < 0.0001, N-1 twoproportion test, two-tailed). The remaining subjects received various other off-label dosing regimens, but the collected data was insufficient to allow for comparison due to limited sample sizes and lack of standardization across the studies. Excluding the difference in proportion of patients who achieved continued disease quiescence following one and two RTX cycles (nominal p < 0.015, N-1 two-proportion test, two-tailed), pair-wise comparisons of the rates of positive therapeutic response, sustained disease remission, and drug-free remission between patients given one, two, and three or more cycles of RTX failed to achieve a nominal p-value of < 0.05. Patients who received one RTX cycle leading to disease remission and were eventually given a second RTX cycle for disease relapse (n = 5; mean of 26.2 months) were categorized into the two-cycle group, which may explain the significant difference in rates of sustained disease remission between the two subgroups. Overall, these findings suggest the rheumatologic protocol could be considered for the treatment of non-infectious/non-malignant orbital inflammationparticularly those patients who fail first-line immunosuppression therapy.
The majority of patients with non-infectious/non-malignant orbital inflammation received RTX as a third-line agent, generally after unsuccessful treatments with corticosteroid and traditional non-corticosteroid immunosuppressive agents. Possible contributing factors for this practice pattern could have been the cost of RTX, a relatively more difficult route of administering RTX, or provider unfamiliarity with RTX as a treatment choice in such patients. Of note, RTX was more likely to be used earlier in the treatment course for patients with IgG4-related disease as compared to patients with GPA. While nearly all patients with orbital inflammation from GPA received RTX as third-line treatment (79/82, 96.3%), those with IgG4-RD received RTX as a third-line agent in just over a third of cases (13/36, 36.1%). In total, 71.4% (10/14) of all cases of orbital inflammation treated with RTX as first-line therapy and 63.0% (17/27) of those given RTX as second-line therapy possessed orbital inflammation attributed to IgG4-related disease alone or concurrently with another etiology. To our knowledge, Plaza et al. first described in 2010 the use of RTX as a first-line immunosuppressant to treat six patients with IgG4-related orbital inflammation, but did not elaborate on their treatment approach [34]. Later, Murakami et al. promptly used RTX to treat a patient with IgG4-related orbital inflammation due to the presence of elevated serum IgG and abnormal lymphoid tissue found on histology [39], and Pomponio et al. used RTX as a first-line agent to treat orbital inflammation from concurrent IgG4-related disease and AOXGD due to the presence of elevated serum plasmablasts and the desire to avoid potential corticosteroid related side effects in a post-menopausal woman [63]. The generally positive outcomes noted here suggest that RTX might be considered earlier in the course of therapy in some patients.
Clinical trials investigating RTX treatment for rheumatologic diseases have found it to be relatively welltolerated with mild to moderate infusion-related reactions as the most common adverse response [68][69][70]. The known severe adverse reactions include tumor lysis syndrome, severe mucocutaneous reactions, progressive multifocal leukoencephalopathy, hepatitis B reactivation, infections, cardiac arrhythmias, renal toxicity, and bowel obstruction and perforation [71]. While there have been a large number of reports detailing the efficacy of RTX in treatment of orbital inflammation, less than half (66/167, 39.4%) included information on drug safety, with onesixth (11/66, 16.7%) of those reported to have experienced an adverse event. According to the common terminology criteria for adverse events, 90.9% (10/11) of reported RTX-associated side effects were grade 1 to 3 reactions. However, one patient (1/11, 9.1%) developed adenovirus pneumonitis 4 months after his second dose of RTX that ultimately caused his death (grade 5). An infusion reaction leading to exacerbation of orbital inflammation, was described early in the treatment course for four patients, three of that later achieved disease remission following RTX treatment. These authors proposed the use of systemic corticosteroids before, during, and after RTX infusions so as to lessen or prevent such inflammatory infusion reactions [36,59,65]. Following orbital exacerbations in the first three of four RTX-treated patients, Suhler et al. implemented use high-dose oral prednisone for 3 days before and after RTX infusions, and prevented further inflammatory flares at and after infusion [36]. Overall, the rate of reported adverse events was similar to or lower than those reported in clinical trials, similar to the rates gleaned from a similarly conducted retrospective review of the use of RTX for non-infectious uveitis (23.7%; nominal p = 0.28, N-1 two-proportion test, two-tailed) and scleritis (14.5%; nominal p = 0.71, N-1 two-proportion test, two-tailed) [13].
The retrospective nature of our analysis was limited by variations in details regarding prior treatment, the definition of positive therapeutic response, treatment protocols, rationale for retreatment, total number of RTX cycles, duration of follow-up and whether disease relapse occurred, and of the occurrence of adverse events. Several larger series provided overall cohort statistics only and so could not be incorporated into our calculations. Finally, the efficacy of the studies reviewed here may not reflect a broader population-based sample of refractory non-infectious/non-malignant orbital inflammation due to referral, selection, treatment, evaluator, and/or publication bias [72]. Safety and adverse events data was explicitly provided for fewer than half of cases.

Conclusions
Rituximab appears to be an effective and well-tolerated treatment option for patients with non-infectious/nonmalignant orbital inflammation. Authors generally favored use of one or two RTX cycles according to the rheumatologic or oncologic dosing. Our findings suggest that the rheumatologic protocol may be superior in inducing disease remission, but similar reported rates of positive response and sustained disease remission were achieved in those given limited and extended treatment courses. Adverse events were reported in about one-sixth of treated patients and tended to be mild to moderate infusion reactions. Overall, the current ophthalmologic literature strongly supports RTX use for non-infectious/non-malignant orbital inflammation refractory to corticosteroid and traditional non-corticosteroid immunosuppressive agents, especially GPA and IgG4-related disease.