Article Text

Download PDFPDF
Trabeculectomy with mitomycin C alone or coupled with intracamerular bevacizumab? A 2-year comparative study
  1. Patrícia José1,2,
  2. Filipa Jorge Teixeira1,2,
  3. Rafael Barão1,2,
  4. David Cordeiro Sousa1,2,3,4,
  5. Raquel Esteves Marques1,2,
  6. Andre Diogo De Oliveira Barata1,2,5,
  7. Carlos Marques-Neves1,2,3,
  8. Marta Alves6,
  9. Ana Luísa Papoila6,
  10. Ingeborg Stalmans7,
  11. José Pedro Silva5,
  12. Luis Abegão Pinto1,2,3,5
  1. 1 Ophthalmology, Centro Hospitalar Lisboa Norte, Lisboa, Portugal
  2. 2 Ophthalmology University Clinic, Faculdade de Medicina, Universidade de Lisboa, Lisbon, Portugal
  3. 3 Vision Sciences Study Center, Faculdade de Medicina, Universidade de Lisboa, Lisboa, Portugal
  4. 4 Vitreoretinal Unit, Royal Victorian Eye and Ear Hospital, Melbourne, VIC, Australia
  5. 5 Ophthalmology, Hospital Lusíadas Lisboa, Lisboa, Portugal
  6. 6 CEAUL, Biostatistics Department, NOVA Medical School, Universidade Nova de Lisboa, Lisbon, Portugal
  7. 7 KU Leuven University Hospitals Leuven, Leuven, Belgium
  1. Correspondence to Dr Patrícia José, Ophthalmology, Hospital de Santa Maria, Lisboa 1649-035, Portugal; patricialopes1{at}


Purpose To compare outcomes of primary trabeculectomy using either mitomycin C (MMC) alone versus MMC augmented with intracamerular bevacizumab in patients with open-angle glaucoma.

Methods Retrospective, cohort, two-centre, comparative study. Patients’ data were screened between October 2015 and March 2019, with inclusion requiring a minimum follow-up of 24 months. Primary outcome was intraocular pressure (IOP) lowering at 24 months, with surgical success defined with different maximum IOP targets (≤18, ≤16 and ≤14 mm Hg) and at least 30% reduction and higher than 5 mm Hg. Absolute success was achieved if no IOP-lowering medication was needed and a qualified success if otherwise. Safety outcomes were analysed.

Results A total of 110 eyes underwent trabeculectomy with MMC, 51 of these combined with intracamerular bevacizumab. Both strategies were effective in terms of IOP lowering (baseline vs 2 years postoperatively: 24.4 (8.0) mm Hg vs 12.1 (5.3) mm Hg in the MMC group; 25.1 (8.7) vs 10.8 (3.8) mm Hg in the MMC+bevacizumab group; p<0.001 in both comparisons). The MMC+bevacizumab group had a significant difference towards higher efficacy on absolute success rates at all targets (IOP≤14 or ≤16 or ≤18 mm Hg; p=0.010, p=0.039 and p=0.007, respectively). The large majority (93%) of the MMC+bevacizumab group was drop-free at 24 months, and 41% had IOP below 10 mm Hg. Complication rates were low and similar between groups, with no systemic adverse events.

Conclusions Intracamerular bevacizumab in MMC-augmented primary trabeculectomy increases the chances of obtaining low IOP outcomes. This strategy may be useful when planning for surgeries aiming at target pressures in the low teens.

Trial registration number ISRCTN93098069.

  • glaucoma
  • aqueous humour
  • optic nerve
  • treatment surgery

Data availability statement

Data are available in a public, open access repository.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


The only proven treatment for glaucoma has been to significantly decrease intraocular pressure (IOP), its main risk factor. Trabeculectomy remains the gold-standard procedure performed with intraoperative cytotoxic agents (such as mitomycin C (MMC)) to achieve significant IOP-lowering effect, to prevent scarring and fistula occlusion.1 2 However, even resorting to these adjuvants is not synonym of success, with long-term success in the range of 60%–80%, depending on the criteria used.3 4 Accordingly, extensive research has been performed to identify other agents that can either replace or be added to the existing MMC, allowing an increase in efficacy without jeopardising the safety of the procedure.

Modulating angiogenesis and inhibiting fibroblast proliferation have been used for this purpose. Indeed, it has been demonstrated that the existence of increased levels of vascular endothelial growth factor (VEGF) in aqueous humour of patients with glaucoma and its receptors are expressed on tenon fibroblasts. Moreover, the use of antiangiogenic factors has been associated with slowing this proliferative and fibrotic process after filtration surgery.5 6 Bevacizumab is a full-length humanised monoclonal antibody against VEGF, but, interestingly, its clinical use as a wound modulator in trabeculectomy has not been consensual. One prospective randomised trial report a positive effect at 1 year, but other studies could not confirm these findings.7–13 Moreover, data beyond 1 year are not available.

