Kaplan-Meier survival analysis of trabeculectomies for 29 eyes of 20 patients with juvenile idiopathic arthritis–related uveitic glaucoma undergoing their first mitomycin C–augmented trabeculectomy. IOP indicates intraocular pressure; TNF, tumor necrosis factor.
eFigure. Trabeculectomy survival (IOP?=?21 mm Hg) in JIA-uveitis with and without prior surgery
Leinonen S, Kotaniemi K, Kivelä T, Majander A. Potential Effect of Tumor Necrosis Factor Inhibitors on Trabeculectomy With Mitomycin C for Patients With Juvenile Idiopathic Arthritis–Related Uveitic GlaucomaA Retrospective Analysis. JAMA Ophthalmol. 2015;133(11):1323-1328. doi:10.1001/jamaophthalmol.2015.3387
The majority of patients with juvenile idiopathic arthritis–related uveitic glaucoma require surgery to control intraocular pressure. Trabeculectomy with mitomycin C (MMC) is a major treatment option, although both chronic inflammation and young age increase risk of filtration failure. Factors that potentially protect from filtration failure are important to identify.
To evaluate the potential effect of treatment with tumor necrosis factor (TNF) inhibitor on the success of an MMC-augmented trabeculectomy for patients with juvenile idiopathic arthritis–related uveitic glaucoma.
Design, Setting, and Participants
In a retrospective observational study with a median follow-up of 7.9 years at the Department of Ophthalmology, Helsinki University Hospital, in Helsinki, Finland, 29 eyes of 29 consecutive patients (3.1-20.4 years of age) underwent an MMC-augmented primary trabeculectomy during the period from April 1996 to January 2014. Fifteen patients were treated with systemic TNF inhibitors at the time of their trabeculectomy to control their uveitis, arthritis, or both. No changes were made to the antirheumatic therapy preoperatively.
Main Outcomes and Measures
Successful trabeculectomies, determined by Kaplan-Meier analysis, in which patients have intraocular pressure of 21 mm Hg or lower without antiglaucomatous medications or further glaucoma procedures.
The success rate of trabeculectomy for patients who were treated with TNF inhibitors was 73% (95% CI, 44%-89%) at 1, 5, and 10 years after surgery, whereas the success rates of trabeculectomy for patients who were not treated with TNF inhibitors were 57%, 16%, and 0% at 1, 5, and 10 years after surgery, respectively (P = .01). The trabeculectomies of patients who were treated with TNF inhibitors were successful for a median of 3.2 years (95% CI, 0.3-9.9 years), whereas the trabeculectomies of patients who were not treated with TNF inhibitors were successful for a median of 1.2 years (95% CI, 0.6-3.6 years) (P = .14). Nine eyes of 9 patients had previously undergone cyclodestruction, intraocular surgery, or both (ie, prior ocular surgery). The effect of TNF was observed especially in the eyes of patients who had not undergone prior ocular surgery (at 5 years: 83% success rate for patients who had not undergone prior ocular surgery vs 19% success rate for patients who had). A difference in the overall success rate between patients who had and patients who had not had prior surgery was not identified.
Conclusions and Relevance
Our retrospective data suggest that better control of intraocular pressure was achieved in the eyes of patients with juvenile idiopathic arthritis–related uveitis glaucoma who were treated with TNF inhibitors at the time of their MMC-augmented primary trabeculectomy. The retrospective design of the present study suggests that our data can be used for the planning of future studies but not for making treatment decisions.
