Uveitis Treatment Trials—A Cross-study Perspective

The previous issue of Ocular Immunology and Inflammation (OII) contains an important and timely report by Martel and colleagues on the 10-year experience of the Francis I. Proctor Foundation with the use of two TNF-α inhibitors, infliximab (Remicade) and adalimumab (Humira), for the treatment of chronic noninfectious uveitis.¹ A considerable strength of the study was the use of standard outcome measures,² including key Kaplan-Meier survival, or responder, analyses over the follow-up period of the study to account for discontinuations.³ These analyses included the proportion of patients with sustained control of inflammation, defined as no or slight intraocular inflammation, as well as the proportion of patients achieving control of inflammation as defined above plus corticosteroid-sparing success, defined as <10 mg/day of prednisone or its equivalent for other corticosteroids. In both instances, success required consistent measurements taken on two separate visits separated by at least 28 days. Additional important data reported included use and tapering of adjunctive immunosuppressive therapies, and drug discontinuation rates and causes.

Roughly 8.5% of patients seen at the Proctor Foundation from 1999 to 2009 were treated with infliximab or adalimumab, among whom the eligibility criteria of the study were met on 41 patients, including 29 treated with infliximab, 10 treated with adalimumab, and 2 who received therapy with 1 followed by the other agent. For infliximab, sustained control of inflammation at 3, 6, and 12 months was achieved in 56, 82, and 70%, respectively, whereas for adalimumab the rates were 38, 75, and 57%, respectively. At these same time points, sustained control of inflammation plus corticosteroid sparing success was achieved in 33, 61, and 61% of patients on infliximab and 38, 63, and 57 on adalimumab. The median time to sustained control was 63 days with infliximab and 151 days with adalimumab.

The discontinuation rate for both groups was 0.26/person-year, with most discontinuations occurring more than 1 year after initiation of therapy. Twenty-five percent of the discontinuations in the infliximab group were due to adverse reactions, including two serious events—pulmonary coccidioidomycosis and bronchial spasm—whereas none of the adalimumab discontinuations were related to adverse reactions. While the use of both corticosteroids and noncorticosteroid immunosuppressive agents tended to decrease in the two groups during the 12 months of the study, patients on infliximab appeared, over time, to be slightly less likely to be on corticosteroids and slightly more likely to be on a noncorticosteroid immunosuppressive agent in addition to the TNF-α inhibitor.

Overall, the results of the Proctor study support growing evidence from smaller retrospective case series that both infliximab and adalimumab can be effective therapeutic adjuncts for patients with chronic noninfectious uveitis refractory to conventional immunotherapy,^5,12 with a reasonable proportion of patients receiving either agent achieving both sustained control of inflammation and corticosteroid-sparing success at 6 and 12 months. TNF treatment appears to come at a real, albeit low, risk of life-threatening, serious adverse events,⁴ a finding similarly supported by the Proctor patients, who experienced an overall proportion of serious adverse events approaching 5%.¹

Isolated therapeutic response rates of the sort described above provide little direct guidance regarding which agent or approach to use for any given patient, however. Moreover, cross-trial comparisons are generally invalid due to differences in study populations, treatments used, and inclusion/treatment bias.^13–15 Such caveats notwithstanding, the growing use of standardized Kaplan-Meier outcome metrics in a number of recent uveitis treatment studies invites a broader interpretation, albeit with caution. Specifically, the Systemic Immunosuppression Therapy for Eye Disease (SITE) Study Group has now published long-term, retrospective, multicenter data on over 8500 patients with noninfectious ocular inflammatory disease, including uveitis, followed at selected tertiary referral centers in the United States.¹⁶ The reports have summarized the investigators experience with 5 different immunosuppressive agents to-date: azathioprine,¹⁷ methotrexate,¹⁸ cyclophosphamide,¹⁹ cyclosporine,²⁰ and mycophenolate mofetil.²¹ Similarly, the Multicenter Uveitis Steroid Treatment (MUST) Trial Research Group has recently published 24-month, randomized, multicenter trial data comparing systemic therapy (corticosteroids plus immunosuppression when indicated) to Bausch & Lomb’s fluocinolone acetonide intravitreal implant, 0.59 mg (Retisert; implant therapy) for noninfectious intermediate, posterior, or panuveitis.^22–24 The proportion of patients achieving sustained control of inflammation, sustained control of inflammation plus corticosteroid-sparing success, and discontinuing therapy in each of these studies is summarized in Table 1.