This study aimed to determine 24-month outcomes comparing primary trabeculectomy using MMC of 0.4 mg/mL with similar protocol but augmented with intraoperative intracamerular bevacizumab injection (1.25 mg/0.05 mL).

Materials and Methods


A retrospective cohort study was performed to evaluate the efficacy and safety of primary trabeculectomy in primary open angle glaucoma with MMC of 0.4 mg/mL (group 1) or MMC (same dosage) plus intraoperative intracamerular bevacizumab injection (50 µL–1.25 mg) (group 2). Patients from two surgical centres were included: Hospital Santa Maria (HSM) and Hospital Lusíadas (HLL), Lisbon, Portugal, between October 2015 and March 2019. All surgeries were a primary filtering surgery and were performed by the same glaucoma surgeons (LAP and ADDDOB in both centres) using identical conditions and techniques of the surgery in two settings. Of note, group allocation relates to real-life settings as hospital pharmacy approval for bevacizumab in glaucoma surgery at one site (HSM) was achieved only in March 2019. Accordingly, all surgeries performed at that centre during this period were with MMC only, while in the other centre (HLL), surgeries were routinely done using intracamerular bevacizumab in addition to MMC.

This study was registered on ISRCTN database, prior to any data collection (9 April 2020).


Indications for surgery were based on (1) IOP associated with high probability of glaucoma progression, (2) progressive glaucoma damage according to visual field loss or signs in optic disc, and (3) allergy or intolerance to current maximal topical therapy.

Inclusion criteria for analysis were age>18 years, absence of combination with cataract surgery, absence of history of another eye disease, anbsence of secondary glaucoma (such as uveitis or neovascular) and a minimum follow-up of at least 24 months. Any contraindications for the use of an anti-VEGF drug, such as uncontrolled blood pressure (systolic>180 mm Hg and/or diastolic>100 mm Hg while a patient is at rest), severe cardiovascular disease (including a stroke or a myocardial infarction 6 months before), and a known allergic reaction to bevacizumab or MMC were the exclusion criteria.

Background history and ocular-related characteristics including baseline IOP (measured by Goldmann applanation tonometry in the last visit before surgery), preoperative visual acuity, glaucoma severity (based on visual field mean deviation) and number of ocular hypotensive medication were retrieved. Safety parameters, and follow-up from visits on day 1; at weeks 1 and 4; at months 3, 6, 12 and 24 after trabeculectomy were collected from charts. The number of postoperative IOP-lowering medication and surgical interventions were also examined. Only one eye per patient was included; in case two eyes fit the criteria, the first eye to be operated was included.

Surgical technique

Surgeries were performed under subtenon, retrobulbar or general anaesthesia as per the local legal requirements. The same protocol was performed in all cases in groups 1 and 2 by the same surgeons. At 1 mm anterior from the limbus in the superotemporal quadrant, a corneal 7–0 silk corneal traction suture was placed. A fornix-based conjunctival flap was dissected and diathermy was performed. A scleral flap measuring 4×3 mm was delineated with a 15° blade and was dissected at half thickness with a crescent knife. MMC-soaked sponges (0.4 mg/mL) were applied subconjunctivally and above the scleral flap for 90 s, followed by abundant washing with saline. A corneal paracentesis was made using 15° blade, and an ocular viscoelastic device (1% sodium hyaluronate) was injected into the anterior chamber. A trabeculectomy measuring approximately 0.5 mm2 was performed using a 15° blade and vannas scissors, followed by a peripherical iridectomy. The scleral flap was fixed with two 10–0 nylon suture in each vertex of rectangular flap. The conjunctiva was closed with two 10–0 nylon suture at the limbus using a ‘purse string’ technique.14 The anterior chamber was sealed by paracentesis hydration after removing the viscoelastic substance. In the MMC+bevacizumab group, intracamerular bevacizumab (50 µL–1.25 mg) was injected through the paracentesis using a 30-gauge needle. A topical preparation containing antibiotic prophylaxis and corticotherapy drops (Tobradex; Novartis, Basel, Switzerland) was prescribed to both groups, postoperatively four times per day, during 1 month and then was slowly tapered.8

Surgical success and failure criteria

Surgical success was defined as IOP of either ≤18, ≤16 or ≤14 mm Hg with at least 30% reduction from baseline at 24 months. Furthermore, a minimum IOP of 6 mm Hg was also required, as per the World Glaucoma Association’s Guidelines on Design and Reporting of Surgical Trials recommendations.15 Absolute success was considered if the conditions were met with no IOP-lowering medication, while success was deemed qualified if medications were needed.15

Failure criteria were defined as IOP above the predefined thresholds or not reduced 30% below baseline, need for a new filtering procedure or loss of vision (loss of two Snellen lines or loss light perception). Suture lysis and needling were not a criteria for surgical failure at any time point.

