Trabeculectomy Function after Extraction

Philip P. Chen, MD,1 Yaffa K. Weaver, MD,2 Donald L. Budenz, MD,2 William J. Feuer, MS,2 Richard K. Parrish II, MD2

Objective: To examine the effect of cataract extraction (CE) after on intraocular pressure (IOP) control. Design: Retrospective noncomparative case series. Participants: A total of 115 consecutive patients who underwent extracapsular CE (N 5 58) or phacoemul- sification (N 5 57) with intraocular (IOL) placement after trabeculectomy were studied. Intervention: Cataract extraction with IOL after trabeculectomy was performed. Main Outcome Measures: Preoperative, intraoperative, and postoperative factors were evaluated for association with loss of IOP control requiring additional medications, bleb needling, or further surgery, using Kaplan–Meier survival analysis and Cox multivariate proportional hazards survival regression. Results: After mean postoperative follow-up of 21.1 6 14.3 months, additional glaucoma medication or needling of the filtering bleb to maintain IOP control was required in 35 (30.4%) and was significantly associated with intraoperative manipulation and early postoperative peak IOP greater than 25 mmHg. Additional was eventually required in 11 eyes (9.6%) and was significantly associated with age of 50 years or younger, preoperative IOP greater than 10 mmHg, and early postoperative peak IOP greater than 25 mmHg. The cumulative proportion of patients who did not require reoperation for glaucoma was 93% and 90% at 1 and 2 years, respectively. The mean IOP at last visit had increased 1.6 mmHg above the pre-CE level and did not vary significantly after the first postoperative month. The median interval from CE to the addition of glaucoma medication or bleb needling was 1.6 months (within 3 months in 20 of 33 eyes) and that from nonsurgical intervention to further glaucoma surgery was 3.6 months (before the 7th postoperative month in 6 of 11 eyes). Of 19 eyes with hypotony (IOP ∂ 6 mmHg) before CE, 11 eyes remained hypotonous after CE despite an increase in the mean IOP from 4.6 to 7.5 mmHg. Conclusions: When CE is performed after trabeculectomy, age of 50 years or younger, preoperative IOP greater than 10 mmHg, intraoperative iris manipulation, and early postoperative IOP greater than 25 mmHg are associated with worsened postoperative IOP control. Most bleb failures occur soon after CE. Resolution of pre-existing hypotony after CE is unpredictable. 1998;105:1928–1935

Glaucoma-filtering surgery may accelerate cataract forma- to maintain IOP control after extracapsular cataract extrac- tion in eyes with postoperative hypotony, shallow anterior tion (ECCE) with (IOL) placement.5–9 Min- chamber, or excessive inflammation.1–4 Cataract extraction imizing intraoperative tissue trauma and postoperative in- (CE) after glaucoma-filtering surgery may decrease bleb flammation may play a role in maintaining bleb function size and function and result in higher intraocular pressure after CE.10,11 Uncomplicated phacoemulsification (PE) has (IOP); previous studies have reported 10% to 38% of eyes been shown to result in significantly lower aqueous flare and require additional medication or further glaucoma surgery cells after surgery than ECCE12,13 and fewer filtering bleb failures.14,15 Other authors have noted resolution of post- trabeculectomy hypotony after CE.16,17 We retrospectively Originally received: November 24, 1997. examined preoperative, intraoperative, and postoperative Revision accepted: April 10, 1998. Manuscript no. 97810. factors to identify risk factors associated with loss of IOP 1 Department of Ophthalmology, University of Washington, Seattle, Wash- control after ECCE or PE with IOL placement. We also ington. investigated the resolution of post-trabeculectomy hypotony 2 Department of Ophthalmology, University of Miami School of Medicine, after CE. Bascom Palmer Institute, Miami, Florida. Supported in part by an unrestricted grant from Research to Prevent Blindness, Inc, New York, New York; and Public Health Service Research Grant EY10410, Department of Health and Human Services, National Methods Institutes of Health, National Eye Institute, Bethesda, Maryland. None of the authors have any proprietary interest in any of the products Before medical records were reviewed, approval for this study was mentioned in this article. given by the Medical Sciences Subcommittee for the Protection of Reprint requests to Philip P. Chen, MD, Department of Ophthalmology, Human Subjects in Research at the University of Miami School of University of Washington, Box 356485, Seattle, WA 98195. Medicine. A computerized search using surgical current proce-

