The Effects of Bevacizumab in Augmenting Trabeculectomy for Glaucoma

Abstract The aim of the study was to assess the effects of bevacizumab in augmenting trabeculectomy for glaucoma. We searched the databases of Cochrane Library, PubMed, Embase, CNKI, and VIP. All the databases were retrieved from the time databases established to September, 2015. The keywords we used were as follows: “bevacizumab,” “anti-VEGF,” “avastin,” “trabeculectomy,” “glaucoma,” and so on. We used a method of the freedom word search and the MeSH search combined, which was recommended by Cochrane Systematic Review Manual 5.1.2. Randomized controlled trails (RCTs) of frequently used bevacizumab in trabeculectomy for glaucoma were included. Study selection, data extraction, quality assessment, and data analysis were performed according to the Cochrane standards. Eight randomized controlled trails involving 212 eyes in the experimental (bevacizumab or bevacizumab + mitomycin C) groups and 214 eyes in the control (mitomycin C or placebo) groups were selected. Compared with placebo, bevacizumab significantly increased the complete success rate [OR = 2.79, 95%CI, (1.47, 5.29), P = 0.002], what else, bevacizumab also significantly decreased the intraocular pressure (IOP) [MD = 3.07, 95% CI, (0.87, 5.27), P = 0.006] at the 6-month after trabeculectomy and the number of antiglaucoma medications [MD = 1.23, 95% CI, (0.66, 1.80), P < 0.0001]. Additionally, it also increased the risk of bleb leak [OR = 5.24, 95% CI, (1.30, 21.10), P = 0.02]. When compared with mitomycin C (MMC), bevacizumab significantly increased the rate of encysted blebs [OR = 4.62, 95% CI, (1.02, 20.91), P = 0.05]. However, there was no significantly difference between the bevacizumab + MMC groups and MMC groups whatever the items were. Bevacizumab was an effective way in trabeculectomy concerning the complete success rate, IOP, and anti-glaucoma medications reduction when compared with placebo; however, it increased the risk of bleb leakage. And it significantly increased the rate of encysted blebs compared with MMC.


INTRODUCTION
G laucoma as the second reason of blindness is a serious threat to human vision health in the world. 1 Most commonly, the clinical course of open angle glaucoma is so insidious that the problem is found only when the visual function suffers serious damage. Thus, for open angle glaucoma, early diagnosis and treatment are very important. Usually, the operation indications of open angle glaucoma are uncontrolled cases with drugs, cases cannot tolerate medications. However, some researchers thought that once the diagnosis was clear, with significant disc and vision changes, filtration operations should be used as the preferred treatment. 2,3 Trabeculectomy is the main technique of open angle glaucoma. 4 There are, however, some limitations of the surgery. Scar formation and fibrosis in the process of wound healing may result in obstruction of filtration tract, leading to the operation failure. 5,6 In recent 3 decades, due to the use of antimetabolites, such as MMC and 5-fluorouracil (5-FU), the rate of operation success has been higher than before. [7][8][9][10] However, antimetabolites may bring some serious complications, such as low intraocular pressure, filtering bleb leakage, filtering bleb-associated endophthalmitis, epithelial toxicity, and so on. 11 Hence, researchers have been searching for more effective and safer ways to inhibit scar formation and fibrosis. Recently, some researchers have found that bevacizumab may work in some ways. [12][13][14] However, studies about this aspect were few, and highquality researches were also seldom seen. Whether glaucoma patients after trabeculectomy could benefit more from bevacizumab than MMC or placebo, it has not been reviewed yet. The purpose of this study is to systemic review the efficacy and safety of bevacizumab in the trabeculectomy, providing more reliable evidences for clinical workers.
the time databases established to September, 2015. The keywords we used were as follows: ''bevacizumab,'' ''anti-VEGF,'' ''avastin,'' ''trabeculectomy,'' ''glaucoma,'' and so on. We used a method of the freedom word search and the MeSH search combined, which was recommended by Cochrane Systematic Review Manual 5.1.2. For a more comprehensive search, a manual search of cited references in published studies was done. Two researchers selected and assessed all included studies independently, and then cross-checked. Due to the fact that all analyses were based on previously published studies, the ethical approval was not necessary for our study.

Data Extraction
Two researchers extracted study characteristics and outcome data independently. If there were some discrepancies, they would be resolved through discussion or a third researcher. Data that we collected were as follows: baseline characteristics, IOP, best-corrected visual acuity (BCVA), complete success rate (CS), quality success rate (QS), failure rate, the number of glaucoma medications, and adverse events.

