outcome of trabeculectomy with intraoperative mitomycin c for uveitic glaucoma

Outcome of trabeculectomy with intraoperative mitomycin C for uveitic glaucomaJason Noble,* MD; Larissa Derzko-Dzulynsky,†‡ MD; Theodore Rabinovitch,‡ MD;

Catherine Birt,†‡ MD

ABSTRACT • RÉSUMÉ

Background: The purpose of this study was to compare the outcomes of mitomycin C (MMC)-augmentedtrabeculectomy in glaucoma patients with uveitis to those without uveitis but with other high-risk characteristics.

Methods: A retrospective comparative cohort analysis consisting of 51 eyes of 51 patients (21 uveitic patientsand 30 nonuveitic patients) was performed. Two outcome classifications were analyzed: absolute success(intraocular pressure [IOP] ≤ 30% baseline without glaucoma medications or 5-fluorouracil (5-FU)injections), and qualified success (IOP ≤ 30% baseline with glaucoma medications or 5-FU injections).Kaplan–Meier survival curves were constructed for both models.

Results: After a mean follow-up of 52 months, uveitis emerged as a negative predictor of success. In thequalified success model, uveitic patients demonstrated survival rates of 90% at 1 year and 79% at 2 yearscompared with 100% for all time points in the control group (Wilcoxon test,p = 0.005).Uveitic patients weremore likely to require postoperative 5-FU injections than the control group (33% vs. 10%, p = 0.04) and weremore likely to require glaucoma medications postoperatively for IOP control (38% vs. 3%, p = 0.001).

Interpretation: Uveitic glaucoma patients are more likely to require postoperative therapeuticinterventions to maintain adequate pressure control in the short-term and are at higher riskof surgical failure in the long-term.

Contexte : Cette étude a pour objet de comparer le résultat de la trabéculectomie avec mitomycine C(MMC) chez des patients glaucomateux atteints et non atteints d’uvéite mais ayant d’autres caractéristiquesde risque élevé.

Méthodes : Une analyse comparative et rétrospective a été effectuée chez une cohorte de 51 yeux de 51patients (21 avec uvéite et 30 sans uvéite). Deux catégories de résultats ont été analysés : réussite absolue(pression intraoculaire [PIO] avec base de référence ≤30 % sans médicaments contre le glaucome ouinjections de fluorouracil 5 (FU-5)) ou réussite qualifiée (PIO avec base de référence ≤30 % sans médicamentou injections de FU-5). On a élaboré des courbes de survie Kaplan-Meier pour les deux modèles.

Résultats : Après un suivi de 52 mois en moyenne, l’uvéite s’est avérée être un prédicteur négatif de réussite.Selon le modèle de réussite qualifiée, les patients atteints d’uvéite ont présenté des taux de survie de 90 %après 1 an et de 79 % après 2 ans comparativement à 100 % en tout temps chez le groupe témoin (testWilcoxon, p = 0,005). Les patients atteints d’uvéite était plus sujets à avoir besoin d’injectionspostopératoires de FU-5 que le groupe témoin (33 % c. 10 %, p = 0,04) et de médicaments contre laglaucome après l’opération pour maîtriser la PIO (38 % c. 3 %, p = 0,001).

Interprétation : Les patients atteints de glaucome uvéitique sont plus sujets à avoir besoin d’in-terventions thérapeutiques postopératoires pour maîtriser adéquatement la pression à courtterme et sont à risques plus élevé d’échec chirurgical à long terme.

From *the Faculty of Medicine, University of Toronto, †the Departmentof Ophthalmology, Sunnybrook Health Sciences Centre, Toronto, Ont.,and ‡the Department of Ophthalmology and Vision Sciences, Universityof Toronto, Toronto, Ont.

Originally received Feb. 22, 2006. Revised Nov. 17, 2006Accepted for publication Nov. 28, 2006

Correspondence to: Catherine Birt, MD, Sunnybrook Health SciencesCentre, Suite M1 302A, 2075 Bayview Ave., Toronto ON M4N 3M5;[email protected]

This article has been peer-reviewed. Cet article a été évalué par les pairs.