Table 1  Kaplan-Meier estimates of the proportion of patients (95% confidence intervals) with control of inflammation and control of inflammation plus corticosteroid-sparing success as reported at various time points across different studies.

Study	Control of inflammation^a			Control of inflammation^a plus corticosteroid sparing success^b				Discontinuation rate/person-year (PY)^c
	3 months (%)	6 months (%)	12 months (%)	24 months (%)	3 months (%)	6 months (%)	12 months (%)
SITE Study Group^d—Azathioprine^Citation17		57 (46–69)	73 (61–83)			32 (24–43)	47 (37–58)	0.45/PY (0.37/PY–0.55/PY)
SITE Study Group—Methotrexate^Citation18		63 (56–70)	74 (68–80)			37 (32–43)	58 (52–65)	0.42/PY (0.37/PY–0.49/PY)
SITE Study Group—Cyclophosphamide^Citation19		59 (51–67)	83 (75–90)			30 (24–38)	61 (53–70)	0.85/PY (0.72–0.99)
SITE Study Group—Cyclosporine^Citation20		61 (54–68)	76 (70–83)			22 (18–27)	36 (31–42)	0.26/PY (0.23/PY–0.31/PY)
SITE Study Group—Mycophenolate mofetil^Citation21		78 (68–86)	91 (83–97)			41 (34–49)	55 (47–64)	0.28/PY (0.22/PY–0.35/PY)
MUST Trial Research Group^e—Fluocinolone acetonide 0.59mg (Retisert®)	74 (66–82)	79 (71–86)	85 (78–90)	88 (80–94)
MUST Trial Research Group—Systemic Therapy^Citation24	53 (43–63)	56 (46–65)	64 (55–74)	71 (61–79)
Proctor Study^e—Infliximab^Citation1	56 (35–75)	82 (63–94)	70 (47–87)		33 (17–54)	61 (41–79)	61 (39–80)	0.26/PY
Proctor Study—Adalimumab	38 (9–76)	75 (35–97)	57 (18–90)		38 (9–76)	63 (25–92)	57 (18–90)	0.26/PY

^aThe definition of control of inflammation differed in different studies and, as shown above, was either defined as no inflammation (MUST Trial) or as no or slight inflammation (SITE Cohort Studies, Proctor Study) on two separate visits separated by at least 28 days.

^bCorticosteroid sparing success defined as ≤10 mg/day.

^cDiscontinuation rates as shown include all reasons, including remission, which was generally low for all agents except cyclosphosphamide, where the discontinuation rate due to remission was 0.32/PY (0.24/PY.0.41/PY). Discontinue due to adverse events was generally one-third the overall discontinuation rate, except for cyclophosphamide, where discontinuation due to adverse events accounted for nearly half of all discontinuations.

^dSITE Study Group rates for all chronic noninfectious ocular inflammation, including, but not limited to, anterior, intermediate, and posterior/panuveitis. Includes scleritis, mucous membrane pemphigoid, and other less-common conditions.

^eMUST Trial Research Group rates for chronic noninfectious intermediate, posterior, and panuveitis.

^fProctor study for chronic noninfectious uveitis, including anterior, intermediate, anterior and intermediate, and posterior/panuveitis.

Several important points may be gleaned from this cross-study perspective. First, moderate success at controlling inflammation and lowering overall systemic corticosteroid exposure may be achieved in most patients with acceptable risk using a number of different agents and approaches. This is fortunate, since the relative indications and contraindications for choosing any given immunosuppressive agent vary considerably from patient to patient, placing a premium on therapeutic flexibility. It should be noted, however, that the success rates shown in Table 1 were generated by uveitis specialists with extensive experience in the use of both corticosteroid and noncorticosteroid immunosuppressive therapies. Lower rates of success, particularly with the use of immunosuppressive agents to achieve acceptable levels of corticosteroids, may be expected when severe uveitis is treated by nonspecialists.^25,26 It should also be noted that the SITE Cohort Study rates in Table 1 were overall rates for all forms of chronic noninfectious uveitis, scleritis, mucous membrane pemphigoid, and other less-common noninfectious ocular inflammatory disorders, whereas the MUST Trial was for chronic noninfectious intermediate, posterior, and panuveitis and the Proctor study was for all forms of chronic noninfectious uveitis. Treatment success rates for any given approach might differ for specific uveitic entities or complications.

