Treatment of Posterior Uveitis With a Fluocinolone Acetonide Implant

CLINICAL TRIALSSECTION EDITOR :  ROY W. BECK, MD, PhD

Treatment of Posterior Uveitis

With a Fluocinolone Acetonide Implant

Three-Year Clinical Trial Results

David G. Callanan, MD; Glenn J. Jaffe, MD; Daniel F. Martin, MD; P. Andrew Pearson, MD; Timothy L. Comstock, OD

Objectives:  To evaluate the safety and efficacy of 0.59-mg and 2.1-mg fluocinolone acetonide (FA) intravitreous implants in noninfectious posterior uveitis.

Design:  A 3-year, multicenter, randomized, historically controlled trial of the 0.59-mg FA intravitreous implant in 110 patients and the 2.1-mg FA intravitreous implant in 168 patients.

Main Outcome Measures:  Recurrence rate, vision, and complications.

Results: Uveitis recurrence was reduced in implanted eyes from 62% (during the 1-year preimplantation period) to 4%, 10%, and 20% during the 1-, 2-, and 3-year postimplantation periods, respectively, for the 0.59-mg dose group (P_.01) and from 58% to 7%, 17%, and 41%, respectively, for the 2.1-mg dose group (P_.01).

More implanted eyes than nonimplanted eyes had improved visual acuity (P_.01). Implanted eyes had higher incidences of intraocular pressure elevation (_10 mm Hg) than nonimplanted eyes (P_.01), and glaucoma surgery

was required in 40% of implanted eyes vs 2% of nonimplanted eyes (P_.01). Cataracts were extracted in 93% of phakic implanted eyes vs 20% of phakic nonimplanted eyes (P_.01).

Conclusions:  The FA implant significantly reduced uveitis recurrence and improved or stabilized visual acuity in subjects with noninfectious posterior uveitis.

 Most subjects required cataract extraction, and a significant proportion required intraocular pressure–lowering surgery.

Application to Clinical Practice:  The FA implant provides an alternative therapy for prolonged control of inflammation in noninfectious posterior uveitis.

Trial Registration: clinicaltrials.gov

Identifier:

NCT00407082

Arch Ophthalmol. 2008;126(9):1191-1201

UVEITIS IS OFTEN A CHRONIC

disease in which longstanding inflammation can result in structural damage and loss of vision. The prevalence of uveitis has been estimated at 38 to 730 people per 100 000 worldwide1 and approximately 115 people per 100 000 in the United States.2 Posterior uveitis is responsible for most uveitis cases that result in blindness.3 In most cases, vision loss arises from cumulative damage to ocular tissues that results from recurrent or chronic inflammation rather than from isolated or acute inflammatory episodes.4 Control of inflammation in noninfectious posterior uveitis (NIPU) is critical to minimize vision loss; however, currently available therapies may not be fully effective or may have treatmentlimiting adverse effects. In general, topical therapy is ineffective for posterior uveitis.

Systemic toxic effects are a significant problem with most maintenance therapies. The fluocinolone acetonide (FA) intravitreous implant (Retisert; Bausch & Lomb, Rochester, New York) is a sustained-release system designed to deliver corticosteroid inside the eye for up to 30 months, thereby minimizing systemic toxic effects. The objective of this 3-year, multicenter,

clinical trial was to evaluate the safety and efficacy of the 0.59-mg and 2.1-mg FA intravitreous implants in subjects with unilateral or bilateral NIPU. The interim 34- week safety and efficacy results reported previously showed that the FA intravitreousn implant effectively controlled inflammation secondary to NIPU and that it successfully avoided the risk of systemic adverse events often associated with other therapies. 5 Herein we report the results of the safety and efficacy analyses of the same subjects for the full 3-year study period.

For editorial comment

see page 1287

METHODS

STUDY

This 3-year, multicenter, randomized, double-masked, historically controlled safety and efficacy study of an intravitreous implant containing FA (0.59 mg or 2.1 mg) in subjects with either bilateral or  nilateral NIPU was conducted at 27 clinical centers (26 in the United States and 1 in Singapore). The study was approved by the institutional review board at each center. All of the subjects provided written informed consent prior to participation, and an independent data and safety monitoring committee monitored the study results throughout the duration of the study. The main subject inclusion and exclusion criteria are shown in Table 1.

IMPLANTS

Both the 0.59-mg and redesigned 2.1-mg FA intravitreous implant contain a polymer-based, sustained-release formulation of FA within a 1.5-mm drug core encased in a silicone elastomer cup incorporating either a single (0.59 mg) or dual (2.1 mg) release orifice and a 98% hydrolyzed polyvinyl alcohol (PVA) membrane across the orifice that allows diffusion of the drug substance. The entire assembly is attached using silicone adhesive, which also forms an impermeable layer, to a heat-cured PVA suture tab with a suture hole to anchor the implant within the eye. In the original 2.1-mg implant, a PVA layer served both as a binder of the silicone cup assembly to the suture tab and as the pathway of diffusion from the pellet. Additionally, the silicone cup was attached to the PVA tab using silicone adhesive