Can J Ophthalmol 2007;42:89–94doi:10.3129/can j ophthalmol.06-124

MMC trabeculectomy in uveitic glaucoma—Noble et al 89

Glaucoma associated with chronic uveitis is wellknown to carry a high risk of surgical failure.1–3

Although trabeculectomy augmented with mitomycinC (MMC) is the suggested surgical approach for refrac-

tory glaucoma in high-risk eyes,4 the success of intraop-erative MMC for glaucoma specifically associated withuveitis has not been adequately documented. In thepresent study, uveitic glaucoma patients were compared

with high-risk glaucoma patients without ocular inflam-mation over a relatively long follow-up period followingMMC trabeculectomy.

METHODS

Approval from the research ethics board ofSunnybrook Health Sciences Centre was obtainedbefore data collection began. A review of patients whohad undergone an MMC trabeculectomy atSunnybrook for uveitis-related glaucoma betweenDecember 1994 and February 2002 was performed. Alluveitic patients were diagnosed with uveitis at least 6weeks preoperatively and were seen by uveitis specialists(L.D.D., T.R.). The control group consisted of all glau-coma patients within the same time period with nohistory of ocular inflammation who were treated withMMC-augmented trabeculectomy for other high-riskcharacteristics (such as prior failed filter, pseudophakia,or African descent) and who were followed for at least 9months. Patients who underwent a simultaneouscataract extraction with the trabeculectomy procedurewere excluded from the study. If the clinical record indi-cated that a patient underwent surgery to both eyes, thefirst eye operated on was used in the analysis.

All trabeculectomy operations were performed by onesurgeon (C.B.). After peribulbar anaesthetic administra-tion, a superior limbal peritomy was performed andhemostasis of the scleral bed was achieved using bipolarcautery. A piece of Weck-Cel sponge was soaked in 0.2 mg/mL of MMC and placed under the conjunctivafor a period of time varying between 30 and 180 seconds.After removal of the sponge, the area was copiously irri-gated with balanced salt solution. A diamond blade set at0.25 mm was used to incise a 4 mm × 4 mm rectangularscleral flap. The scleral flap was lifted with a crescentknife, the anterior chamber was entered with a paracen-tesis blade, and the trabeculectomy ostomy was createdunder the flap with the blade and scissors, followed by aperipheral iridectomy. The scleral flap was laid back andsutured with four interrupted 10-0 nylon sutures.Balanced salt solution was irrigated through the paracen-tesis to show filtration. The conjunctiva was replaced tothe limbus and held with a combination of running andinterrupted 8-0 Vicryl sutures. Balanced salt solution wasagain irrigated through the paracentesis to demonstratethe formation of a watertight bleb, after whichtobramycin and dexamethasone eye ointment wereinstilled. After surgery, patients were placed on a stan-dard regimen of prednisolone acetate every hour whileawake, tobramycin and atropine drops 4 times a dayeach, and dexamethasone ointment at night. Drops were

tapered and stopped depending on the clinical examina-tion at each visit and with the goal of maintainingcontrol of the uveitis and function of the trabeculectomy.

Patient charts were reviewed for pertinent informa-tion and the data were entered into a computer data-base. The intraocular pressure (IOP) measured at thetime of surgical booking was considered the baselineIOP for each respective patient. Patients were followedup and IOP levels were recorded at postoperative day 1,and 3 months, 6 months, 1 year, and 2 years postsurgi-cally, when possible.

In the uveitic group, 4 patients were on a preoperativecourse of immunosuppressive agents, includingmethotrexate and cyclosporin A. Also, 8 patients in theuveitic group were taking oral steroids preoperatively,and 5 were on oral nonsteroidal antiinflammatory drugs(NSAIDs).

Statistical analysis consisted of 2 full Cox model analy-ses comparing 5 predictor variables between the uveiticand control groups. The predictor variables included age,ancestry, history of previous ocular surgery, aphakic orpseudophakic state, and the number of glaucoma med-ications preoperatively. χ2 tests were done comparing theuse of postoperative subconjunctival injections of 5-fluo-rouracil (5-FU) and the number of glaucoma medica-tions between the 2 groups. Outcome was classifiedacross 2 distinct models of success: (1) absolute success(IOP reduction of 30% of baseline or more withoutglaucoma medications or postoperative subconjunctival5-FU injections), and (2) qualified success (IOP reduc-tion of 30% of baseline or more with glaucoma medica-tions or postoperative subconjunctival 5-FU injections).When patients were found to have failed in a particularmodel, the time at which this failure event took place wasrecorded. Kaplan–Meier survival analysis was done usingthe two different outcome classifications to determinethe survival probability at various time points.

