A comparative study of mitomycin C and 5-Fluorouracil trabeculectomy in west Africa.

PURPOSE: To report the comparative efficacy and safety of intraoperative 5-fluorouracil (5-FU) or mitomycin C (MMC) in primary trabeculectomy in Nigeria.
MATERIALS AND METHODS: Retrospective chart review of patients undergoing primary antimetabolite trabeculectomy in Lagos, Nigeria between 1996 and 2003. We included 129 patients (132 eyes) of the 210 patients with greater than one year postoperative follow-up. Success rates between groups were compared by Kaplan-Meier survival analysis. IOP changes between groups were compared with ANOVA test. Non-parametric comparisons were performed with the Chi-square test with Yates correction or Fisher exact test. A P value less than 0.05 was considered statistically significant.
RESULTS: Seventy-three eyes underwent 5-FU (5-FU group) and 59 eyes underwent MMC augmentation (MMC group) during primary trabeculectomy. The 5-FU group had longer mean follow-up of 53 ± 26 months than the MMC group (38 ± 18 months, P<0.001). The preoperative intraocular pressure was 25.4 ± 6.2 in the MMC group and 25.8 ± 6.0 mm Hg in the 5FU group (P=0.8). Postoperative IOPs were significantly lower (P<0.05) in the MMC group at all follow up visits except between 30-35 months (P=0.07). The probability of maintaining an IOP less than 19 mmHg and 15 mmHg without additional medication or needle revisions at 2 and 3 years postoperatively was 71 (95% CIs, 54-82%) and 64% (95% CIs, 53-76%) respectively for the 5FU group and 81 (95% CIs,71-92%) and 79% (95% CIs,69-90%) respectively for the MMC group. The MMC group had significantly better survival times, both for IOP less than 19 mm Hg (P=0.03) and IOP less than 15 mm Hg (P= 0.006). At last follow up, 40 eyes (30.3%) had lost more than 2 lines of Snellen visual acuity, 24 from 5-FU and 16 from the MMC group (P=0.8). The MMC group was statistically less likely than the 5-FU group to require medications (18.5% vs. 41.1%, P =0.007) or needle revisions (5.1% vs. 17.8%, P=0.03) to control IOP. Blebitis and endophthalmitis developed in one eye each in both groups. Persistent hypotony was observed in 4 eyes (6.8%) in the MMC group only.
CONCLUSIONS: In this study of Nigerian patients, intraoperative application of MMC was more efficacious than 5-FU in lowering IOP following primary trabeculectomy. However, delayed ocular hypotony was only seen with MMC use.

INTRODUCTION

Glaucoma is the world's second most common blinding condition causing irreversible visual loss.1 Epidemiologic studies in the West suggest that the prevalence of primary open angle glaucoma is the highest in people of African descent.2–5 The prevalence may be as high as 8% in the population over 30 years of age in West Africa.6 In West Africa, medical treatment is often impractical due to a variety of reasons including cost and lack of patients’ understanding of the disease.78 Surgery is often performed too late in the course of disease and the results are often poor due to an aggressive healing response.910 Studies on African populations have suggested a high failure rate of trabeculectomy that is not augmented with a metabolite.11–13 A randomized, prospective study from South Africa reported a 30% risk of failure after primary trabeculectomy compared to a 5% failure rate at one year with beta-irradiation application during surgery.14 Intraoperative 5-FU application appears to improve the outcomes of trabeculectomy in African patients.15–17

Short and long-term IOP outcomes from the same cohort of patients in Ghana suggest better results with MMC than 5-FU.1819 The aim of this study is to compare the long-term outcomes of trabeculectomy augmented with either 5-FU or MMC in a cohort of Nigerian patients.

MATERIALS AND METHODS

We performed a retrospective chart review, searching for patients who underwent glaucoma surgery at the Maja Eye Hospital at Lagos, Nigeria. Primary trabeculectomies between January 1998 and January 2004 performed by the two authors (NA and VKD) who visited this center from the United Kingdom were included. Eyes with less than 12 months follow-up were excluded. Postoperatively, local doctors managed the patients. The same surgeons (NA and VKD, who visited the hospital in every three-six months) performed needle revisions. The cohort included 129 of 210 patients (61.4%) with follow up greater than one year. In cases of bilateral surgery, the eye that underwent the first surgery was included. However if different antimetabolites were used during surgery in bilateral cases, both eyes of the same patient were included.

A Microsoft Access™ (Microsoft Corp., Redmond, Wa., USA) database was used to collect data. Data collected included patients’ details, ocular diagnosis, co-existing ocular pathology, previous ocular surgery, preoperative IOP, Snellen visual acuity (VA) and glaucoma medications. Operative details and intraoperative complications were noted. The followings were recorded for each postoperative visit: VA, IOP, complications and medications. Subsequent surgeries were noted. Visual field tests were performed sporadically with a tangent screen and therefore the visual field data are not useable for this study.