1928 Chen et al z Cataract Extraction after Trabeculectomy

Table 1. Outcomes of Trabeculectomy Function after that when additional medication was added (28.6 6 4.6 vs. 27.1 6 Cataract Extraction 8.3 mmHg). This indicated that blebs that underwent needling were not failing to a greater degree than blebs that had medications Outcome Definition added. Eyes that underwent further glaucoma surgery were con- Complete No additional medications,* bleb needling, or further sidered failures. success glaucoma surgery Based on the above criteria, each variable was evaluated for Qualified Additional medication* or bleb needling needed for association with trabeculectomy survival (i.e., time to qualified success IOP control success or failure) using Kaplan–Meier survival analysis. Al- Failure Further glaucoma surgery needed for IOP control though survival analysis predicates inclusion of all patients regard- less of length of follow-up, to maintain clinical relevance we IOP 5 intraocular pressure. excluded patients with less than 6 months of follow-up. Cox * Additional medications used more than 2 months after cataract extrac- multivariate proportional hazards survival regression using for- tion. ward-stepwise variable selection was used to determine which risk factors were independent predictors of qualified success or failure. All preoperative and intraoperative variables were candidates for entry into the model (not only those identified as significant by dural terminology codes was used to identify patients who had univariate analysis). Differences in pre- and post-CE IOP were undergone CE after trabeculectomy between January 1991 and examined for statistical significance using paired and two-sample, December 1995 at the Anne Bates Leach Eye Hospital, Miami, two-tailed Student’s t test. Results are given as the mean 6 Florida. Variables investigated included age at time of CE, gender, standard deviation where applicable. race, eye, type of glaucoma, best-corrected Snellen , number and type of glaucoma medications, previous incisional ocular surgeries, time elapsed since the most recent trabeculec- tomy to the cataract surgery, type of antifibrosis agent used with Results the most recent trabeculectomy, bleb appearance (pre- and post- CE), IOP (the average IOP during the 4 months preceding CE, the The search by CPT codes identified 125 eyes of 125 consecutive maximum IOP during the first 2 postoperative weeks, and the IOP patients who had CE after trabeculectomy. Two patients were at 1, 2, 3, 6, 9, 12, 18, 24, and 36 months and at the last visit after excluded from analysis because their glaucoma (due to traumatic surgery), method of CE (ECCE vs. PE), position of cataract hyphema) had resolved despite a flat bleb, and the cataract incision incision relative to the filtering bleb, cataract incision location was made at the trabeculectomy site without attempt to maintain (clear corneal vs. limbal vs. scleral), length and closure (number bleb function. Eight patients were excluded because of follow-up and type of sutures), IOL type and placement, intraoperative iris of less than 6 months, resulting in a study population of 115 eyes manipulation, intraoperative and postoperative complications, of 115 patients. postoperative medications and antifibrosis agents administered, The mean age at the time of CE was 68.9 6 12.0 years (range, and the dates when new glaucoma medicines were added or further 17–88 years) and differed significantly between patients having glaucoma surgery was performed. ECCE (71.5 6 8.6 years, N 5 58) and PE (66.2 6 14.2 years, N 5 Patients undergoing planned ECCE had a 7.0- to 11.5-mm 57; P 5 0.02). Demographic data and characteristics of the pa- incision made either in peripheral or through the surgical tients in this study are listed in Table 2. No patients had bleb limbus, either adjacent to or in a separate quadrant from the revision in conjunction with CE, but one patient had choroidal filtering bleb, followed by can-opener anterior , nu- effusions drained at the time of ECCE and one patient had pars clear mobilization and expression, manual or automated irrigation plana performed on the same date as PE for retained and aspiration of cortical remnants, placement of a polymethyl- nuclear fragments. methacrylate (PMMA) IOL, and wound closure with 10–0 nylon The mean follow-up was 21.1 6 14.3 months (range, 6–60 sutures. Patients undergoing PE had a tunnel incision (length, months) and was significantly longer for ECCE (24.5 6 16.3 3.2–7.0 mm) made in peripheral cornea or , followed by months) compared to PE (17.6 6 10.9 months, P , 0.001). Nine continuous-tear , nuclear and emul- patients were included in the survival analysis only until the last sification, automated irrigation and aspiration of cortical remnants, office visit before the date when nonglaucoma-related intraocular placement of a silicone foldable or PMMA IOL, and wound surgery was performed, which included repair of traumatic cataract closure with 10–0 nylon sutures in some cases. Iris manipulation wound dehiscence on postoperative day 16 (1), pars plana vitrec- (posterior synechiolysis, stretching, sphincterotomies, sector iri- tomy (for vitreous hemorrhage [1, at 3.2 months], bleb-related dectomy, or use of iris retraction hooks [Grieshaber, Kennesaw, [1, at 13 months], and epiretinal membrane [1, at Georgia]) was performed as necessary to enable CE. Intraocular 31.3 months]), bleb revision for hypotony (3, all within 3.5 lenses were implanted in the posterior capsular bag or the ciliary months), and penetrating keratoplasty for pseudophakic bullous sulcus; in one patient, an anterior chamber IOL was used. keratopathy (2, at 6 and 6.5 months). We defined complete success as the absence of additional Thirty-five eyes (30.4%) required additional medication or glaucoma medications, bleb needling, or further glaucoma surgery filtering bleb needling (i.e., were qualified successes) to maintain for long-term IOP control after CE (Table 1); however, additional IOP control after surgery. Eleven eyes (9.6%) required additional glaucoma medicines used within the first 2 postoperative months glaucoma surgery (i.e., were failures), including trabeculectomy were permitted. Eyes that required additional medications for IOP (6), glaucoma drainage device placement (5), and pars plana control after the second postoperative month or that underwent vitrectomy for aqueous misdirection (1). All but one patient in the bleb needling at the were considered qualified successes. failure group was a qualified success before reoperation. We included bleb needling in the same category as additional The cumulative proportion of patients who did not need further medications because some preferred to needle a previ- glaucoma surgery (i.e., were not failures) was 93% at 1 year (N 5 ously functioning bleb rather than start new medications (1 sur- 72), 90% at 2 years (N 5 45), and 85% at 3 years (N 5 20) (Fig geon performed 8 of 9 needlings), and the mean IOP at the time 1). The cumulative proportion of patients who also did not require bleb needling was performed was not significantly different from additional glaucoma medicines or bleb needling (i.e., were com-