RESULTS

In total, 51 eyes of 51 patients were reviewed. Thepatient cohort consisted of 21 uveitic glaucoma patientsand 30 nonuveitic glaucoma patients of mean (SD) age 52(16) years (range 18–89). Twenty-six patients were menand 25 were women. The patients had a mean (SD)follow-up time of 52 (25) months (range 9–114 months).All uveitic patients had adequate inflammatory controlpreoperatively (<5 cells per high power field in the anteriorchamber). Uveitis diagnoses are listed in Table 1. All nonu-veitic patients had at least 1 high-risk characteristic requir-ing the use of intraoperative MMC. These risk factorsincluded 20 (67%) cases of prior intraocular surgery

MMC trabeculectomy in uveitic glaucoma—Noble et al

90 CAN J OPHTHALMOL—VOL. 42, NO. 1, 2007

(including cataract extraction, trabeculectomy, or vitrec-tomy), 7 patients (23%) were of African ancestry, and 3patients (10%) were of younger age (i.e., age ≤ 18 years).

IOP decreased significantly in both the uveitic andthe control group after trabeculectomy surgery (Table2). In the uveitic group, IOP fell from 33.9 (SD 11.1)mm Hg preoperatively to 15.2 (6.4) mm Hg (a 55%decrease) after an average follow-up of 53.1 months(range 9–91). In the control group IOP fell from 30.3(8.4) mm Hg preoperatively to 14.2 (7.4) mm Hg (a53% decrease) after an average follow-up of 51.5months (range 12–114). There was no statistically sig-nificant difference in the preoperative IOP levelsbetween the two groups (p > 0.05, t test).

Table 3 shows the data on the 5 predictor variables forthe uveitic and control groups. No significant differencein age, ancestry, previous history of ocular surgery,aphakic or pseudophakic state, or the number of glau-coma medications preoperatively was detected betweenthe uveitic and the control group. Also, MMC exposuretimes for the uveitic (71.4 [0.3] seconds) and controlgroup (75.2 [0.3] seconds) were not significantly differ-ent. There was a significant difference in the need forpostoperative 5-FU injections between the 2 groups,with 33% of uveitic patients requiring 5-FU injections

compared with 10% of controls (p = 0.04). Further,38% (8/21) of patients in the uveitic group requiredglaucoma medications postoperatively for IOP controlcompared with 3% (1/30) in the control group (p =0.001). At 1-year follow-up, however, there was no sta-tistical difference in the number of patients taking glau-coma medications between the uveitic and controlgroups (38% vs. 17%, respectively, p = 0.08).

Complications occurred in 4 (19%) of the uveiticpatients: 2 cases of hypotony, 1 case of choroidal hem-orrhage, and 1 case of presumed (culture negative)endophthalmitis. Three (10%) of the patients in thecontrol group had complications: 1 case of hypotony, 1wound leak, and 1 case of late blebitis with associatedvitritis that occurred 2 years after the surgery. There wasno difference in the overall rate of complicationsbetween the 2 groups.

Uveitic patients were found to require significantlymore postoperative procedures compared with thecontrol group (p = 0.007). Eleven (52%) of the uveiticpatients underwent a postoperative procedure: 4 repeattrabeculectomy operations, 2 cataract extractions, 2wound revisions, 2 5-FU injections, and 1 requiredYAG (yttrium aluminum garnet) laser to the internalostomy. In comparison, 5 (17%) of the control patients


CAN J OPHTHALMOL—VOL. 42, NO. 1, 2007 91

Table 1—Uveitic diagnoses and IOP measurements for the uveitic group

Patient Age,years Uveitic diagnosis

IOP at baseline

IOP at last follow-up

1 37 Psoriatic 30 122 18 Tubulointerstitial nephritis and uveitis syndrome 35 11 3 35 Reiter's disease 37 144 28 Vogt-Koyanagi-Harada syndrome 41 10 5 45 Idiopathic 26 14

32 04 cihtapoidI 43 67 66 Vogt-Koyanagi-Harada syndrome 24 5

6 04 cihtapoidI 35 89 49 Fuch's heterochromic iridocyclitis 31 25

41 62 sisodiocras demuserP 14 0111 42 HLAB27 negative spondyloarthropathy 17 23

41 83 cihtapoidI 83 2113 47 Idiopathic 23 8



81 05 suriv retsoz sepreH 56 8119 72 Herpes zoster virus 36 2020 43 Fuch's heterochromic iridocyclitis 64 24 21 48 HLAB27 positive 22 12

Note: IOP = intraocular pressure (mm Hg).