at 2 points that were not part of the diffusion pathway. Once the cup, which did not include a diffusion orifice, was attached with PVA and silicone to the suture tab, the entire implant

was coated with multiple layers of PVA to further unitize the assembly. Drug diffusion occurred through the combination of the PVA adhesive layer, the PVA suture tab, and the PVA overcoat. Approximately 9 months after commencement of this study, a release rate higher than the upper limit specified for the 2.1-mg implant was reported to the study sponsor by the manufacturer of the clinical trial materials. Study enrollment was then suspended and the 2.1-mg implant was redesigned, following which enrollment into the study resumed. As indicated earlier, the redesigned 2.1-mg implant is similar in structure to the 0.59-mg implant, which was unchanged throughout the study. The implants were initially designed to release FA at a rate of approximately 2 μg/d for the 2.1-mg implant and 0.5 μg/d for

the 0.59-mg implant over 1000 days. The redesigned 2.1-mg implant was shown to initially release FA at approximately 2 μg/d, decreasing to approximately 1 μg/d over a 3-year period. Given the different release rates, data for uveitis recurrences are shown separately for the original and redesigned 2.1-mg implants.

SURGICAL PROCEDURE

ThesurgicalprocedureforimplantingtheFAintravitreousimplant was described previously.6 Briefly, the implant was sutured into place after the creation of a scleral opening at the pars plana. It was

anchored with an 8-0 Prolene suture so that the top surface of the implant and release orifice was facing the front of the eye.

STUDY DESIGN

The study design was described previously.5 All of the subjects received implants containing 1 of the 2 FA doses in their study eye. In subjects with bilateral disease, the more severely affected eye was assigned to 1 of the implant groups; these subjects were considered for enrollment only if the investigator felt that it would be possible to control fellow-eye ocular inflammation with local therapy (topical medication or periocular injection). Subjects were randomized in a ratio of 2:3 to receive the 0.59-mg or 2.1-mg FA implant (eFigure 1; http://www.archophthalmol.com). Following implantation, uveitis  edications were tapered. Systemic corticosteroid dosages were decreased by 30% per week to 2.5 mg/d for 1 week and then discontinued. Corticosteroid daily eyedrop frequency was decreased

by 1 eyedrop at weekly intervals and then discontinued. Immunosuppressive agents were tapered within a 6-week period and discontinued, as medically appropriate, at the investigator’s discretion. Follow-up visits occurred at 1 day after implantation (day 2), at weeks 1, 4, 8, 12, 18, 24, 30, 34, and 52, and every 3 months thereafter through year 3.

Assessments included best-corrected visual acuity (VA) as measured by the protocol developed for the Early Treatment Diabetic Retinopathy Study and adapted for the Age-Related Eye Disease

Study,7 applanation tonometry, slitlamp biomicroscopy, indirect ophthalmoscopy, automated visual field testing (Humphrey 24-2), hematology, and serum chemistry testing. Patients also underwent fluorescein angiography at screening, week 8, week 34, 1 year, 2 years, and 3 years with the evaluation of macular hyperfluorescence under a protocol developed by the Retinal Diseases and Image Analysis Center, Case Western Reserve University, Cleveland, Ohio, and described previously.

PRIMARY EFFICACY OUTCOME

The primary efficacy outcome was the difference in uveitis recurrence rate before and after implantation in the implanted eye (study eye). The analysis of recurrence was binary in structure, ie, answer of yes or no to the question of whether the subjects experienced a recurrence during that time. Postimplantation uveitis recurrences were defined as follows: (1) an increase of 2 or more steps in the number of anterior chamber cells compared with baseline; (2) an increase of 2 or more steps in vitreous haze compared with baseline; or (3) a deterioration of at least 0.30 log- MAR in VA from screening or baseline without an obvious alternate cause. Preimplantation recurrences were captured on a uveitis history clinical report form and were confirmed only if they met the same definition used for postimplantation recurrences. Postimplantation recurrences observed during the 12 months, 24 months, and 36 months following implantation surgery (referred to as the 1-, 2-, and 3-year postimplantation periods) were compared with  the preimplantation uveitis recurrences with onset within 1 year prior to implantation.

SECONDARY EFFICACY OUTCOMES

Secondary efficacy outcomes included within-subject comparisons of the implanted eye with the fellow nonimplanted eye for postimplantation uveitis recurrence rate; between–dose group comparisons of postimplantation uveitis recurrence rate in the implanted eye; the time to postimplantation uveitis recurrence in the implanted eye and fellow nonimplanted eye; the need for

adjunctive uveitis treatment (systemic therapy, periocular injections, or topical corticosteroids) for the implanted eye (before vs after implantation); and within-subject comparisons of the implanted

eye with the fellow eye for best-corrected VA and area of cystoid macular edema (CME) (masked reading of fluorescein angiography at 300 seconds by reading center).

SAFETY OUTCOMES

Safety outcomes included intraocular pressure (IOP), lens opacity classification system II lens opacity scores, visual fields, ocular and systemic adverse events, VA, and ophthalmoscopic examination

findings.