1929 Ophthalmology Volume 105, Number 10, October 1998

Table 2. Demographic Data and Characteristics of Eyes Having Cataract Extraction (CE) after Trabeculectomy

N% Gender Male 57 (50) Female 58 (50) Race White 56 (49) Hispanic 42 (37) Black 17 (15) Asian 1 (1) Glaucoma type Primary open-angle 57 (50) Chronic angle closure 15 (13) Pseudoexfoliation 11 (10) Normal tension 10 (9) Mixed mechanism 7 (6) Pigmentary 5 (4) Uveitic 4 (3) Traumatic 3 (3) Neovascular 2 (2) Steroid-induced 1 (1) Trabeculectomy location Superior 180° 112 (97) Inferior 180° 3 (3) Antifibrosis agent with last trabeculectomy Figure 1. Kaplan–Meier survival curves for intraocular pressure control in None 29 (25) eyes undergoing cataract extraction after trabeculectomy. 5-Fluorouracil (10–45 mg) 36 (31) Mitomycin C 50 (43) No. of incisional surgeries before CE 1 trabeculectomy 98 (85) within the first postoperative week, 5 at 1.0–4.3 months), six of 2 trabeculectomies 15 (13) whom used additional medications either before or after needling 3 trabeculectomies 2 (2) and two of whom eventually had further glaucoma surgery. The No. of glaucoma medications prior to CE median time to qualified success was 1.6 months (range, 0.03–58.0 1 11 (9) months) and was before 3 months in 10 (77%) of 13 eyes after PE 2 1 (1) and in 10 (45%) of 22 eyes after ECCE. The median time from 3 3 (3) qualified success to failure was 3.6 months (range, 0.3–19.1 Interval from trabeculectomy to CE (mos) months). Six (55%) of 11 failures occurred within 6.5 months All eyes 19.9 6 17.3 after CE. ECCE (N 5 58) 18.6 6 15.5 PE (N 5 57) 21.2 6 19.1 Incision location: ECCE Clear cornea 42 (72) Limbus 16 (28) Incision location: PE Clear cornea 12 (21) Sclera 45 (79) Incision length and type: PE Small (#4 mm) 42 (74) Unsutured 21 (37)

ECCE 5 extracapsular cataract extraction; PE 5 phacoemulsification.

plete successes) was 75% at 1 year, 67% at 2 years, and 59% at 3 years (Fig 1). Cumulative survival did not differ significantly by type of CE (ECCE vs. PE) (Fig 2), but Cox multivariate propor- tional hazards regression, adjusted for other statistically significant variables, estimated a risk ratio of 1.1 for qualified success (95% confidence interval [CI] 0.6, 2.2), and 3.0 for failure (95% CI 0.7, 12.8), for eyes having ECCE compared to PE. Forty-nine patients (43%) required from 1 to 4 medications for IOP control; 32 patients (28%) used medications beyond the second postoperative month, 8 of whom eventually had further glaucoma surgery. Of 15 patients who used glaucoma medications before surgery, 7 used the same number at the last follow-up, 3 Figure 2. Kaplan–Meier survival curves for intraocular pressure control in required more, and 5 used fewer medications. Nine patients (7.8%) eyes undergoing extracapsular cataract extraction and phacoemulsification underwent needling revision of the filtering bleb at the slit lamp (4 (Phaco) after trabeculectomy.

1930 Chen et al z Cataract Extraction after Trabeculectomy

Table 4. Factors Associated with Failure in Eyes Having Cataract Extraction (CE) after Trabeculectomy

Factor N Survival (%)* P† Age at time of CE #50 yrs 7 64 6 21 0.062§ .50 yrs 108 95 6 2 Glaucoma type POAG 57 95 6 3 Uveitic 4 38 6 29 0.007 Time from trabeculectomy to CE #6 mos 18 75 6 11 0.038 .6 mos 97 97 6 2 Early postoperative maximum IOP‡ #25 mmHg 83 97 6 2 .25 mmHg 31 83 6 7 0.008