Table 2—IOP measurements before and after trabeculectomy with mitomycin C

Group Preoperative 1 year 2 years last follow-up

Uveitis 33.9 (11.1) 16.1 (6.7) 14.6 (5.3) 15.2 (6.4)

Control 30.3 (8.4) 11.7 (5.0) 13.4 (3.8) 14.2 (7.4)

Note: values are presented as mean (SD) intraocular pressure (IOP), mm Hg.

underwent postoperative procedures: 2 wound revisions,1 IOL repositioning, 1 YAG to the internal ostomy, and1 cataract extraction.

Cox analysisIn both outcome models, there was no evidence that

any of the 5 predictor variables had a significant influ-ence on outcome. On the other hand, uveitis was foundto be a negative predictor of outcome in each model.Although not statistically significant, uveitic patientswere found to be 1.8 times more likely to fail in theabsolute success model (i.e., require glaucoma medica-tions or 5-FU injections for IOP control) than thecontrol group (p = 0.22).

Kaplan–Meier analysisKaplan–Meier survival curves were constructed for

the 2 models at an average follow-up time of 52 months.Figs. 1 and 2 show the survival curves for each outcomeclassification across the study time points.

In the absolute success model, the uveitic group demon-strated worse survival across all time points (Fig. 1), with

6-month, 1-year and 2-year survival rates of 67%, 51%,and 51%, respectively, compared with 90%, 80%, and70%, respectively, in the control group. Nonetheless, thesurvival curves for the 2 groups were not found to be sig-nificantly different (Wilcoxon test, p = 0.06).

A statistically significant difference in survival wasdetectable in the qualified success model (Fig. 2), withthe control group demonstrating a 100% survival prob-ability rate at all time points compared with a survivalrate of 90% and 79% at 1 and 2 years, respectively, inthe uveitic group (Wilcoxon test, p = 0.005).

INTERPRETATION

Currently, the initial treatment for uveitic glaucoma ismedical therapy. If this fails, filtration procedures such atrabeculectomy are indicated. The outcome of unaug-mented trabeculectomy surgery in uveitic glaucomapatients has been previously investigated. Hill et al5

reported an 81% qualified success rate at 1 year. Stavrouet al7 reported a qualified success rate of 78% and anabsolute success rate of 53% at 5 years. Thus, it appearsthat unaugmented trabeculectomy surgeries are initiallyeffective after surgery but have a substantial risk offailure over the longer term.

In recent years, surgeons have documented positiveresults using antimetabolite therapy in difficult-to-treatcases such as uveitic glaucoma. In particular, the cytosta-tic antimetabolite 5-FU has been shown to produce goodresults. Studies looking at 5-FU–enhanced trabeculec-tomy in uveitic patients have demonstrated qualifiedsuccess rates of 82%, 71%, and 67% at 24-, 36-, and 60-month follow-up, respectively. These results suggest 5-FU



Fig. 1—Kaplan–Meier survival curves for uveitic and control groups in the absolute success model.

Table 3—Predictor variables by study group*

PredictorUveitis (n = 21)

Nonuveitis(n = 30)

Age, mean (SD), years 47.9 (14) 54.7 (17) African descent 1 (5) 7 (23) Previous ocular surgery 8 (38) 7 (23) Pseudophakic/aphakic state 6 (29) 5 (17) Preoperative glaucoma medications, mean (SD) 3.8 (0.7) 3.4 (1.0)

Note: values are number (percentage) unless otherwise stated. *No statistical difference between groups was detectable (p > 0.05).

improves outcomes when compared with unaugmentedsuccess rates.6,8 Shunt procedures have also been shown toimprove success in uveitic glaucoma. Da Mata et al9

reported a 95% 1-year qualified success rate for Ahmedvalve implants, and Hill et al5 reported 1- and 2-yearqualified success rates of 79% using Molteno implants.Molteno et al10 demonstrated a 5- to 10-year long-term(success rate of 76% with Molteno implants.