STATISTICAL ANALYSES

Initial sample size calculations were based on a hypothesis that 0.59-mg implants would have a higher recurrence rate than that observed with 2.1-mg implants (as suggested in pilot studies) and were performed to ensure a sufficient sample size to adequately define the safety profile. Primary and secondary efficacy end points were analyzed using the McNemar test for correlated proportions, with the following exceptions: the time to uveitis recurrence was determined using Kaplan-Meier analysis, and the between–dose group comparison of the postimplantation uveitis recurrence rate was performed using the Cochran-Mantel-Haenszel _2 test. The safety data (IOP, lenticular opacity, and visual fields) were analyzed using the _2 test stratified by investigator and prior therapy; the distribution of time

to first IOP elevation of 10 mm Hg or more was also assessed using proportional hazards regression stratified by investigative site and prior therapy. Adverse events, classified by the Medical Dictionary for Drug Regulatory Affairs, were compared using the Fisher exact test. Efficacy and safety analyses were performed on the intention-to-treat population, which included all of the enrolled subjects who received an implant and attended at least 1 postimplantation examination. Descriptive statistics were calculated for efficacy and safety data. Data for randomized patients were entered and verified in Clintrial database version 4.2 (Phase Forward, Waltham, Massachusetts) and were analyzed using SAS version 8.2 statistical software (SAS Institute Inc, Cary, North Carolina).

In the assessment of the uveitis recurrence rate, analyses were conducted using both observed recurrences and “imputed recurrences,” in which case any subject not seen within 10 weeks of his or her final scheduled visit was assigned a positive recurrence even though he or she may not have had one. Because the imputed recurrence rates were similar to the observed recurrence rates and using imputed data did not change the statistical significance of differences in recurrence rates in implanted study eyes or in fellow nonimplanted eyes, only observed data are shown.

RESULTS

Enrollment began in December 2000 and the study was completed in September 2005. The subjects’ demographic and baseline characteristics are shown in Table2.

A total of 278 subjects were randomized to receive either the 0.59-mg (n=110) or the 2.1-mg (n=168) FA intravitreous implant. Within the 2.1-mg FA implant group, 70 subjects received the original 2.1-mg implant and 98 received the redesigned 2.1-mg implant. Subjects were primarily female (72%) and had a mean age of 43.5 years. Most (77%) of the subjects had bilateral disease. No statistically significant difference was found between the 2 dose groups for any baseline characteristic. The most frequently used medications in the 1 year prior to  mplantation in both dose groups were glucocorticoids (91% of subjects) and anticholinergics (80% of subjects). A total of 241 subjects (87%) completed the study, including 98 (89%) in the 0.59-mg FA implant group and 143 (85%) in the 2.1-mg FA implant group (eFigure 1).

Seven subjects in each implant group discontinued participation owing to adverse events, including uncontrolled IOP (2 subjects), lysis of the anchoring suture (1 subject), endophthalmitis (1 subject), unassociated intraocular lymphoma (1 subject), spontaneous dissociation of the cup and strut (1 subject), and death (1 subject) in the 0.59-mg

FA implant group and uncontrolled IOP (3 subjects), necrotizing scleritis (1 subject), persistent hypotony  (1 subject), depletion of the study medication (1 subject), and cytomegalovirus infection with corneal decompensation (1 subject) in the 2.1-mg FA implant group. Three subjects in the 0.59-mg FA implant group and 7 subjects in the 2.1-mg FA implant group

were lost to follow-up. Two additional subjects in the 0.59-mg FA implant group and 11 in the 2.1-mg FA implant group discontinued participation for other reasons, including withdrawal of consent (3 subjects), receiving a second implant (6 subjects), removal of the implant (3 subjects), and enucleation of the implanted study eye because of persistent pain and

decreased vision associated with ciliary body shutdown (1 subject).

EFFICACY OUTCOMES

Uveitis Recurrence Rates in the Study Eye and Fellow Eye

Table 3 shows uveitis recurrence rates for implanted eyes and fellow nonimplanted eyes for the 1-year preimplantation period and the 1-, 2-, and 3-year postimplantation periods. These are cumulative recurrence rates following implantation (ie, the period from implantation to 12 months, from implantation to 24 months, and from implantation to 36 months, respectively). The FA implant greatly reduced uveitis recurrence in implanted eyes. The 1-, 2-, and 3-year postimplantation recurrence rates for the 0.59-mg group were 4%, 10%, and 20%, respectively, compared with the 1-year preimplantation rate of 62% (P_.01 for all). The 1-, 2-, and 3-year postimplantation recurrence rates for the 2.1-mg group (original and redesigned implants combined) were 7%, 17%, and 41%, respectively, compared with the 1-year preimplantation rate of 58% (P_.01 for all). Of note, uveitis recurrence rates with the original 2.1-mg FA implant during the 2- and 3-year postimplantation periods were higher than with the redesigned 2.1-mg FA implant, and the recurrence rate for the original 2.1-mg FA implant during the 3-year postimplantation period was not different compared with the 1-year preimplantation rate (P=.06). In contrast to the implanted eye, postimplantation uveitis recurrence rates in nonimplanted fellow eyes were increased compared with the 1-year preimplantation rate.