IOP 5 intraocular pressure; POAG 5 primary open-angle glaucoma. Figure 3. Mean intraocular pressure in eyes undergoing cataract extrac- * Kaplan-Meier survival analysis estimate of 12-month survival rate. tion after trabeculectomy, for all patients and for those eyes that were † Log rank P value. complete successes at last follow-up. ‡ Maximum IOP measured within the first 2 postoperative weeks; not included in multivariate survival regression; data missing from one eye. § Remained significant (P 5 0.033) after Cox multivariate proportional The average preoperative IOP was 10.5 6 3.9 mmHg (range, hazards survival regression analysis of preoperative and intraoperative 3–19). Eyes that became qualified successes or failures had higher factors. mean preoperative IOP compared to those that did not (12.1 6 3.6 vs. 9.8 6 3.8 mmHg; P 5 0.005) (Fig 3). Although preoperative IOP as a continuous variable was not associated with trabeculec- tomy survival, preoperative IOP greater than 10 mmHg was sig- increased to 12.1 6 5.1 mmHg at last visit and was increased nificantly associated with qualified success (P 5 0.036) (Table 3), significantly at each postoperative timepoint through 18 months and this association remained significant (risk ratio, 2.2; 95% CI compared to the pre-CE IOP but did not vary significantly after the 1.0, 4.5; P 5 0.030) after multivariate analysis. first postoperative month. In eyes that were complete successes, The mean maximum IOP during the first 2 postoperative weeks the mean IOP increased significantly from 9.8 6 3.8 to 10.8 6 4.1 was 19.5 6 10.0 mmHg. However, IOP greater than 25 mmHg mmHg at last visit. during this period was noted in 31 patients (27%) in whom Other preoperative variables associated with further glaucoma qualified success (P , 0.001) and failure (P 5 0.008) were surgery by univariate (Kaplan–Meier) analysis included uveitic subsequently more likely compared to those patients who had IOP glaucoma (vs. POAG, P 5 0.007), CE at6months or less after less than 25 mmHg (Tables3and 4). The mean IOP in all eyes trabeculectomy (P 5 0.038), and age of 50 years or younger (P 5 0.062) (Table 4). After multivariate analysis, age of 50 years or younger remained significant (risk ratio, 8.4; 95% CI 5 1.7–40.9; P 5 0.033), and CE greater than 6 months after trabeculectomy Table 3. Factors Associated with Qualified Success in Eyes was borderline protective against failure (risk ratio, 0.5; 95% CI Having Cataract Extraction (CE) after Trabeculectomy 0.2, 1.1; P 5 0.087). No other preoperative variables were shown to be associated with qualified success or failure after CE, includ- Factor N Survival (%)* P† ing use of antifibrosis agents with the most recent trabeculectomy, Preoperative IOP number of prior trabeculectomies, or number of glaucoma medi- #10 mmHg 58 83 6 5 cations. Preoperative bleb appearance was noted to be “avascular,” .10 mmHg 57 66 6 7 0.036\ “good,” “elevated,” or “large” in 91 eyes (79%); these terms were Intraoperative iris manipulation used for 65 (59%) of 111 eyes in which postoperative bleb ap- None 62 83 6 5 pearance was described. The bleb was “small” or “flat” in 8 eyes Any‡ 53 64 6 7 0.027\ (7%) before CE; these terms were used for 25 eyes (22%) after Early postoperative maximum IOP§ surgery. No significant difference in survival was found between #25 mmHg 83 85 6 4 the two sets of terms. ,0.001 Intraoperative iris manipulation was associated significantly .25 mmHg 31 45 6 9 with the need for additional medications or bleb needling after surgery (P 5 0.027, Table 3), and this association withstood IOP 5 intraocular pressure. multivariate analysis (risk ratio, 2.3; 95% CI 5 1.1, 4.5; P 5 * Kaplan-Meier survival analysis estimate of 12-month survival rate. 0.019). Neither the location (clear cornea vs. limbus in ECCE, or † Log rank P value. vs. sclera in PE), length, nor closure of the incision was associated ‡ Includes posterior synechiolysis, stretching, sphincterotomy, iri- with different survival. Only 1 (2.4%) of 42 patients who under- dectomy, iris suturing, and/or use of iris retraction hooks. went small incision PE (length # 4 mm) with foldable silicone § Maximum IOP measured within the first 2 postoperative weeks; not IOL was a failure because of postoperative aqueous misdirection included in multivariate survival regression; data missing from one eye. necessitating pars plana vitrectomy, after which the bleb continued to function. If this case is not considered to be a failure of bleb \ Maintained significance after Cox multivariate proportional hazards sur- vival regression analysis of preoperative and intraoperative factors. function, then large incision (length . 4 mm) CE is associated significantly with bleb failure (10 [13.7%] of 73 eyes; P 5 0.035).