More recently, the antiproliferative agent MMC hasbeen suggested as an adjunct for use in trabeculectomy inhigh-risk cases such as aphakic or pseudophakic eyes,eyes with a previous history of ocular surgery, and eyeswith uveitis.1 Indeed, studies have suggested that MMCmay be superior to 5-FU as an antimetabolite therapy.MMC avoids the corneal risk of 5-FU, may reduce theneed for repeated subconjunctival injections, and alsoimproves outcomes in other high-risk eyes.11–14 Forexample, both Skuta et al11 and Katz et al12 demonstratedthat IOP levels and rate of dependence on glaucomamedications postoperatively were both lower in patientstreated with MMC than in those treated with 5-FU.

There is currently a dearth of studies investigating theoutcome of MMC trabeculectomy in uveitic glaucomapatients. In one study, Prata et al15 looked at the short-term 10-month results of trabeculectomy in glaucomaassociated with uveitis and reported an absolute successrate of 75% and a qualified success rate of 91.7%. Thesestatistics were not derived from Kaplan–Meier analysis.Wright et al16 performed a similar review for a slightlylonger follow-up period of 14 months and found lowerabsolute and qualified success rates of 62% and 75%,respectively. Recently, Ceballos et al17 have reported a

qualified success rate of 78% at 1 year and 62% at 2 years.Unfortunately, the above studies did not compare thesuccess rates of the uveitic patients with a control groupwithout uveitis. Furthermore, all the aforementionedstudies defined success as an IOP lower than 21 mm Hg.

In the present study, patients with uveitic glaucomademonstrated worse surgical outcomes after MMC tra-beculectomy compared with high-risk glaucomapatients without ocular inflammation. Given that allhigh-risk surgical features (i.e., the predictor variables)were not significantly different between the 2 groups,the worse outcome in the uveitis group is likely second-ary to the history of intraocular inflammation.

Although not statistically significant, the uveitic groupwas found to have worse absolute success rates than thecontrol group, with 1-year survival rates of 67% com-pared with 90%, respectively. Furthermore, uveiticpatients were more likely to require postoperative 5-FUinjections and glaucoma medications to maintain ade-quate IOP control early in the postoperative period. Thissuggests that a significant proportion of patients withuveitic glaucoma require postoperative therapeutic inter-ventions to control IOP after MMC trabeculectomy.

Patients with uveitic glaucoma were also found tohave significantly worse qualified success rates whencompared with controls. The 1- and 2-year survival ratesfor the uveitic group were 90% and 79%, respectively,while the control group maintained a 100% qualifiedsuccess rate throughout the entire study period. Thisimplies that even with postoperative therapeutic inter-ventions a significant proportion of uveitic patients areunable to maintain adequate pressure control after


CAN J OPHTHALMOL—VOL. 42, NO. 1, 2007 93

Fig. 2—Kaplan–Meier survival curves for uveitic and control groups in the qualified success model.

MMC trabeculectomy. Furthermore, the data suggestthat qualified success rates for uveitic patients progres-sively decline over time. This may indicate that the riskof complete surgical failure necessitating a repeat tra-beculectomy becomes progressively greater in the longerterm in uveitic glaucoma patients.

Unique attributes of this study were the use of a high-risk control group for comparison and the use of a morerelevant definition of success. Traditionally, success hasbeen somewhat arbitrarily defined as achieving an IOPless than 21 mm Hg. As target IOP levels are idiosyn-cratic and related to baseline measurements,18 it may bemore relevant to measure success in terms of a reductionfrom baseline IOP in outcome analyses.

Patients in the study exhibited relatively high rates ofpostoperative complications, including a relatively highincidence of wound revisions. This is likely due to thehigh-risk characteristics of the patients enrolled in thestudy. Two of 51 patients were diagnosed with possibleendophthalmitis during the postoperative period. Oneof the patients was diagnosed with blebitis 2 years post-operatively, however, and the 2nd case, although treatedas a presumed endophthalmitis, presented as ahypopyon 2 months postoperatively. Since cultures ofthis 2nd case were negative, this may have representedsterile endophthalmitis or a uveitis flare.