All patients underwent a Cairns-type trabeculectomy with slight modifications in technique described previously.16 A fornix-based conjunctival flap was dissected. Cellulose sponge fragments soaked in 5-FU, 50 mg/ml were placed between the dissected conjunctiva and sclera for 5 min. MMC 0.2 mg/ml was similarly applied for 3 min. The exact area of antimetabolite application could not be determined from the case-notes. The antimetabolite was then washed with 10-20 ml of balanced salt solution. The scleral flap was either rectangular (4 × 3 mm, NA) or triangular (3 mm at base, VKD) and the depth of dissection was half to two-thirds scleral thickness. After the excision of a block of corneo-scleral tissue, a peripheral iridectomy was performed and the scleral flap was closed with one to five 10/0 nylon sutures. The conjunctival flap was closed with 10/0 nylon sutures. Subconjunctival betamethasone 1% and an antibiotic were instilled at end of surgery. Patients used a topical steroid-antibiotic combination for at least six weeks postoperatively. In the absence of the visiting surgeons, the local doctors took decisions on glaucoma therapy.

Two criteria for success were used for survival analyses. IOP of less than 19 and 15 mmHg, a decrease of 20% from preoperative IOP were used for Kaplan–Meier survival analyses. Loss of light perception and any subsequent procedure to lower IOP, including needle revision, was considered to be a failure. Survival data was censored when extra-capsular cataract extraction (ECCE) was performed due to the unfavorable effect of cataract surgery procedures on bleb function.20–23 Univariate ccomparison of actuarial success between groups was performed by the log-rank test. IOP changes between groups were compared with the ANOVA test. Non-parametric comparisons were performed by the Chi-square test with Yates correction or Fisher exact test as indicated. All tests were two-tailed with a significance level of P< 0.05. All statistical analyses were performed using Medcalc™ Software (Broekstraat 52, 9030 Mariakerke, Belgium).

RESULTS

We included 132 eyes of 129 patients with a follow-up of a year or more. There were 73 eyes in the 5-FU group and 59 in the MMC group. Preoperative characteristics are shown in Table 1. The 5-FU group had a significantly longer follow-up (P=0.004). IOP changes after surgery are presented in Figure 1. The MMC group had significantly lower pressures at all postoperative visit except between 30 and 35 months (P=0.07). Figures 2 and 3 show the outcomes of survival analyses. The probability of maintaining an IOP less than 19 mmHg at 1, 2 and 3 years after surgery was 78% (95% confidence intervals (CI): 69-88%,), 71% (95% CI: 54-82%) and 59% (95% CI: 48-73%) in the 5-FU and 88% (95% CI: 80-97%), 81% (95% CI: 71-92%) and 76% (95% CI: 66-89%) in the MMC group, respectively. The probability of maintaining an IOP less than 15 mmHg was 71% (95% CI: 62-82%), 64% (95% CI: 53-76%) and 55% (95% CI: 44-70%) in the 5-FU, and 86% (95% CI: 78-96%), 79 % (95% CI: 69 -90%) and 76% (95% CI: 66-89%) in the MMC group, respectively. The MMC group had significantly better survival times, both for IOP less than 19 mm Hg (P=0.03) and IOP less than 15 mm Hg (P= 0.006).

Table 1

Preoperative parameters

Figure 1

Postoperative change in intraocular pressure (IOP) after trabeculectomy. Error bars represent 95% confidence intervals. Broken lines represent IOP changes in the 5-Fluorouracil (5-FU) group

Figure 2

Survival times by treatment group with an intraocular pressure less than 19 mm Hg. Group 0 (broken line) represents the 5-Fluorouracil (5-FU) group and Group 1 (solid line) represents the Mitomycin C (MMC) group

Figure 3

Survival times by treatment group with an intraocular pressure less than 15 mm Hg. Group 0 (broken line) represents the 5-Fluorouracil (5-FU) group and Group 1 (solid line) represents the Mitomycin C (MMC) group

Early and late complications are presented in Table 2. The 5-FU group had a higher incidence of early conjunctival flap edge leaks, shallow anterior chambers and hyphema; yet, none of the differences were statistically significant (P>0.05, all cases). Cystic areas within blebs were noted more frequently in the MMC group (40 eyes (68%)) compared to the 5-FU group (33 eyes (44%), P=0.01). Delayed bleb leaks were seen in 3 eyes of the 5-FU and 4 eyes of the MMC group. Delayed hypotony, defined as an IOP less than 6 mmHg which developed at least 6 months postoperatively and persisted for more than 6 months, was observed in 4 eyes (6.8%) of the MMC group. Maculopathy with hypotony was observed in only one of these eyes, which also had a delayed bleb leak. This patient underwent bleb repair, ECCE and then a posterior capsulotomy. The VA decreased from 6/6 preoperatively to 6/36 at last visit. There was no significant difference between groups in the frequency of late complications.

Table 2

Complications after trabeculectomy

Preoperatively, VA of 6/18 or less was noted in 55 eyes (41.7 %). By last follow up, 40 eyes (30.3 %) had lost more than 2 lines of Snellen VA, 24 (32.8%) from the 5-FU group and 16 (29.3%) from the MMC group (P=0.8). Probable cause of loss of 2 lines or more of VA are listed in Table 3. Cataract and cataract surgery were implicated in at least 25 eyes (18.9%).