1931 Ophthalmology Volume 105, Number 10, October 1998

Table 5. Studies on Intraocular Pressure (IOP) Control in Eyes Having Extracapsular Cataract Extraction (ECCE) or Phacoemulsification (PE) with Intraocular Lens (IOL) after Glaucoma Filtering Surgery*

Follow-up % Success Study N (ECCE/PE) (months) Definition of Success (ECCE/PE) Comments Alpar15 (1979) 3/7 .12 Not defined 66/86 Excluded if ,1 year follow-up or during CE Binkhorst5 (1981) 26/0 Mean “almost 36” “Normal” IOP and no See comment 3 of 33 eyes required additional glaucoma medications or reoperation, not medications noted whether these were after ICCE or ECCE Obstbaum6 (1986) 15/0 $9 IOP unchanged and 80/NA Excluded if ,9 mos follow-up no increase in glaucoma medications Antonios28 (1988) 29/0 Median 13 Same as current study CS-69/NA QS-31/NA Excluded if ,8 mos follow-up; 7 eyes F-0/NA had IOLs; all had ECCE .1 yr after filtering surgery Murchison7 (1989) 22/0 Mean 22.3 Same as current study CS-68/NA QS-23/NA Excluded if ,6 mos follow-up F-9/NA Brooks30 (1992) 43/0 12 No further glaucoma 98/NA 43 eyes of 33 patients; pre- and post- surgery cataract extraction medications not provided Yamagami8 (1994) 36/9 24 No increase in 67 (see comment) Excluded 6 patients with post-CE iritis; (before exclusion) glaucoma included 13 patients without medications functioning filtering bleb pre-CE; separate analysis of ECCE and PE not provided Dickens9 (1996) 23/0 Median 70 Same as current study CS-65/NA QS-26/NA Excluded if ,4 yrs follow-up F-9/NA Seah14 (1996) 16/6 Mean 13.6 See comment CS-38/67 QS-31/0 Excluded if ,6 mos follow-up; CS 5 F-31/33 IOP # 19 without intervention; QS 5 IOP # 19 with medication; F 5 IOP . 19 or further surgery Present study 58/57 24.5/17.6 See text/Table 1 CS-64/74 QS-22/19 Excluded if , 6 mos follow-up Mean 21.1 F-14/5

NA 5 not applicable; CS 5 complete success; QS 5 qualified success; F 5 failure. * With minimum follow-up of (mean) 6 months.

Intraoperative complications occurred in 7 eyes (6%) and post- 7.5 6 3.6 mmHg (P 5 0.002), no factor was associated signifi- operative complications developed in 31 eyes (27%), including cantly with resolution of hypotony. After CE, two eyes required posterior capsular tear with (3) and without (1) vitreous loss, glaucoma medications for less than 1 month and two eyes under- retained nuclear fragments (2), retained cortical remnants (3), went bleb needling (at 0.3 and 1 month). One eye had pre-CE hyphema (5), transient bleb leak (3) or cataract wound leak (6), hypotony maculopathy, which had not resolved after 7.4 months’ traumatic wound dehiscence (1), aqueous misdirection (1), persis- follow-up. tent corneal edema (5), persistent inflammation (3), IOL capture Of 115 eyes, 110 (96%) had corrected preoperative Snellen (2), vitreous hemorrhage (1), cystoid (4), epiretinal visual acuity of 20/50 or worse. Postoperative visual acuity was membrane formation (1), late bleb-related endophthalmitis (1), and 20/40 or better in 82 eyes (71%) and 20/100 or worse in 16 eyes ptosis (2). Of two eyes noted to have bleb leaks before surgery, one (14%). Only two eyes (1.7%) had worse postoperative visual leaked intermittently after CE. No intraoperative or postoperative acuity during follow-up. The method of CE was not associated complication was associated significantly with different trabecu- significantly with different postoperative visual acuity. lectomy survival, nor was use of postoperative subconjunctival 5-fluorouracil (5-FU) injections (9 patients; total dose, 5–25 mg). Nineteen eyes (17% of all eyes) had hypotony (IOP # 6 mmHg) before CE (mean IOP, 4.6 6 1.1 mmHg). Trabeculectomy had been performed with mitomycin C in 13 eyes and with 5-FU Discussion in 3 eyes. Phacoemulsification was performed in 13 eyes and ECCE in 6 eyes. Eleven eyes (58%) continued to have hypotony Several reports have described the effect of CE on IOP after CE, including 3 eyes that required revision of trabeculectomy control after trabeculectomy.5–9,14,15 Table 5 summarizes for hypotony-related corneal folds, after mean follow-up of 14.6 6 11 months (17 6 10.5 months if eyes censored for trabeculectomy studies with follow-up of at least 6 months and sufficient revision are excluded). The mean peak IOP in the first 2 postop- information to allow comparison with our patients. The few erative weeks was 15.8 6 11.5 mmHg and did not differ signifi- that have investigated outcomes after PE are limited by cantly for those eyes in which hypotony resolved after CE. Al- small sample size (6–9 eyes).8,14,15 Variations in exclusion though the mean IOP at last follow-up increased significantly to and success criteria, follow-up time, and methods of statis-