The recurrence rates in the 0.59-mg implant group were 44%, 52%, and 59% over the 1-, 2-, and 3-year postimplantation periods, respectively, compared with the 1-year preimplantation rate of 30% (P_.01 for all). For the 2.1-mg implant group (original and redesigned combined), the recurrence rates were 47%, 51%, and 55% over the 1-, 2-, and 3-year postimplantation periods, respectively, compared with the 1-year preimplantation rate of 22% (P_.01 for all).

As expected, further comparison of uveitis recurrence rates in implanted eyes vs those in nonimplanted fellow eyes showed that recurrence rates were significantly lower in implanted eyes compared with fellow nonimplanted eyes in both the 0.59-mg and 2.1-mg dose groups at the 1-, 2-, and 3-year follow-up periods (P_.01 for all).

Between–Dose Group Comparisons in Uveitis Recurrence

Comparison of the uveitis recurrence rate in the implanted eye over the 3-year postimplantation period by dose group indicated that therewerenosignificant differencesbetween

the0.59-mgand2.1-mg(originalandredesignedcombined) dose groups over the 1- and 2-year postimplantation periods (P=.32 andP=.06, respectively). However, there was a significant differencebetweendose groups over the 3-year postimplantation period, with a higher recurrence rate in the 2.1-mg group (P_.01).

Time to Recurrence of Uveitis

The time to uveitis recurrence was evaluated using Kaplan-Meier curves and is shown in Figure 1 (depicted as freedom from recurrence of uveitis) with mean (standard error) time to recurrence for each group given. There was a difference between implanted and fellow eyes in the freedom from uveitis curves for both the 0.59-mg implant group and the 2.1-mg implant group (original and redesigned combined) (Figure 1) (P_.01). In fellow nonimplanted eyes, uveitis recurrences were observed to a greater extent during the first 250 days after implantation in the contralateral study eye; for implanted eyes, uveitis recurrences were not seen until approximately 1000 days after implantation. There was a significant difference

in the freedom from uveitis curves for the 0.59-mg, original 2.1-mg, and redesigned 2.1-mg FA implant groups (Figure 1) (P_.01), clearly related to the faster time to recurrence in the original 2.1-mg implant group.

Uveitis Recurrences by Subject

As indicated before, the analysis of uveitis recurrence was binary (yes or no). However, all of the recurrences were documented. During the 1 year prior and 3 years subsequent

to implantation, the mean (SD) numbers of uveitis recurrences in the study eye per subject were 1.4 (1.6) and 0.5 (0.9), respectively, for the 0.59-mg dose group and 1.1

(1.3) and 0.7 (1.0), respectively, for the 2.1-mg dose group (original and redesigned implants combined). Nine subjects in the 0.59-mg group and 24 subjects in the 2.1-mg dose group had more than 1 recurrence in the study eye in the 3 years after implantation. In the fellow nonimplanted eye, the mean (SD) numbers of uveitis recurrences per subject in the 1 year prior to and 3 years after implantation were 0.57 (1.1) and 1.8 (2.1), respectively, for the 0.59-mg dose group and 0.40 (1.0) and 1.6 (1.8), respectively, for the 2.1-mg dose group.

Need for Adjunctive Uveitis Treatment

The proportion of subjects requiring adjunctive treatment to control inflammation before vs after implantation is shown in Table 4. The FA intravitreous implant significantly reduced the need for systemic therapy or periocular injections, with a nearly 80% reduction in the number of subjects requiring systemic medications to control uveitis regardless of dose or study visit. There was an approximate 95% reduction in periocular injections during the 1-year postimplantation period, decreasing during the 2- and 3-year postimplantation periods but still

much reduced from the 1-year preimplantation period (P_.01 for all). Finally, the proportion of FA-implanted eyes requiring topical corticosteroids was reduced by about 50% at the 1-year postimplantation visit (P_.01) but increased to preimplantation levels at the 2- and 3-year postimplantation visits. In comparison, the proportion of nonimplanted fellow eyes equiring periocular injections or topical corticosteroids increased during the 1-,2-, and 3-year postimplantation periods relative to the 1-year preimplantation period.

Visual Acuity

The mean (standard deviation) logMAR best-corrected VAs at baseline and at the 1-, 2-, and 3-year postimplantation visits for the implant groups and the nonimplanted fellow eyes are shown in Table 5. There was no significant difference in mean logMAR VA at the 1-or 3-year postimplantation visit as compared with baseline for either the 0.59-mg implant group or the 2.1-mg implant group. However, there was a significant improvement in logMAR VA at the 2-year postimplantation visit for both dose groups and a deterioration in mean log-

MAR VA in fellow nonimplanted eyes at all 3 postimplantation visits (P_.01). In the 0.59-mg and 2.1-mg implant groups at the 3-year visit, 23% (22 of 94) and 18% (26 of 141) of implanted

eyes, respectively, improved by at least 3 lines or more over baseline compared with 6% (5 of 90) and 4% (6 of137) of nonimplanted fellow eyes, respectively (P_.01 for both). In contrast, there was no significant difference between implanted eyes and fellow nonimplanted eyes in the proportion of eyes that deteriorated by 3 lines or more from baseline at the 3-year visit. In the 0.59-mg and 2.1-mg implant groups, VA decreased by 3 or more lines in 13% (12 of 94) and 21% (29 of 141) of implanted eyes, respectively, compared with 19% (17 of 90) and 18% (25 of 137) of nonimplanted fellow eyes, respectively (Figure 2).