Table 3

Causes of decreased Snellen visual acuity after trabeculectomy

All procedures done after trabeculectomy are listed in Table 4. Extracapsular cataract extraction (ECCE) surgery was the most common postoperative procedure, done in 55 eyes (34.1%), of which 24 (32.9%) were in the 5-FU and 21 (35.6) in the MMC group (P=0.6). ECCE was associated with significant complications, which are listed in Table 5. Needle revision was performed more frequently in the 5-FU group (P=0.01). Glaucoma medications were commenced postoperatively for 30 eyes (41%) of the 5-FU and 11 (20.4%) of the MMC group (P=0.02).

Table 4

Surgical procedures after trabeculectomy

Table 5

Complications of extra-capsular cataract extraction after trabeculectomy

DISCUSSION

In African patients, a successful outcome of trabeculectomy may be compromised by an aggressive healing response.924 Therefore, antimetabolites such as 5-FU or MMC can be used. MMC has a more profound effect on inhibiting ocular wound healing than 5-FU in eyes with an increased risk of failure.2526

In a short-term, comparative, prospective study of 81 patients undergoing primary trabeculectomy in Ghana, Singh et al., reported that the final mean IOP was 13.7 mmHg in the MMC group and 16.3 mmHg in the 5-FU group.27 More eyes achieved an IOP less than 21 mmHg in the MMC group (P=0.1). No differences were noted in visual acuity and complication rates.27 In a subsequent retrospective study with three years or longer follow-up, they reported a significantly higher proportion of eyes with MMC augmentation achieved an IOP less than 21 mmHg without medications.18 Our study confirms Singh et al.'s observations.

Trabeculectomy accelerates cataract formation.2829 This is particularly true in eyes with early complications such as flat anterior chamber, hyphema and marked anterior chamber inflammation.30 In most of West Africa, ultrasonic biometry and phacoemulsification are unavailable. Extracapsular cataract extraction (ECCE) is the standard technique. Cataract extraction in an eye with a trabeculectomy may lead to reduction of bleb function and loss of IOP control.29 There are numerous technical problems in performing ECCE in eyes with posterior synechiae, small pupils and overhanging filtration blebs. Superior corneal sections tend to be more anterior as the surgeon tries to avoid the filtration blebs. The surgically induced astigmatism can be quite high. Patients will not often get corrective spectacles because of anisometropia or cultural perceptions. The incidence of posterior capsular tears in our patients was quite high at 11%, due to variety of reasons, including difficulty in performing an adequate capsulotomy, poor visibility of the posterior capsule and the difficulty of delivering nuclei through small pupils. Combined surgery is an option. A recent study from East Africa showed improvement of visual acuity in 75% of patients after combined trabeculectomy and cataract surgery.31 Phacoemulsification was performed in 80% of the cases. IOP data was limited as only 30% had a follow up of three months or more. The IOP was between 5 and 21 mmHg in 81% of these patients. The duration of follow-up was influenced by how far the patient lived from the treatment center.31

Studies indicate that the visual acuity may decrease after trabeculectomy. Cataract may be a common cause. Kirwan et al.,14 in their prospective study in South Africa reported that 26-28% of the patients lost more than 0.25 logMAR units (approximately 2 Snellen lines) after a mean follow-up of about two years after beta-irradiation and trabeculectomy. About 30% of patients lost VA of 2 lines or more in this case series.

There are significant limitations of this study. This was a retrospective study and the drop-out rates were high with 38% not included in analyses, and the 5-FU group had significantly longer follow-up times. Hence there may be an inherent bias against intraoperative 5-FU application. High dropout rates imply that we may be underestimating the incidence of delayed complications such as bleb leaks and infections. The cause of decreased visual acuity was not always clear from case-notes. There were no visual field data. Assessment of bleb morphology can be very subjective. The ability to recognize encapsulated bleb by non-glaucoma specialists is doubtful. For example, encapsulated blebs were reported more frequently by specialists in university hospitals than in district hospitals in the UK.32

Cystic areas on the bleb, on the other hand, are easily seen and recognized by all ophthalmologists. In this study, cystic blebs were noted more frequently after MMC trabeculectomy (68%). Although bleb leaks were observed with equal frequency in both groups, delayed hypotony was observed only in eyes with MMC augmentation. Mermoud et al., had noted that 63% of blebs were avascular after MMC trabeculectomy in South African blacks and 13 % had a late bleb leak.33 This is in marked contrast to the reports from Ghana where no cases of bleb leaks or hypotony were noted after trabeculectomy with MMC.182734 This may be due to differences in MMC application technique or limited follow-up in the reports from Ghana. The finding of a significantly higher number of needle revisions and postoperative glaucoma medications in the 5-FU group in this study have to be interpreted with caution as the 5-FU group had a much longer follow-up than the MMC group.

In conclusion, this study suggests that better IOP outcomes may be achieved by using MMC than 5-FU during primary trabeculectomy in West African patients. However, there is a significant risk of delayed bleb leaks with use of both antimetabolites. Ocular hypotony may also be of concern in the long-term with MMC use during trabeculectomy.