1932 Chen et al z Cataract Extraction after Trabeculectomy tical analysis make comparisons with previous studies dif- mean follow-up of 24.5 months (Fig 2). However, Kaplan– ficult. In addition, practice patterns may vary by , Meier survival analysis and Cox multivariate proportional and the pre-existing severity of glaucomatous damage un- hazards regression showed no statistically significant differ- doubtedly influences the threshold for adding medications ence between ECCE and PE in progression to qualified or performing further glaucoma surgery, so the use of med- success (risk ratio, 1.1 for ECCE) or failure (risk ratio, 3.0 ical or surgical intervention as an indicator for trabeculec- for ECCE) (the risk ratio is larger for failure, although the tomy failure may not provide the most accurate benchmarks difference between ECCE and PE in Fig2appears larger for for comparison. Nonetheless, the proportion of patients who qualified success, because the standard errors are smaller were in each outcome category in our study was similar to when survival is close to 1). The lack of significance may be that in several previous studies that included primarily because of the small number of failures in our study popu- ECCE with IOL after glaucoma-filtering surgery.7–9,14 We lation, the significant difference in follow-up time between found 22% of patients required only medical intervention or the two methods of CE, and the total sample size. A clini- bleb needling, and 9.6% required further glaucoma surgery. cally important difference in failure between small (#4 We found an encouraging cumulative failure-avoidance mm) and large incision size (2.4% vs. 13.7% failure, re- rate of 93% and 90% at 1 and 2 years, respectively, but the spectively) may not have been statistically significant be- cumulative complete success rate was lower at 75% and cause the one failure after small-incision PE occurred 2 67% at 1 and2years, respectively (Fig 1). Yamagami et al8 weeks after surgery. also used Kaplan–Meier survival analysis and found IOP Some authors have noted mitomycinCto be associated control was maintained in 22 (56%) of 39 eyes 2 years after with long-term protection from subsequent failure despite ECCE or PE; however, 6 eyes with postoperative iritis were severe inflammation.21 Use of antifibrosis agents with the excluded from analysis. most recent trabeculectomy did not influence survival after We identified several preoperative variables that were CE in our study. Both 5-FU and mitomycin C affect local associated significantly (by univariate Kaplan–Meier sur- fibroblast proliferation rather than migration,22,23 and al- vival analysis) with loss of IOP control, including uveitic though mitomycin C may have considerably prolonged ef- glaucoma, period less than 6 months between trabeculec- fects compared to those of 5-FU,23 its half-life after topical tomy and CE, pre-CE IOP greater than 10 mmHg, and age administration is limited,24 and inflammation occurring af- of 50 years or younger. Patients with uveitis have been ter CE may result in further recruitment and migration of reported to have an exaggerated postoperative inflammatory cicatrix-forming cells. In this study, use of 5-FU injections response after CE with IOL implantation.18 Although the after CE did not influence bleb survival and IOP control, but diagnosis of uveitic glaucoma includes numerous open- the number of patients was small and patient selection was angle and angle-closure etiologies of elevated IOP, chronic likely biased toward those with more intraocular inflamma- intraocular inflammation is the underlying factor that may tion or external vascularization. A randomized, prospective contribute to bleb failure after CE. An interval between study would be needed to properly investigate the role of trabeculectomy and CE of 6 months or less was associated 5-FU injections after CE in patients with functioning filter- significantly with reoperation for glaucoma. Other authors ing blebs. have reached similar conclusions and believe the filtering Intraoperative iris manipulation was associated signifi- bleb needs sufficient time to develop properly19,20; the in- cantly with the need for additional medications or bleb flammation associated with cataract surgery presumably needling, probably through long-term postoperative inflam- curtails this process. In our study, patients who maintained mation due to blood–aqueous barrier breakdown25 and the IOP control without additional intervention had a signifi- underlying cause of the posterior synechiae (miotic use, cantly lower mean pre-CE IOP than did those who were diabetes mellitus, previous uveitis, previous surgery, or qualified successes or failures (9.8 vs. 12.1 mmHg). Patients pseudoexfoliation syndrome). In the current study, 46% of with higher IOP before CE may have filtering blebs with patients required iris manipulation. The surgical challenge borderline function, which are likely more susceptible to these patients may pose is reflected in the relatively high fibrosis after CE. Relative youth is a well-recognized risk complication rate seen in this and previous studies.7,8,14 One factor for filtering surgery failure10 and is likely related to study found intraoperative complications frequently led to loss of IOP control after CE for similar reasons. After bleb failure,14 but this association was not seen in our study. multivariate analysis, only IOP greater than 10 mmHg and The mechanism for loss of IOP control in most patients age of 50 years or younger remained significantly associated with previously functioning filtering blebs is external scar- with further glaucoma surgery. An interval greater than 6 ring at the level of Tenon capsule and episclera.26,27 Other months between trabeculectomy and CE was of borderline authors have observed bleb scarring and shrinkage after significance as protective from failure. CE,5,8,9 sometimes with worsened IOP control.8 Reduced Of 57 patients who had PE, 3 (5.3%) required further bleb size was noted in approximately 18% of eyes in our glaucoma surgery after mean follow-up of 18 months. Two study, but change in bleb appearance was not associated patients had uveitic glaucoma and had their cataract inci- with different survival, although this is difficult to assess sions enlarged to 6 and 7 mm for PMMA lens placement. In accurately in a retrospective study. the third patient, small-incision PE precipitated aqueous Different authors have noted changes in mean IOP after misdirection that required pars plana vitrectomy for resolu- CE ranging from a decrease of 0.8 mmHg to an increase of tion, but bleb function was not lost. In comparison, 8 6.6 mmHg, after mean or median follow-up ranging from 8 (13.7%) of 58 patients who had ECCE were failures after to 70 months.5–9,14,28–31 We found an increase in mean IOP