This study demonstrates that uveitic glaucomapatients treated with MMC trabeculectomy have worseoutcomes when compared with high-risk patientswithout a history of intraocular inflammation. Thesepatients are at increased risk of requiring postoperativetherapeutic interventions to maintain adequate pressurecontrol in the short term, and remain at high-risk of sur-gical failure in the longer term.

The authors gratefully thank Marko Katic for help with thestatistical analysis.

REFERENCES

1. Reynolds AC, Skuta GL. Clinical perspectives on glaucoma-filtering surgery: antiproliferative agents. Ophthalmol ClinNorth Am 2000;13:501–16.

2. Kok H, Barton K. Uveitic glaucoma. Ophthalmol Clin NorthAm 2002;15:375–87.

3. Panek WC, Holland GN, Lee DA, et al. Glaucoma in patientswith uveitis. Br J Opthalmol 1990;74:223–7.

4. Yanoff M, Duker JS. Ophthalmology. 1st ed. London, UK:Mosby International Ltd; 1999.

5. Hill RA, Nguyen QH, Baerveldt G, Forster DJ, Minckler DS,Rao N, Lee M, Heuer DK. Trabeculectomy and Moltenoimplantation for glaucomas associated with uveitis.Ophthalmology 1993;100:903–8.

6. Towler HM, McCluskey P, Shaer B, Lightman S. Long-termfollow-up of trabeculectomy with intraoperative 5-fluo-rouracil for uveitis-related glaucoma. Ophthalmology2000;107:1822–8.

7. Stavrou P, Murray PI. Long-term follow-up of trabeculectomywithout antimetabolites in patients with uveitis. Am JOphthalmol 1999;128:434–9.

8. Patitsas CJ, Rockwood EJ, Meisler DM, Lowder CY.Glaucoma filtering surgery with postoperative 5-fluorouracilin patients with intraocular inflammatory disease.Ophthalmology 1992;99:594–9.

9. Da Mata A, Burk SE, Netland PA, Baltatzis S, Christen W,Foster CS. Management of uveitic glaucoma with Ahmedglaucoma valve implantation. Ophthalmology 1999;106:2168–72.

10. Molteno AC, Sayawat N, Herbison P. Otago glaucoma surgeryoutcome study: long-term results of uveitis with secondaryglaucoma drained by Molteno implants. Ophthalmology2001;108:605–13.

11. Skuta GL, Beeson CC, Higginbotham EJ, et al: Intraoperativemitomycin versus postoperative 5-fluorouracil in high-riskglaucoma filtering surgery. Ophthalmology 1992;99:438–44.

12. Katz GJ, Higginbotham EJ, Lichter PR, et al: Mitomycin Cversus 5-fluorouracil in high-risk glaucoma filtering surgery:extended follow-up. Ophthalmology 1995;102:1263–9.

13. Kitazawa Y, Kawase K, Matsushita H, et al: Trabeculectomywith mitomycin: a comparative study with fluorouracil. ArchOphthalmol 1991;109:1693–8.

14. Lamping KA, Belkin JK: 5-Fluorouracil and mitomycin C inpseudophakic patients. Ophthalmology 1995;102:70–5.

15. Prata JA Jr, Neves RA, Minckler DS, Mermoud A, Heuer DK.Trabeculectomy with mitomycin C in glaucoma associatedwith uveitis. Ophthalmic Surg 1994;25:616–20.

16. Wright MM, McGehee RF, Pederson JE. Intraoperative mit-omycin-C for glaucoma associated with ocular inflammation.Ophthalmic Surg Lasers 1997;28:370–6.

17. Ceballos EM, Beck AD, Lynn MJ. Trabeculectomy withantiproliferative agents in uveitic glaucoma. J Glaucoma 2002;11:189–96.

18. The Advanced Glaucoma Intervention Study (AGIS): 7. Therelationship between control of intraocular pressure and visualfield deterioration. The AGIS Investigators. Am J Ophthalmol2000;130:429–40.

Key words: glaucoma, uveitis, outcome, trabeculectomy, mito-mycin C



outcome of trabeculectomy with intraoperative mitomycin c for uveitic glaucoma

Documents