Secondary glaucoma is not only a common complication of pediatric uveitis (with a frequency rate of 14%-42%) but also an important risk factor for severe vision loss.1- 4 Changes in the conventional outflow pathway in the chamber angle that result from chronic uveal inflammation and corticosteroid use can lead to an increase in intraocular pressure (IOP).5 In up to two-thirds of children, the IOP elevates within the first 2 years of uveitis, and the risk of secondary glaucoma is further increased after 7 years of active uveitis.6,7 The diagnosis of glaucoma for children is based on elevated IOP and optic disc cupping, the latter of which may develop quickly together with visual field loss.3,8 Antiglaucomatous medication is often insufficient, and 59% to 63% of eyes eventually undergo an average of 2.6 surgical procedures to control the IOP.1,2 The surgical approaches include a trabeculectomy with or without an antimetabolite such as mitomycin C (MMC), drainage implants, a deep sclerectomy, and, in some medical centers, a goniotomy.1,9- 13
Tumor necrosis factor (TNF) is one of the proinflammatory and profibrogenic cytokines expressed not only by multiple inflammatory cells but also by ocular tissues.14 The TNF levels are elevated in the aqueous humor of patients with idiopathic uveitis, and TNF inhibitors have proved to be an efficient treatment for uveitis.14,15 In the Helsinki University Hospital in Helsinki, Finland, they have been part of the treatment of juvenile idiopathic arthritis (JIA) and JIA-related uveitis since 1995. During the period from 2004 to 2014, nearly 60% of patients with JIA-related uveitis were treated with TNF inhibitors if the uveitis or arthritis was not controlled by nonbiologic antirheumatic therapy. Long-term safety reports on treating children are still needed.16,17 For example, potential increases in overall and cancer-related mortality after TNF inhibitor therapy have been suspected.18
The success of a trabeculectomy depends on multiple factors modifying the wound healing process, and chronic inflammation and young age are risk factors for trabeculectomy failure.19,20 Tumor necrosis factor enhances the proliferation of fibroblasts of the human Tenon capsule, and elevated levels of TNF in the aqueous humor before surgery are associated with early trabeculectomy failure.21,22 The administration of a topical TNF inhibitor after a trabeculectomy suppressed the postoperative subconjunctival healing response in a recent experimental animal model.23 To the best of our knowledge, clinical studies on the effect of TNF inhibition on trabeculectomies have not been published. Our aim was to evaluate whether concurrent use of TNF inhibitors had any effect on the success rate of MMC-augmented primary trabeculectomies for patients with JIA-related uveitic glaucoma.
Secondary glaucoma is a major risk for vision loss for patients with juvenile idiopathic arthritis–related uveitis. Medication is often insufficient to control the intraocular pressure.
Trabeculectomy with mitomycin C is one treatment option. Both chronic inflammation and young age are risk factors for filtration failure, however.
Treatment with tumor necrosis factor inhibitors was associated with higher success rates of trabeculectomy with mitomycin C for patients with juvenile idiopathic arthritis–related uveitis.
All consecutive patients with JIA-related uveitic glaucoma diagnosed at 16 years of age or younger who were undergoing an MMC-augmented primary trabeculectomy in 1 or both eyes at the Department of Ophthalmology, Helsinki University Hospital during the period from April 1996 to January 2014 were eligible for enrollment in our study. Thirty patients with JIA-related uveitic glaucoma (who underwent a total of 55 MMC-augmented primary trabeculectomies) were identified in our database. Uveitis was classified and graded according to the recommendations of the Standardization of Uveitis Nomenclature (SUN) Working Group.24 One patient was excluded from our study because he or she received a biologic treatment (abatacept) other than a TNF inhibitor at the time of the trabeculectomy. Sixteen patients were male, and 13 female. One eye was drawn at random for analysis from 14 patients who had underwent a sequential bilateral MMC-augmented trabeculectomy. Indication for surgery was IOP higher than 21 mm Hg under maximal tolerated medication and concomitant glaucomatous optic disc changes. All patients were receiving antirheumatic medications to achieve remission of JIA and a SUN grade of 0.5+ or lower for uveitis. This medication was not changed preoperatively. The ethics committee of the Department of Surgery, Hospital District of Helsinki and Uusimaa, approved this study, and its design complies with the Declaration of Helsinki. A retrospective study model requires an ethical evaluation of the study, and the data are deidentified for analysis, but no patient consent form is needed.