Statistical Analysis
Revman 5.0 (the Cochrane collaboration; http://www.cochrane.org/) was used for statistical analysis of the data. For continuous outcomes, mean difference (MD) or standard mean difference (SMD) with 95% confidence intervals (CI) was used to calculate the results; however, odds ratio (OR) with 95% confidence intervals (CI) was used for dichotomous outcomes. We used the chi-square test to assess heterogeneity between trials and the I2 statistic to assess the extent of inconsistency. If there was a significant heterogeneity, a random-effect statistical model would be used to confirm the case results. A fixed-effect model for calculations of summary estimates was applied, unless there was a significant heterogeneity. Subgroup analysis was intended to explore clinical differences among trials.

Search Results
We obtained 101 publications through searching literature databases and cited references. According to the inclusion criteria, only the RCTs for patients using bevacizumab during trabeculectomy were included. We eliminated the 74 articles by reading the title and abstract. Through further reading the full text, we ruled out 19 published papers, including 2 nonrandomized controlled trials, 7 retrospective case series, 6 retrospective controlled trials, and 4 prospective case series. Finally, we included 8 RCTs 15-22 about use of bevacizumab in augmenting trabeculectomy for glaucoma in the metaanalysis. The process of literature screening was shown in Figure 1.   Table 1 described the specific information of the RCTs. A total of 426 eyes with 212 eyes in the experimental (bevacizumab or bevacizumab þ MMC) groups and 214 eyes in the control groups separately were included in them. Figure 2 showed the methodological quality of the included RCTs, which was assessed by using the Cochrane Handbook 5.0.2. Seven studies [16][17][18][19][20][21][22] of the included studies offered adequate descriptions of the randomization process. Five studies 16,17,19,21,22 reported that masking was done either for the patients or for the practitioners; only 4 studies 16,18,21,22 adequately stated allocation concealment. Six of included studies [15][16][17][18]21,22 had stated incomplete outcome data. Furthermore, none of the papers adequately described other bias.

Studies and Baseline Characteristics
Characteristics of the 8 trials [15][16][17][18][19][20][21][22] were shown in Table 1. Trabeculectomy were performed under local anaesthesia by experienced surgeons. All patients received bevacizumab, MMC, normal saline, and bevacizumab þ MMC after trabeculectomy in the groups of bevacizumab, MMC, placebo, and bevacizumab þ MMC respectively. Six trials reported antiglaucoma medications before trabeculectomy, and 3 studies reported the type of glaucoma including primary open-angle glaucoma (POAG) and pseudoexfoliation glaucoma (PEXG). The baseline characteristics of participants were displayed in Table 1. All studies were published between 2011 and 2015. Follow-up ranged from 3 to 24 months. There were comparable throughout age, IOP, BCVA, glaucoma medications, and glaucoma type in the papers.

Complete Success Rate
Seven studies [15][16][17][18]20,21 reported the complete success rate. The complete success rate of the bevacizumab groups was significantly higher than the placebo groups (OR ¼ 2.

Adverse Events
We could analyze 7 studies 15-18,20-22 for adverse events including bleb leak, hyphema, encysted blebs, anterior chamber shallowing, and so on. Fortunately, there was no statistically significant difference between bevacizumab and control groups, including the placebo groups (OR ¼ 1.

Bleb Leak
We could analyze 5 studies 15,17,18,20,22 for the bleb leak, and there was statistically significant difference between bevacizumab and placebo groups (OR ¼ 5.  Figure 10B). Therefore, bevacizumab was associated with significantly increased the rate of bleb leak compared with placebo groups.

Hyphema
There were 5 studies 15,16,18,20,22 reported the rate of hyphema. There was no statistically significant difference between bevacizumab and control groups, including the placebo groups (OR ¼ 0.

Bleb Morphology
There were 5 studies 15,18,[20][21][22] reported bleb characteristics. Two 15,20 showed no significantly difference between the experimental (bevacizumab or bevacizumab þ MMC) groups and control groups. Two 21,22 found that the vascularity scores of the experimental (bevacizumab or bevacizumab þ MMC) groups were significantly lower when compared with the control groups at the 1-month follow-up. But these were not retained for longer time. One 18 showed a statistically significant difference between 2 groups in regard to maximal bleb area, with the control group exhibiting more diffuse bleb area.