Cystoid Macular Edema

The proportion of eyes with a reduction in the area of CME(a reduction characterized as any decrease from baseline) was greater in implanted eyes vs nonimplanted fellow eyes at the 1- and 3-year postimplantation visits (the proportion of eyes with CME reductions at the 2-year postimplantation visit was not calculated). At the 1- and 3-year postimplantation visits, there was a reduction in the area of CME in 86% (31 of 36) and 73% (29 of 40) of implanted eyes, respectively compared with 28% (10 of 36) and 28% (11 of 40) of fellow nonimplanted eyes, respectively, in the 0.59-mg implant group and 70% (39 of 56) and 45% (25 of 55) of implanted eyes, respectively, compared with 27% (15 of 56) and 22% (12 of 55) of fellow nonimplanted eyes, respectively, in the 2.1-mg FA implant group (P_.01). As reported previously, the mean area of CME in eyes receiving the 0.59-mg FA implant decreased from 33mm2 to 7 mm2 from screening to 34 weeks after implantation5 and remained statistically significantly lower than baseline at the 1-, 2-, and 3-year postimplantation visits (P_.01) (Table 6). Similarly, the mean area of CME in eyes that received the 2.1-mg implant (original and redesigned implants combined) decreased from 30mm2 to 4 mm2 from screening to 34 weeks after implantation5; however, while the mean area of CME at the 1- and 2-year postimplantation visits was significantly lower than at baseline, it no longer was by the 3-year visit. In contrast, the area of CME in nonimplanted fellow eyes fluctuated over a very narrow range (±7 mm2) during the 3-year follow-up period. Finally, the mean (SD) area of CME in eyes receiving the original 2.1-mg implant was 29 (43) mm2 at screening and 30 (48) mm2 at 3 years after implantation (P=.30), whereas the mean (SD) area of CME in eyes receiving the redesigned 2.1-mg implant

decreased from 30 (45)mm2 at screening to 19 (44) mm2 at 3 years after implantation (P=.14).

SAFETY OUTCOMES

Intraocular Pressure

At study entry, mean IOP values for both implanted study eyes and fellow eyes as well as for each dose group were within ±0.5mmHg and ranged from 14.2mmHg to 14.7 mm Hg. No statistically significant differences in mean IOP were observed between the 2 dose groups at any postimplantation study visit. Mean (SD) IOP values for the 0.59-mg and 2.1-mg FA implant dose groups at the 3-year postimplantation visit were 15.4 (8.1)mmHg and 13.9 (7.4) mm Hg, respectively, for implanted eyes and 14.2 (5.7) mm Hg and 14.9 (5.6) mm Hg, espectively, for fellow nonimplanted eyes. During the 3-year study period, 67% (74 of 110) of eyes receiving the 0.59-mg FA implant and 79% (133 of 168) of eyes receiving the 2.1-mg FA implant had IOP elevated by 10 mm Hg or more from baseline. In comparison, the incidence of an IOP increase of 10 mm Hg or more from baseline in fellow nonimplanted eyes was

23% (25 of 109) in the 0.59-mg FA implant group and 25% (42 of 166) in the 2.1-mg FA implant group. There was no significant difference in the incidence of an IOP increase of 10 mm Hg or more between the original and redesigned 2.1-mg FA implant groups. During the course of the 3-year postimplantation period, 78% of implanted eyes (both dose groups combined) required IOPlowering eyedrops compared with 36% of fellow eyes (P_.01), and 40% of implanted eyes (both dose groups combined) required IOP-lowering surgery compared with

2% of fellow eyes (P_.01). The frequency of IOPlowering surgery began to increase by postimplantation week 12 for implanted eyes and month 27 for fellow eyes. Six implants were removed during the course of the study due to elevated IOP. Details of the IOP elevation and its management have been published elsewhere.8

Lens Opacification

Lens opacification was graded at each visit for phakic eyes. The incidence of posterior subcapsular cataract progression (defined as an increase of_2 grades on the lens opacity classification system II scale) was 67% (88 of 131) in phakic implanted eyes compared with 18% (32 of 178) in phakic fellow nonimplanted eyes (P_.01). In the 0.59-mg and 2.1-mg (original and redesigned implants combined) implant groups, 70% (35 of 50) and 65% (53 of 81) of phakic implanted eyes, respectively, compared with 27% (18 of 66) and 13% (14 of 112) of phakic nonimplanted fellow eyes, respectively, experienced posterior subcapsular cataract progression. Nuclear and cortical cataract progression occurred in 21% (27 of 130) of implanted eyes in the 0.59-mg group and 12% (15 of 129) of implanted eyes in the 2.1-mg group (original and redesigned combined) vs 12% (22 of 178) of phakic nonimplanted fellow eyes in the 0.59-mg group and 7% (12 of 177) of phakic nonimplanted fellow eyes in the 2.1-mg group. During the 3-year postimplantation period, 93% (132 of 142) of phakic implanted eyes underwent cataract surgery compared with only 20% (37 of 186) of phakic nonimplanted fellow eyes (P_.01). Most of the surgical procedures performed on implanted eyes occurred between

postimplantation week 24 and month 24, whereas for fellow eyes, the overall frequency of cataract surgery increased progressively throughout the 3-year postimplantation period.