The following data were recorded: onset of uveitis, onset of JIA diagnosed by a pediatric rheumatologist, the age of the patient and the antirheumatic medication used at the time of trabeculectomy, previous ocular surgery, preoperative and postoperative IOP and activity of uveitis, perioperative and postoperative complications, IOP increase of more than 21 mm Hg postoperatively, best-corrected visual acuity at the last follow-up, and later IOP-lowering surgery. The main outcome measure was trabeculectomy survival defined as IOP of 21 mm Hg or lower without antiglaucomatous medication or later IOP-lowering surgery. Intraocular pressure was measured either with the Goldmann applanation tonometer or the iCare TA01i rebound tonometer (Icare Finland Oy), depending on the age and cooperation of the patient. Best-corrected visual acuity was recorded in decimal notation with a Snellen eye chart or a letter chart for older patients and with a Lea symbol chart (Good-Lite) for younger patients. Counting fingers was converted to decimal notation.25 Visual fields were recorded for all patients capable of performing the test; these results were not included in our study.
Of the 29 eyes of 29 patients, 26 were operated on by 2 surgeons, and 3 were operated on by 2 other surgeons. The conjunctival incision was performed superiorly. An instrument wipe 3 × 3 mm in size saturated with mitomycin C (0.4 mg/mL) was applied on episcleral and corresponding sub-Tenon surfaces, followed by rinsing with a 30-mL balanced salt solution. A partially thick, limbus-based scleral flap was formed. A sclerectomy was performed using either a 1-mm motorized trephine (Geuder AG) or a 0.6-mm Crozafon–De Laage punch (Moria SA), followed by a peripheral iridectomy. The scleral flap was closed with two to four 10-0 polyamide sutures, and the conjunctiva with a continuous 9-0 polyglactin suture. All patients were seen on the first postoperative day, every 1 to 7 days until the filtration stabilized, every 4 to 10 weeks after the immediate postoperative period, or more frequently if needed. Postoperative medication included a topical antibiotic and cyclopentolate hydrochloride for 2 weeks and topical corticosteroids up to 8 times a day, tapering off slowly during the 3 months after surgery.
The Statistical Package for the Social Sciences (SPSS, version 19; IBM Corp) and Stata, version 13 (StataCorp), were used. Uveitis activity during follow-up was dichotomized (SUN grade of <1+ or ≥1+). Distributions of continuous variables were compared using the Mann-Whitney U independent-samples test, and distributions of categorical variables were compared using singly ordered contingency tables. All tests were 2 sided, and P < .05 was considered to be statistically significant. Kaplan-Meier survival analysis with a log-rank test was used for the analysis of trabeculectomy survival. Stratification was used to guard against potential bias introduced by imbalances in the type of JIA (oligoarticular or polyarticular), history of prior surgery (yes or no), surgeon, age (tertiles: <9, 9-12, and ≥13 years), and duration of uveitis (tertiles: <4, 4-7.9, and ≥8 years) at the time of trabeculectomy in the association of TNF inhibitor treatment with trabeculectomy survival.26 Subgroup analysis based on previous surgical procedures was planned a priori; no other subgroup analyses were performed. Comparisons between subgroups were adjusted according to the Bonferroni test.
At the time of MMC-augmented primary trabeculectomy, 14 patients did not receive TNF inhibitors for their uveitis or JIA, and 15 patients were treated with TNF inhibitors: 12 received 3 to 6 mg/kg of infliximab every 4 to 8 weeks, 1 received 24 mg/m2 of adalimumab every 2 weeks, 1 received 0.8 mg/kg of etanercept weekly, and 1 received 50 mg of golimumab every 4 weeks. All patients were treated with topical corticosteroids, and 26 patients were treated with nonbiologic disease-modifying antirheumatic drugs (Table 1). Inhibition of TNF continued uninterrupted for a median of 4.1 years (range, 0.8-11.7 years) after trabeculectomy for the patients who received TNF inhibitors. Because of relapses of JIA, uveitis, or both, 5 patients started receiving TNF inhibitors a median 2.8 years(range, 1.5-9.4 years) after the trabeculectomy. They were treated for a median of 3.3 years (range, 0.8-8.0 years). Three patients started receiving TNF inhibitors after trabeculectomy failure, and 2 before its eventual failure.
Trabeculectomy with MMC was the first surgical intervention for 20 eyes, whereas 9 eyes had previously undergone cyclodestruction, intraocular surgery, or both (7 eyes had undergone cyclophotocoagulation, 4 cataract surgery, 2 viscocanalostomy, and 1 posterior vitrectomy). Three eyes were pseudophakic, and 1 eye was aphakic, at the time of trabeculectomy.