Second, response rates for controlling inflammation and for corticosteroid-sparing success generally increased over the first 12 months of therapy. This was most convincingly demonstrated for the implant therapy arm of the MUST Trial, at least with regard to controlling inflammation.²⁴ Use of the Retisert implant in this trial was, however, associated with nearly universal cataract formation in phakic eyes and a cumulative percent use of IOP-lowering medications and surgery of 61 and 26%, respectively, over the period of the study. These data are consistent with published data from the 3-year, prospective, multicenter Retisert implant trial, which reported an overall glaucoma surgery rate of 40%,²⁷ as well as with data from other prospective and retrospective series.^28–31

Response rate trends from both the SITE Cohort Study reports and the Proctor study suggest similar, although perhaps less dramatic, 12-month improvements in controlling inflammation and corticosteroid sparing success when systemic immunosuppressive therapy was used. In addition, the SITE study data, derived from the same treatment centers using the same analytic approach,¹⁶ suggest the intriguing possibilities, by no means substantiated statistically, that the modest success at controlling inflammation with cyclosporine²⁰ may have come at the cost of increased corticosteroid exposure, and that mycophenolate mofetil²¹ may have been particularly effective at controlling inflammation by 12 months. Cyclophosphamide¹⁹ also appeared to be effective at controlling inflammation, but in a different patient population and with what appeared to be more toxicities and a higher discontinuation rate relative to other agents. Of note, cyclophosphamide produced drug-free remission in 61 of the 215 (28%) patients studied, an uncommon occurrence with antimetabolites, cyclosporine, or TNF-α inhibitors. The extent to which cyclophosphamide-induced remissions occurred in patients with uveitis vs. other ocular inflammatory diseases was not reported (Table 1).

The third conclusion is that, even in quite experienced hands, inflammation remained incompletely controlled in 30–40% of patients. Moreover, in patients who did achieve control, some required >10 mg/day of prednisone or its equivalent to achieve that goal^1,17–21—this despite the generally held view by such specialists that an acceptable maintenance dose of oral prednisone is below 10 mg/day.³² While the cause of such partial patient responses is clearly complex and undoubtedly includes incomplete individual responses to the specific agents and approaches used in the various studies, including relative resistance to corticosteroids,³³ perhaps we should all give more consideration to other less generally used options for treating patients with chronic sight-threatening uveitis, such as short-term, high-dose chlorambucil^34–36 and systemic interferon therapy,^37,38 both of which, like cyclophosphamide, appear to offer the possibility of drug-free remission.

The importance of considering all approaches to the treatment of chronic noninfectious uveitis and its complications is highlighted in a second article in this issue of Ocular Immunology and Inflammation, by Butler and colleagues,³⁹ entitled “Interferon alpha 2b in the treatment of uveitic cystoid macular edema.” While interferon therapy for uveitis generally is not used in the United States, these authors reported dramatic and sustained improvement in both central macular thickness and vision in 8 eyes of 4 patients with longstanding, refractory CME following treatment with subcutaneously injected interferon alpha 2b. Of note, all 4 patients had previously failed to respond to various combinations of local and systemic corticosteroids together with systemic noncorticosteroid immunosuppressive drugs. Systemic symptoms such as malaise, arthralgias, fatigue, nausea, and anorexia occurred in each case, but in no patient were the symptoms so severe that interferon therapy was discontinued. Similar success has been reported in Europe with interferon alpha 2a for the treatment of both uveitic^40,41 and pseudophakic macular edema (Irvine-Gass syndrome).⁴²

Taken together, the two studies highlighted here offer further support for the use of biologic agents, specifically the TNF-α inhibitors and interferon, in the treatment of sight-threatening uveitis and its complications. They also point to the benefits of using standardized outcome metrics in reporting and interpreting uveitis treatment studies, such as control of inflammation, corticosteroid-sparing success, and drug discontinuation rates and causes, including specific mention of the rate and durability of drug-free remission.

ACKNOWLEDGEMENT

Supported in part by The Pacific Vision Foundation (ETC), The San Francisco Retina Foundation (ETC), and an unrestricted grant from Research to Prevent Blindness, New York, NY, USA (DAG). We thank Nisha R. Acharya, John H. Kempen, and James T. Rosenbaum for thoughtfully reading and commenting on an early version of this editorial. Three and six month response rates and three, six, nine, and twelve month confidence intervals for the MUST Trial arms provided by Elizabeth Sugar, PhD on behalf of the Must Trial Research Group.

Declaration of interest: DAG: Advisory board, research support, honoraria: Abbott, Allergan, Bausch & Lomb, GSK, Lux biosciences, Santen. MZ: Advisory board: Allergan, Bausch & Lomb, GSK, Lux Biosciences. ETC has no financial interests.