Visual Fields

Mean deviation was used as the quantitative measure for evaluating visual field changes. Statistically significant reductions in mean deviation (worsening visual fields) were

seen from baseline to the final 3-year visit both in eyes with the 0.59-mg FA implant (−1.42 dB) and their fellow  nonimplanted eyes (−1.05 dB) and in eyes with the 2.1-mg FA implant (−2.36 dB) and their fellow nonimplanted eyes (−0.76 dB) (P_.05 for all comparisons of baseline to 3-year visit). There was no significant difference in the change in mean deviation from baseline between the 0.59-mg and 2.1-mg (original and redesigned combined) implant groups (P=.28).

OTHER OCULAR ADVERSE EVENTS

Ocular adverse events were reported in 98% (273 of 278) of implanted study eyes and 86% (239 of 278) of fellow nonimplanted eyes. In addition to increased IOP and cataract

formation, the most frequently observed ocular adverse events reported for eyes in the 0.59-mg and 2.1-mg FA implant groups were eye pain (52% [57 of 110] and

60% [101 of 168], respectively), conjunctival hyperemia (31% [34 of 110] and 38% [64 of 168], respectively), conjunctival hemorrhage (29% [32 of 110] and 34% [57 of 168], respectively), blurred vision (30% [33 of 110] and 33% [55 of 168], respectively), and reduced VA (26% [29 of 110] and 35% [58 of 168], respectively). In fellow nonimplanted eyes, the most frequently

observed ocular adverse events in the 0.59-mg and 2.1-mg FA implant groups were eye pain (26% [29 of 110] and 26% [44 of 168], respectively), vitreous floaters (25% [27 of 110] and 23% [39 of 168], respectively), blurred vision (21% [23 of 110] and 20% [33 of 168], respectively), and reduced VA (15% [16 of 110] and 22% [37 of 168], respectively). The incidence of hypotony (IOP_6mmHg) was significantly higher in eyes in the 0.59-mg group (34% [37 of 110]) and 2.1-mg group (46% [78 of 168]) compared with fellow nonimplanted eyes (15% [41 of 275]; both dose groups combined) at any time during the 3-year follow- up (P_.01 for both). The prevalence of hypotony at 3 years after implantation is shown in eFigure 2. At

3 years after implantation, hypotony was present in 11% (11 of 96) and 15% (21 of 144) of eyes in the 0.59-mg and 2.1-mg implant groups, respectively, compared with 6% (14 of 235) of fellow nonimplanted eyes (both dose groups combined). This was significant only for the 2.1-mg group (P_.01).

Retinal detachments occurred in 4% (4 of 110) and 5% (9 of 168) of eyes in the 0.59-mg and 2.1-mg FA implant groups, respectively, compared with 3% (9 of 278) of the fellow nonimplanted eyes (both dose groups combined). Endophthalmitis that was attributed to incomplete wound closure following implantation occurred in 1 subject (1%) in the 0.59-mg FA implant group and in no subjects in the 2.1-mg FA implant group. There were no cases of endophthalmitis in fellow nonimplanted eyes.

NONOCULAR ADVERSE EVENTS

Nonocular adverse events were reported in 94% (261 of 278) of subjects. The most frequently observed nonocular adverse events were headache (37% [102 of 278]), sinusitis (19% [52 of 278]), nasopharyngitis (16% [45 of 278]), nausea (16% [45 of 278]), and arthralgia (15% [42 of 278]). One subject in the 0.59-mg FA implant group died from preexisting cancer. For most of these subjects, the events were reported by the investigator to be unrelated (64% [168 of 261]) or unlikely to be related (22% [58 of 261]) to the study treatment. No serious systemic adverse events related to the implant were reported.

EXPLANTS

Overall, a total of 22 subjects had the FA implant surgically removed over the course of the 3-year postimplantation follow-up period (8 subjects in the 0.59-mg FA implant group and 14 in the 2.1-mg FA implant group) for the following reasons: 6 for uncontrolled elevated IOP (of which 1 was at the request of the subject), 6 for suspected depletion of study medication (5 of these subjects were reimplanted with new FA implants), 3 for spontaneous dissociation of the implant from its anchoring strut (these implants were produced prior to the adoption

of an improved adhesive process), and 1 each for spontaneous expulsion of the implant (this subject was reimplanted with a new FA implant), inadvertent lysis of the anchoring suture, endophthalmitis, necrotizing scleritis (in the area of the implant in a subject with sarcoidosis), unassociated intraocular lymphoma, cytomegalovirus infection,9 and hypotony. Five  implants were removed in each of the first 2 years, with a further 12 implants removed during the third year of the study.