One eye had a perioperative complication, vitreous prolapse to the iridectomy. One choroidal detachment was surgically corrected, otherwise no surgical intervention was necessary postoperatively (Table 2). Postoperative hypotony (IOP ≤ 5 mm Hg), observed in 17 eyes, resolved within a median of 14 days (range, 1-48 days). It did not differ in duration between patients who received and patients who did not receive TNF inhibitors (P = .42, determined by use of the Mann-Whitney U test). The median postoperative best-corrected visual acuity was 0.8 (range, 0.005-1.1) (Snellen equivalent, 20/25 [range, 20/18 to 20/4000]) for eyes in which the hypotonic phase lasted 2 weeks or longer and 1.0 (range, 0.2-1.0) (Snellen equivalent, 20/20 [range, 20/20 to 20/100]) for eyes with a shorter hypotonic phase (P = .47, determined by use of the Mann-Whitney U test). None of our patients developed bleb leakage or a bleb-related infection.
The median follow-up time after the trabeculectomy was 7.9 years (range, 1.0-18.7 years). At the end of follow-up, 13 MMC-augmented trabeculectomies survived, while 16 had failed, and 18 additional IOP-lowering procedures had been performed in the latter eyes. The only factor identified as being associated with a successful trabeculectomy was TNF inhibition (P = .003 for the full follow-up, determined by use of the Fisher exact test) (Table 3).
The overall median survival of the trabeculectomy was 1.5 years (95% CI, 0.9-5.0 years [range, 0.04-11.7 years]). The trabeculectomies performed on patients who received TNF inhibitors at the time of surgery were successful for a longer length of time (median, 3.2 years [95% CI, 0.3-9.9 years]; range, 0.04-11.7 years) than those performed on patients who did not receive TNF (median, 1.2 years [95% CI, 0.6-3.6 years]; range, 0.04-9.3 years), a median difference of 2 years (P = .14, determined by use of the Mann-Whitney U test).
By Kaplan-Meier analysis, the success rates of trabeculectomies were 73% (95% CI, 44%-89%) at 1, 5, and 10 years for patients who received TNF inhibitors and 57% (95% CI, 28%-78%) at 1 year, 16% (95% CI, 3%-40%) at 5 years, and 0% at 10 years for patients who did not receive TNF inhibitors (P = .01, determined by use of the log-rank test) (Figure). The effect of treatment with TNF inhibitors persisted when stratifying the analysis by type of JIA (adjusted P = .01 for TNF inhibition), history of prior surgery (P = .03), surgeon (P = .03), age of the patient (P = .03), and duration of uveitis (P = .02) at the time of trabeculectomy.
A difference in the overall survival of trabeculectomies in eyes with vs without prior surgery was not identified (P = .21) (eFigure, A, in the Supplement). The success rates of trabeculectomies were 83% (95% CI, 48%-96%) at 1 and 5 years for patients without prior ocular surgery and 38% (95% CI, 9%-67%) at 1 year and 19% (95% CI, 1%-54%) at 5 years for patients who did not receive TNF inhibitors (P = .01, determined by use of the log-rank test with Bonferroni correction for 2 comparisons) (eFigure, B, in the Supplement).
Differences in the median corresponding survival rates of trabeculectomies performed on patients who had prior ocular surgery with or without TNF inhibitors were not identified (17% with vs 33% without TNF inhibitors at 5 years; P > .99) (eFigure, C, in the Supplement). No differences were observed between patients who had prior ocular surgery and patients who did not with regard to the following baseline characteristics: TNF inhibitor treatment (P = .25, determined by use of the Fisher exact test), nonbiologic disease-modifying antirheumatic treatment (P = .22), age at onset of uveitis (P = .24, determined by use of the Mann-Whitney U test), age at onset of JIA (P = .45), duration of uveitis (P = .12), preoperative IOP (P = .23), preoperative SUN grade (P = .87), SUN grade of 1+ or higher for uveitis activity during follow-up (P = .54), presence of postoperative hypotony (P = .63, determined by use of the Fisher exact test), or the duration of postoperative hypotony (P = .37, determined by use of the Mann-Whitney U test). The patients who had prior ocular surgery were marginally older at the time of trabeculectomy than the patients who did not (median age, 13.7 years vs 10.0 years; P = .07).