1933 Ophthalmology Volume 105, Number 10, October 1998 of 1.6 mmHg overall, 1.0 mmHg for patients who were 5. Binkhorst CD, Huber C. Cataract extraction and intraocular complete successes. Of note, the mean IOP stabilized lens implantation after fistulizing glaucoma surgery. J Am quickly and did not vary significantly after the first postop- Intraocular Implant Soc 1981;7:133–7. erative month. 6. Obstbaum SA. Glaucoma and intraocular lens implantation. J A marked postoperative IOP rise is not uncommon after Cataract Refract Surg 1986;12:257–61. 7. Murchison JF Jr, Shields MB. An evaluation of three surgical uncomplicated ECCE in patients with glaucoma.32 In our approaches for coexisting cataract and glaucoma. Ophthalmic study, eyes with a postoperative IOP spike greater than 25 Surg 1989;20:393–8. mmHg subsequently had significantly worse IOP control. 8. Yamagami S, Araie M, Mori M, Mishima K. Posterior cham- These patients should be monitored carefully as their bleb ber intraocular lens implantation in filtered or nonfiltered function may be tenuous. Some IOP spikes may be related glaucoma eyes. Jpn J Ophthalmol 1994;38:71–9. to retained viscoelastic material after CE. Mechanical 9. Dickens MA, Cashwell LF. Long-term effect of cataract ex- causes of decreased filtration after CE, such as iris, vitreous, traction on the function of an established filtering bleb. Oph- or IOL haptic incarceration into the internal sclerostomy, thalmic Surg 1996;27:9–14. were not noted in our study. 10. Skuta GL, Parrish RK II. Wound healing in glaucoma filtering Cataract surgery has been advocated to treat postfiltering surgery. Surv Ophthalmol 1987;32:149–70. surgery hypotony maculopathy16,17 and chronic choroidal 11. Araujo SV, Spaeth GL, Roth SM, Starita RJ. A ten-year detachment33 in those patients with coexisting cataract. In follow-up on a prospective, randomized trial of postoperative # after trabeculectomy. Ophthalmology 1995; our series, 19 patients had hypotony (IOP 6 mmHg) after 102:1753–9. trabeculectomy, 11 (58%) of whom continued to have hy- 12. Oshika T, Yoshimura K, Miyata N. Postsurgical inflammation potony or needed bleb revision for hypotony at last follow- after phacoemulsification and extracapsular extraction with up. Although the mean IOP increased significantly from 4.6 soft or conventional intraocular lens implantation. J Cataract to 7.5 mmHg, none of the factors investigated was associ- Refract Surg 1992;18:356–61. ated significantly with resolution of hypotony. Cataract sur- 13. Pande MV, Spalton DJ, Kerr–Muir MG, Marshall J. Postop- gery did not reliably resolve hypotony after filtration sur- erative inflammatory response to phacoemulsification and ex- gery in our study. tracapsular cataract surgery: aqueous flare and cells. J Cataract A decrease in IOP control over time after successful Refract Surg 1996;22(Suppl 1):770–4. trabeculectomy, even without intervening nonglaucoma-re- 14. Seah SKL, Jap A, Prata JA Jr, et al. Cataract surgery after lated surgical intervention, has been reported.2,3,11,34–37 trabeculectomy. Ophthalmic Surg Lasers 1996;27:587–94. Loss of bleb function related to such attrition was not 15. Alpar JJ. Cataract extraction and lens implantation in eyes with pre-existing filtering blebs. J Am Intraocular Implant Soc distinguished from that due to CE in the current study. 1979;5:33–5. Nonetheless, our findings suggest that CE in the presence of 16. Sibayan SAB, Igarashi S, Kasahara N, et al. Cataract extrac- a functioning filtering bleb results in a low cumulative rate tion as a means of treating postfiltration hypotony maculopa- of severe loss of IOP control, and few patients (9.6%) thy [case reports]. Ophthalmic Surg Lasers 1997;28:241–3. require reoperation for glaucoma. Patients 50 years of age or 17. Allingham RR. Treatment of hypotonous maculopathy. In: younger, with IOP of 11 mmHg or greater, or who need Epstein DL, Allingham RR, Schuman JS, eds. Chandler and intraoperative iris manipulation for CE, are at significantly Grant’s Glaucoma, 4th ed. Baltimore: Williams & Wilkins, higher risk for loss of IOP control, as are those with early 1997;549. postoperative IOP greater than 25 mmHg. The mean post- 18. Foster RE, Lowder CY, Meisler DM, Zakov ZN. Extracapsu- operative IOP did not vary significantly after the first post- lar cataract extraction and posterior chamber intraocular lens operative month, and most eyes that needed additional med- implantation in uveitis patients. Ophthalmology 1992;99: ications or needling, or further glaucoma surgery, required 1234–41. 19. Shields MB. Combined cataract extraction and guarded intervention before the 3rd and 7th postoperative months, sclerectomy. Reevaluation in the extracapsular era. Ophthal- respectively. For most patients with glaucoma with func- mology 1986;93:366–70. tioning filtering blebs, CE may be recommended with the 20. Shields MB. Textbook of glaucoma, 4th ed. Baltimore: Wil- anticipation of both restoration of visual acuity and contin- liams & Wilkins, 1998;566. ued IOP control. 21. Yaldo MK, Stamper RL. Long-term effects of mitomycin on filtering blebs. Lack of fibrovascular proliferative response following severe inflammation. Arch Ophthalmol 1993;111: 824–6. 22. Yamamoto T, Varani J, Soong HK, Lichter PR. Effects of References 5-fluorouracil and mitomycin C on cultured rabbit subcon- junctival fibroblasts. Ophthalmology 1990;97:1204–10. 1. Sugar HS. Postoperative cataract in successfully filtering glau- 23. Khaw PT, Doyle JW, Sherwood MB, et al. Prolonged local- comatous eyes. Am J Ophthalmol 1970;69:740–6. ized tissue effects from 5-minute exposures to fluorouracil and 2. D’Ermo F, Bonomi L, Doro D. A critical analysis of the mitomycin C. Arch Ophthalmol 1993;111:263–7. long-term results of trabeculectomy. Am J Ophthalmol 1979; 24. Kawase K, Matsushita H, Yamamoto T, Kitazawa Y. Mito- 88:829–35. mycin concentration in rabbit and human ocular tissues after 3. Mills KB. Trabeculectomy: a retrospective long-term fol- topical administration. Ophthalmology 1992;99:203–7. low-up of 444 cases. Br J Ophthalmol 1981;65:790–5. 25. Ferguson VMG, Spalton DJ. Continued breakdown of the 4. Vesti E. Development of cataract after trabeculectomy. Acta blood aqueous barrier following cataract surgery. Br J Oph- Ophthalmol (Copenh) 1993;71:777–81. thalmol 1992;76:453–6.