COMMENT

The results from this study demonstrate that the FA intravitreous implant is highly effective in controlling inflammation secondary to NIPU. Results of the primary efficacy outcome analysis showed that uveitis recurrence rates in implanted eyes during the 1-, 2-, and 3-year postimplantation periods (ie, during the 12-, 24-, and 36- month periods following implantation, respectively) were significantly lower than the 1-year preimplantation rate in the same eyes regardless of dose group. As expected, the rate of uveitis recurrence was low for the period corresponding to the active life of the implant but started to increase near the end of the drug delivery period. In the 0.59-mg FA implant and redesigned 2.1-mg FA implant groups, uveitis recurrence rates were extremely low during the 1- and 2-year postimplantation periods but started to increase during the 3-year postimplantation period even though the rates were still significantly lower than the 1-year preimplantation rates. In contrast, in eyes receiving the original 2.1-mg FA intravitreous implant, which was found to have a faster drug delivery rate than intended, the recurrence rates were significantly lower during the 1- and 2-year postimplantation periods but returned to preimplantation levels during the third year. Results of the secondary efficacy outcomes were consistent with the primary efficacy outcome. Uveitis recurrence rates in implanted eyes were lower than those in fellow nonimplanted eyes even though approximately onefourth of the patients had not been diagnosed with uveitis in their fellow nonimplanted eye at the time of enrollment, and in those subjects with bilateral NIPU the more severely affected eye received the FA implant. However, the protocol-required withdrawal of systemic medications following implantation may have contributed in part to the difference in uveitis recurrence rates between implanted eyes and nonimplanted fellow eyes (ie, by contributing to an increase in uveitis recurrences in the fellow nonimplanted eyes of subjects with bilateral uveitis). There was no difference in the postimplantation uveitis recurrence rates between the 0.59-mg and 2.1-mg FA implant dose groups during the 1- or 2-year

postimplantation period, but there was a significant difference between the 2 dose groups during the 3-year postimplantation period, again likely due to faster drug delivery with the original 2.1-mg implant and the contribution of that implant to the uveitis recurrence rate observed for the combined 2.1-mg dose group (original plus redesigned). In agreement with a prior study in which the mean time to uveitis recurrence was delayed by about 12 months in eyes with the FA intravitreous implant compared with fellow nonimplanted eyes,10 Kaplan-Meier freedom from uveitis curves were significantly different for implanted eyes vs fellow nonimplanted eyes regardless of dose group. In addition, differences between the 0.59-

mg, original 2.1-mg, and redesigned 2.1-mg implant groups were also significant. Finally, in agreement with previous studies,6,10 adjunctive therapy to manage uveitis was significantly reduced after implantation, with a near 80% reduction in the need for systemic medications to control inflammation. These results exceed those reported in a study of eyes receiving intravitreous triamcinolone acetonide (IVTA) injections for uveitis in which only 54.5% of subjects were able to reduce or discontinue use of systemic anti-inflammatory agents.11

Visual acuity remained stable or improved over baseline in most eyes during the 3-year study period. A significantly greater proportion of implanted study eyes compared with fellow nonimplanted eyes in either dose group had an improvement in VA of at least 3 lines over baseline at the 3-year postimplantation visit, and mean logMAR VA for implanted eyes was either improved or unchanged from baseline at the 1-, 2-, and 3-year postimplantation visits. These differences should be interpreted with some caution given that more implanted eyes

than nonimplanted fellow eyes had a VA of 0.3 logMAR or worse at baseline (72% vs 39%, respectively) due to the protocol requirement of the more severely affected eye being the implant study eye; hence, implanted eyes may have had more room for improvement. On the other hand, implanted eyes had a greater incidence of cataract formation, which would be expected to lower VA at some point in these eyes. Other studies have reported similar VA improvements with the FA implant,6,10,12 and there was a concurrent significant deterioration in the mean logMAR VA in fellow nonimplanted eyes at the 1-, 2-, and 3-year postimplantation visits. In addition, the proportion of implanted eyes with a reduction in the area ofCMEwas consistently greater than that of fellow nonimplanted eyes regardless of implant dose, and reduction in the area of CMEwas correlated with improvement in VA. Thus, post hoc analysis revealed a significant correlation between CME and VA with correlation coefficients of 0.32, 0.21, and 0.21 at baseline, 34 weeks after implantation, and 3 years after implantation, respectively (P_.05 for all). Cataract formation was nearly universal in implanted phakic eyes in this study, consistent with other