Treatment with inflammation-modifying TNF inhibitors at the time of trabeculectomy was associated with a higher success rate of MMC-augmented trabeculectomy in our small retrospective series of patients with JIA-related uveitic glaucoma who typically show a vigorous wound healing response because of their uveitis and young age.10- 12 This finding suggests that TNF inhibitors may influence wound healing after trabeculectomy.21 Tumor necrosis factor inhibitors were given to patients with more active JIA or JIA-related uveitis (ie, when control was not achieved with nonbiologic antirheumatic treatment). In our series, the shortest duration of TNF inhibitor treatment following an MMC-augmented trabeculectomy was 8 months. Our small sample size did not allow us to analyze the minimum duration of TNF inhibitor treatment associated with a successful trabeculectomy. Five patients who did not receive TNF inhibitors eventually received them after relapses of JIA, uveitis, or both that occurred after the trabeculectomy, which had failed or subsequently failed for all of them.
Our study is retrospective and uncontrolled in design and, thus, subject to bias. No potential effect of TNF inhibitors could be predicted at the time of trabeculectomy. A randomized, controlled TNF inhibitor trial cannot easily be designed for patients with JIA-related uveitis because of the relapsing nature of both JIA and uveitis. The potential adverse effects of TNF inhibitors also limit their experimental use.18
Treatment with TNF inhibitors was associated with a high trabeculectomy success rate for patients who did not have prior ocular surgery but not for patients who had prior ocular surgery. Prior ocular surgery has frequently been associated with a less successful trabeculectomy survival.27- 29 Given a clinically meaningful 35% point difference in survival rates at 5 years but a nonsignificant log-rank test, the effect of prior ocular surgery on the outcome of the trabeculectomy remains open in our small retrospective series.
Postoperative hypotony accompanied with hypotonic maculopathy was common in our patients, but it was unassociated with TNF inhibition or the lack of it. Younger age and secondary glaucoma, as such, are known risk factors for hypotonic maculopathy and for late low IOP after MMC-augmented trabeculectomies.30 Prolonged hypotony is reported to be a rare complication in children; however, it is likely the result of a vigorous wound healing response.19,31 In our patients, postoperative hypotony and hypotonic maculopathy were not associated with any adverse effects or lower final visual acuity. In line with safety reports on MMC-augmented trabeculectomy,32 no bleb-related infections or other trabeculectomy-related complications caused permanent visual loss in our study group.
Our results suggest that MMC-augmented trabeculectomy is a viable surgical approach when treating uveitic glaucoma in patients with JIA who receive TNF inhibitors. Our retrospective design cannot address cause and effect, but previous studies21- 23 support the interpretation that systemic TNF inhibitors used at the time of the MMC-augmented trabeculectomy may have protected patients against scarring of the trabeculectomy bleb.
Submitted for Publication: April 23, 2015; final revision received July 29, 2015; accepted July 30, 2015.
Corresponding Author: Sanna Leinonen, MD, Department of Ophthalmology, Helsinki University Hospital, Haartmaninkatu 4 C, PL 220, FI-00029 HUS, Helsinki, Finland (firstname.lastname@example.org).
Published Online: September 24, 2015. doi:10.1001/jamaophthalmol.2015.3387.
Author Contributions: Drs Leinonen and Kivelä 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.
Study concept and design: Leinonen, Kotaniemi, Majander.
Acquisition, analysis, or interpretation of data: Leinonen, Kivelä, Majander.
Drafting of the manuscript: Leinonen, Majander.
Critical revision of the manuscript for important intellectual content: Kotaniemi, Kivelä, Majander.
Statistical analysis: Leinonen, Kivelä.
Administrative, technical, or material support: Leinonen, Majander.
Study supervision: Kotaniemi, Kivelä, Majander.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.