DISCUSSION
The failure of trabeculectomy is mainly due to fibrosis and scar formation of subconjunctival tissue around the scleral flap and bleb during the wound-healing process. 23,24 Bevacizumab, a humanized nonselective monoclonal antibody against vascular endothelial growth factor (VEGF), has been successfully used for diabetic retinopathy (DR), 25 neovascular glaucoma, 26,27 may work in some ways. As is known to all, tissue growth requires nutrients which provided by blood. Thus, bevacizumab are expected to act a role of inhibiting scar formation and fibrosis through the inhibition of angiogenesis information. 28 On the other hand, the vascularization of conjunctiva is an important reason of bleb filtration failure. What is more, there were also evidences showed that VEGF had a direct effect on fibroblasts, which if inhibited by bevacizumab, scar formation, and fibrosis would be modulated. [28][29][30][31][32] Previous studies found that the VEGF levels were elevated in patients who had a trabeculectomy. 33,34 And the concentration of VEGF was significantly reduced after application of bevacizumab. 32,35 Thus, bevacizumab may have the potential to work in trabeculectomy.
In the present study, 8 RCT studies were reviewed, consisting of 3 studies about bevacizumab vs placebo, 3 about bevacizumab vs MMC, and 2 about bevacizumab þ MMC vs MMC. We found similar efficacy of reduction in the IOP and BCVA in the experimental (bevacizumab or bevacizumab þ MMC) groups and control groups at last visit. Because of the lack of data reported in all phases of follow-up and trials with different durations, we chose the data from the end-point. The operative failure rate was also similar between the 2 groups. There were 5 studies, including 1 in the bevacizumab vs placebo groups, 3 in the bevacizumab vs MMC groups, and 1 in the bevacizumab þ MMC vs MMC groups, reported IOP at the 6month; we found the change of IOP was more remarkable in bevacizumab groups when compared with placebo groups For safety, results of adverse events were reported in 7 studies. [15][16][17][18][20][21][22] Concerned overall adverse events, there was no statistically difference between the experimental (bevacizumab or bevacizumab þ MMC) groups and control groups. And no one died patient was associated with bevacizumab and MMC in including studies. The adverse events included bleb leak, hyphema, encysted blebs, anterior chamber shallowing, hypotony, and so on. This meta-analysis showed bevacizumab not only increased the rate of bleb leak compared with placebo groups, but also increased the rate of encysted blebs compared with MMC.
Concerned with bleb morphology, 2 studies 21,22 found bevacizumab had some advantages in reduce the vascularity scores in 1 month, which was similar to a recent cohort study. 36 This might be associated with mechanism of bevacizumab, inhibiting the angiogenesis information. Akkan and Cilsim 17 reported that the bevacizumab showed less efficiency in diffuse bleb area. This was in contrast with 1 recent study, 36 revealing the bevacizumab group had greater extent.
Despite bevacizumab and MMC had similar efficacy in the IOP reduction and success rate, bevacizumab was much more expensive than MMC, with approximately $450 for each bevacizumab vial. 37 If we use each vial of bevacizumab for multiple injections, the per dose price will potentially much lower than $450, depending on the number of injections per vial. However, each bevacizumab vial was allowed to use for only 1 injection because of the contamination outbreaks, discarding the leftover amount. Therefore, MMC might be the preferred choice concerned cost-effectiveness.
The present study is the meta-analysis that evaluates the efficiency and safety of bevacizumab in trabeculectomy. All the studies we included were RCT studies. Seven studies [16][17][18][19][20][21][22] of the included studies offered adequate descriptions of the randomization process. The randomization process of 6 studies [16][17][18][20][21][22] was generated by computer. Five studies 16,17,19,21,22 reported that masking was done either for the patients or for the practitioners; only 4 studies 16,18,21,22 adequately stated allocation concealment. Six of included studies [15][16][17][18]21,22 had stated incomplete outcome data. Furthermore, none of the papers adequately described other bias. Of course, there are some limitations in our meta-analysis that should be taken into consideration when considering the results. First, the number of RCTs and the sample sizes of these studies were very small, all of the studies 15-22 enrolled only 426 eyes, resulting in the possibility of false-negative statistical error. Second, the varying definitions of surgical success in the literature and absence of patient's stratification into different types of glaucoma and risk of surgical failure should be taken into consideration. Furthermore, the different operative methods and procedures were performed by different surgeons would lead to an unavoidable potential bias. Additionally, the data came from the end-point owing to the lack of data reported in all phases of follow-up and trials with different durations introduced a potential heterogeneity. Finally, publication bias was inevitable.

CONCLUSION
From the current evidences, we found bevacizumab was an effective way in trabeculectomy concerned the complete success rate, IOP, and antiglaucoma medications reduction when compare with placebo, but bevacizumab did not show any advantages when compared with MMC. However, bevacizumab not only increased the rate of bleb leak compared with placebo groups, but also increased the rate of encysted blebs compared with MMC. What is more, there was no difference between bevacizumabþ MMC and MMC whatever the items were. However, MMC might be the preferred choice concerned cost-effectiveness. Further intensive RCTs of large sample, high-quality, multiple centres, and vary phases of follow-up should be carried out to provide more clear and reliable evidence.