1934 Chen et al z Cataract Extraction after Trabeculectomy

26. Addicks EM, Quigley HA, Green WR, Robin AL. Histologic 32. Krupin T, Feitl ME, Bishop KI. Postoperative intraocular characteristics of filtering blebs in glaucomatous eyes. Arch pressure rise in open-angle glaucoma patients after cataract or Ophthalmol 1983;101:795–8. combined cataract-filtration surgery. Ophthalmology 1989;96: 27. Maumenee AE. External filtering operations for glaucoma: the 579–84. mechanism of function and failure. Trans Am Ophthalmol Soc 33. Berke SJ, Bellows AR, Shingleton BJ, et al. Chronic and 1960;58:319–28. recurrent choroidal detachment after glaucoma filtering sur- 28. Antonios SR, Traverso CE, Tomey KF. Extracapsular cataract gery. Ophthalmology 1987;94:154–62. extraction using a temporal limbal approach after filtering 34. Watson PG, Barnett F. Effectiveness of trabeculectomy in operations. Arch Ophthalmol 1988;106:608–10. glaucoma. Am J Ophthalmol 1975;79:831–45. 29. Burratto L, Ferrari M. Extracapsular cataract surgery and intraoc- 35. Costa VP, Katz LJ, Spaeth GL, et al. Primary trabeculectomy ular lens implantation in glaucomatous eyes that had a filtering in young adults. Ophthalmology 1993;100:1071–6. bleb operation. J Cataract Refract Surg 1990;16:315–9. 36. Robinson DIM, Lertsumitkul S, Billson FA, Robinson LP. 30. Brooks AMV, Gillies WE. The effect of cataract extraction Long-term intraocular pressure control by trabeculectomy: with implant in glaucomatous eyes. AustNZJOphthalmol aten-yearlifetable.AustNZJOphthalmol 1993;21: 1992;20:235–8. 79–85. 31. Drolsum L, Haaskjold E. Extracapsular cataract extraction in 37. Chen TC, Wilensky JT, Viani MAG. Long-term follow-up of eyes previously operated for glaucoma. Acta Ophthalmol initially successful trabeculectomy. Ophthalmology 1997;104: (Copenh) 1994;72:273–8. 1120–5.

1935