The Promise and Limitations of TNF-α Inhibition for Uveitis

Since their introduction more than 15 years ago, tumor necrosis factor-alpha (TNF-α) inhibitors have assumed an important role in the management of moderate to severe systemic autoimmune disease, including rheumatoid arthritis, seronegative spondyloarthropathies, inflammatory bowel disease, psoriatic arthritis, and juvenile idiopathic arthritis (JIA). While none of the TNF-α inhibitors is specifically approved by either the Food and Drug Administration (FDA) or the European Medicines Agency (EMEA) for the management of isolated ocular inflammation, anti-TNF-α agents are now used widely off-label for the management of a number of non-infectious inflammatory eye disorders.1–3 Three articles in this issue of Ocular Immunology & Inflammation highlight both the growing importance and limitations of TNF-α inhibitors in the treatment of uveitis.4–6

Shakoor and colleagues describe their experience with attempts to discontinue infliximab in a group of 18 patients seen at the Francis I. Proctor Foundation between 1998 and 2010, 27.7% of the total infliximab treated cohort (18 of 65) cared for at the Proctor Foundation during the period of the study.4 All 18 patients achieved control of inflammation, which the authors defined as up to 0.5+ anterior chamber and/or vitreous inflammation, but no active vasculitis, retinitis, or choroiditis.. The uveitis was idiopathic in eight patients (44.4%), associated with JIA and Behçet’s disease in four patients each (22.2%), and with systemic lupus erythematosus and Vogt-Koyanagi-Harada (VKH) disease in one patient each (5.5%). Of note, the discontinuation was for presumed remission in only 9 patients (50%), with other, unplanned reasons for discontinuation including occurrence of adverse events or intolerance, poor compliance, and pregnancy. Fifteen patients (83.3%) were on either methotrexate (MTX) or mycophenolate mofetil, and three (16.7%) were on both an antimetabolite and 5mg/day of prednisone following discontinuation. The median time on infliximab prior to discontinuation was 1.3 years (interquartile range 0.7–2.4 years). Eleven patients (61.1%) experienced recurrence or worsening of their inflammation after stopping infliximab at a median time of approximately 20 months (95% CI 8 days to 4 years), with just under two-thirds of recurrences occurring within 90 days and nearly three-quarters within one year. Small numbers in the subgroup analyses aside, the recurrence rate appeared to be particularly high in patients with JIA (4 of 4; 100%) and panuveitis (6 of 7; 85.7%), and time to relapse was faster in patients with JIA and in those who took longer to achieve initial control on infliximab. Inflammation recurred in half of the patients with idiopathic uveitis and Behçet’s disease (2 of 4 each), and in the single patient with VKH disease. Unfortunately, the authors did not describe whether patients who discontinued for presumed remission did better than those who were discontinued for other, unplanned reasons. This point notwithstanding, the overall rate of successful, sustained remission following discontinuation was relatively low, included 38.9% of the patients who discontinued (7 of 18), and 10.8% of the entire cohort of those treated with infliximab at the Proctor Foundation during the period of the study (7 of 65). These results complement those previously published by the Proctor group7 and others,2,3,8 including both the therapeutic effectiveness of anti-TNF-α agents and the relatively low rate of sustained drug-free remission in patients with JIA and Behçet’s disease.9 Of note, Deuter and colleagues have reported a drug-free remission rate of nearly 90% in patients with Behçet’s disease whose uveitis was treated and controlled with interferon-α,10 an agent used infrequently outside of Europe.11,12

Miserocchi and associates describe their experience using golimumab, a humanized, subcutaneously administered anti-TNF-α agent engineered from infliximab, in 17 patients with severe, refractory uveitis, including 13 with JIA and 4 with HLA-B27-associated uveitis.5 Eight patients (47.1%) had anterior uveitis and nine (52.9%) panuveitis. All patients included in the study were inadequately controlled by one or more TNF-α inhibitors, rituximab, or abatacept. Overall, 14 patients (82.4%) showed a positive therapeutic response to golimumab, defined as an improvement of at least 1.0+ step in inflammation, and 85.7% of those who responded (12 of 14) were inactive at last clinic visit (mean follow-up 25.1 months; range 18–29 months). Concurrent immunosuppressive agents used at last visit among these 14 patients included MTX in 42.9% (6 of 14) and low-dose prednisone in 28.6% (4 of 14; two of whom were also on MTX). Serious adverse events reported in the cohort included bacterial pneumonia in one patient, which prompted discontinuation of golimumab, and a transient, desquamative, perioral dermatitis that resolved spontaneously without discontinuation of treatment in a second. This same group13 and others14–17 have reported similarly promising results with the use of golimumab for severe, refractory uveitis – most notably in patients with Behçet’s disease, JIA, and seronegative spondyloarthropathy-associated uveitis.

Tsang and colleagues describe a 12-year-old boy with severe JIA-associated uveitis incompletely controlled on prednisone, MTX, infliximab, adalimumab, and abatacept, but who ultimately responded favorably to MTX in conjunction with tocilizumab, a monoclonal antibody targeting the interleukin-6 (IL-6) receptor.6 While the patient’s uveitis was ultimately controlled when tocilizumab was dosed 8 mg/kg every 3 weeks, he had a recurrence when dosed every 4 weeks, the indicated treatment interval, and developed transient neutropenia when dosed every 2 weeks. Additional adverse events associated with tocilizumab include infection, thrombocytopenia, and elevated liver enzymes. Others have described similarly beneficial responses to tocilizumab in selected patients with refractory uveitis.18–23

Taken together, these findings support the notion that the use or addition of an anti-TNF-α agent can help achieve control of uveitis in a high proportion of patient with uveitis, particularly those with Behçet’s disease and JIA, two conditions for which TNF-α inhibitors are now strongly recommended as standard of care in the step-wise management of ocular inflammation.24 The studies cited above also highlight, however, some of the limitations of TNF-α inhibitors, including the unpredictably varied response of each TNF-α inhibitor in an individual patient, both initially and over time, and the disappointingly low rate of successful discontinuation. This appears to be particularly true in patients with Behçet’s disease and JIA – those for whom anti-TNF-α agents appear to be especially effective. Fortunately, a number of other development stage and currently approved agents, such as tocilizumab, appear to show promise for the treatment of moderate to severe, treatment-refractory uveitis.25,26

Declaration of interest

The authors have no financial conflicts. Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC).

Acknowledgements

We thank Dr. Nisha R. Acharya, Dr. Elisabetta Miserocchi, and Dr. Adrian Tsang for thoughtfully reading and commenting on an early version of this editorial.