FA implant studies in which the incidence of cataract formation ranged from 50% to 100%.10,13 In contrast, Kok et al11 reported a 14.3% incidence of cataract formation in eyes receiving IVTA injections with a mean follow- up of 8 months. Given the high likelihood of cataract formation in phakic eyes receiving an FA intravitreous implant, lens extraction should be  onsidered in eyes with mild cataract at the time of implantation. Consistent with previous studies,6,10 IOP elevation was a common complication of the FA intravitreous implant. In eyes with uveitis, elevated IOP may arise as a consequence of the intraocular inflammation itself leading to the occlusion of the aqueous outflow (the presence of inflammatory debris or the formation of peripheral anterior synechia)14,15 or from reduced aqueous outflow associated with corticosteroid use.16-18 In our study, three-fourths of implanted eyes had an IOP elevation

of 10 mm Hg or more above baseline at some time during the 3-year postimplantation period, and 40% of eyes required IOP-lowering surgery (primarily trabeculectomy or placement of a glaucoma drainage device). Success rates were high when surgery was required, with 85% of eyes achieving a postoperative IOP of 6 to 21 mm Hg (see the article by Goldstein et al8 for a detailed analysis of the management of elevated IOP following implantation with the FA intravitreous implant, which includes patients from this study and 2 other studies with the FA

intravitreous implant). While elevated IOP and glaucoma surgery have also been reported following treatment of uveitic patients receiving IVTA injections,19-21 the incidence of elevated IOP with the FA intravitreous implant was higher than that with IVTA injections, which ranged from 36.3% to 48.6% depending on the definition used for elevated IOP.6,11,12,22-25 This may be related to differences between FA and TA at the glucocorticoid receptor and/or differences in exposure between the 2 treatments. Eyes implanted with the FA intravitreous implant are exposed to sustained corticosteroid levels throughout the implant’s 30-month lifespan, whereas in the referenced studies, eyes treated with IVTA injections are exposed to fluctuating corticosteroid levels every 3 months. Of note, the FA implant doses studied in this clinical trial were chosen based on the results of previous studies with FA intravitreous implants ranging in doses from 0.59 mg to 6 mg (conducted under Investigator Investigational New Drug applications) suggesting that implants with release rates of 6.0 μg/d might be more likely to cause excessive elevations of IOP. For this reason, the 0.59-mg FA implant with a release rate of 0.5 μg/d and the 2.1-mg FA implant with an intended release rate of 2.0 μg/d were chosen for this study. Although it is possible that a lower FA dose may have a better safety profile compared with the 0.59-mg and 2.1-mg FA implants, we do not know whether a lower dose would have sufficient efficacy. There were substantially more ocular adverse events in implanted eyes than in fellow nonimplanted eyes. Adverse events in fellow nonimplanted eyes were characteristic of the underlying uveitic process (eye pain, vitreous floaters, and reduced vision), whereas implanted eyes manifested adverse events consistent with implantation (increased IOP, cataract formation, conjunctival hyperemia, endophthalmitis, hypotony, and hemorrhage). Of note, there were no nonocular serious adverse events considered to be treatment related in either implant group. The absence of systemic steroid-related adverse effects is consistent with minimal exposure of FA to the systemic circulation and the short half-life; indeed, plasma levels of FA have been reported to be below the limit of detection following FA intravitreous implantation.

The prevalence of hypotony remained relatively constant for individual dosage groups during the first year following FA implant insertion. These rates gradually increased during the second and third years. The reason for hypotony in these eyes is likely multifactorial and is dependent on the time following initial implant placement. Very early on, eyes may have been transiently hypotonous because of surgical trauma and decreased aqueous production that subsequently improved. Over time, particularly beginning at about 12 months following the original implant surgery, an increasing number of eyes underwent filtration surgery, which is associated with increased incidence of hypotony in eyes that have undergone FA implant

insertion.7 Hypotony in later years of the study could also be related to uveitis recurrences, which occurred more commonly as implants were depleted of drug; in these eyes, decreased IOP would have resulted from inflammationinduced decreased aqueous production. Further follow- up will be needed to determine whether hypotony resolves in eyes with drug-depleted implants when the inflammation is once again controlled. In summary, our results demonstrate that the FA intravitreous implant effectively controls ocular inflammation and improves VA in eyes with NIPU not only over the 34-week postimplantation period previously described5 but for the full 3-year study duration, and they are in agreement with other studies on the FA intravitreous implant.6,10 Approximately 80% of subjects in the 0.59-mg group experienced a 3-year recurrence-free period with a significant reduction in the need for adjunctive therapy, whereas fellow nonimplanted eyes experienced a high rate of early recurrence and a significant increase in the need for adjunctive therapy. The high incidence of elevated IOP in implanted eyes is a significant complication with the FA intravitreous implant but can be managed in most patients with IOP-lowering eyedrops or surgery as documented by Goldstein et al.8 However, strategies directed at limiting the IOP elevation observed with the FA intravitreous implant warrant further investigation. Finally, the practitioner will need to balance the control of ocular inflammation observed with the FA intravitreous implant against the potential need for IOP-lowering surgery and cataract extraction along with any complications associated with these surgical procedures and the FA implantation procedure itself (ie, hypotony,retinal detachment, and endophthalmitis).

Submitted for Publication: January 20, 2008; final revision received May 17, 2008; accepted May 19, 2008.

Correspondence: David G. Callanan, MD, Texas Retina Associates, 1001 N Waldrop Dr, Ste 512, Arlington, TX 76012 (dcallanan@texasretina.com).

Author Contributions: All of the authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.