Primary and secondary outcome measures

The primary outcome was to compare the absolute and qualified surgical success rate of trabeculectomy with MMC or with MMC+bevacizumab at 24 months. The secondary outcomes included safety parameters (intraoperative and postoperative complications).

Statistical analysis

Demographics and clinical characteristics of patients were described using the mean (SD) or median (IQR: 25th–75th percentiles) for quantitative variables and the frequencies (percentages) for categorical variables. To compare IOP values between baseline and 24-month visits, Wilcoxon signed-rank test was used. Group comparisons of quantitative variables were performed using non-parametric Mann-Whitney test, and χ2 test or Fischer’s exact test was used for categorical variables. Kaplan-Meier estimator was used to assess the survival of each group treatment and results were further compared by log-rank test. Univariable and multivariable Cox regression models were applied to study time until failures related to IOP target (≤18, ≤16 and ≤14 mm Hg with at least 30% reduction from baseline) considering a follow-up of 24 months. The level of significance α=0.05 was considered. Data analysis was performed using STATA V.15 and GraphPad Prism V.8 (San Diego, California, USA).


Baseline characteristics

A total of 110 eyes were submitted to trabeculectomy with MMC. In 51 (46%) of these, intracamerular bevacizumab was injected at the end of the procedure. Overall, 88 eyes (80%) reached the 24-month time point with completed follow-up (42 eyes in the MMC group and 46 in the MMC+bevacizumab group). Reasons for exclusion from the analysis and number of patients at each time point are depicted in the flowchart (figure 1). Demographic data and overall baseline characteristics of the study patients are listed in table 1, with both groups being similar at baseline in all variables except in visual acuity (p<0.001). Surgeries were similarly performed in two centres by the two surgeons (LAP, 60 procedures, and ADDDOB 50, procedures). Most of the procedures were under subtenon anaesthesia (88% in the MMC group and 93% in the MMC+bevacizumab group). The remaining were under general anaesthesia in cases of patient request or poor collaboration.

Figure 1

Flowchart of patient progress in the 24-month follow-up study. MMC, mitomycin C.

Table 1

Baseline characteristics of participating patients


Absolute and qualified success rates are presented in table 2. Log-rank tests identified significant differences in the three success criteria of IOP when we considered absolute success rates (IOP≤14 mm Hg: 44.8% in the MMC group vs 66.7% in the MMC+bevacizumab group, p=0.010; IOP≤16 mm Hg: 58.4% in the MMC group vs 76.1% in the MMC+bevacizumab group, p=0.039; IOP≤18 mm Hg: 54.1% in the MMC group vs 78.2% in the MMC+bevacizumab group, p=0.007). Related to qualified success criteria, significant differences between Kaplan-Meier survival curves were found when IOP≤14 mm Hg (46.6% in the MMC group and 70.6% in the MMC+bevacizumab group, p=0.007). From a statistical point of view, this difference became weaker when comparing less strict thresholds targets (IOP target≤16 mm Hg: 63.4% in the MMC group vs 80.1% in the MMC+bevacizumab group, p=0.054; IOP target≤18 mm Hg: 67.8% in the MMC group vs 84.2% in the MMC+bevacizumab group, p=0.052). Figure 2 depicts Kaplan-Meier curve estimates for the three success criteria and corresponding information about failure events, respectively.

Figure 2

(A) Kaplan-Meier survival curve estimates (IOP≤18mmHg): qualified and absolute* success. (B) Kaplan-Meier survival curve estimates (IOP≤16 mmHg): qualified and absolute* success. (C) Kaplan-Meier survival curve estimates (IOP≤14 mmHg): qualified* and absolute* success. *P<0.05. IOP, intraocular pressure; MMC, mitomycin C.

Table 2

Comparison of success rates in both groups

In the MMC group, five patients (8%) at the 12-month visit and two more patients (12%) at the 24-month visit needed further glaucoma surgery (drainage implants). There was no failure related to visual acuity (no one had loss of two Snellen lines or loss light perception).

Only 10 patients were taking IOP-lowering medication at 24 months’ visit, 7 of which were patients in the MMC group.

Results of the univariable statistical models (table 3) showed that, in the more lenient criteria (IOP≤18 mm Hg), treatment intervention, age, gender and preoperative IOP were candidates (p<0.25) to the multivariable model. In the outcome IOP of ≤16 mm Hg, this list was shortened to only the treatment intervention. Finally, with the IOP target of ≤14 mm Hg, the univariable analysis identified treatment intervention (ie, addition of bevacizumab) to be associated with a more positive outcome (p=0.011). No multivariable model was obtained for any of the success criteria.

Table 3

Univariable analysis of success for each criteria (p values)

IOP lowering

IOP was effectively reduced in both groups at the 24-month visit compared with baseline (MMC: 24.4 (8.0) mm Hg vs 12.1 (5.3) mm Hg, MMC+bevacizumab: 25.1 (8.7) mm Hg vs 10.8 (3.8) mm Hg; p<0.001 in both comparisons), with no significant difference between the groups at the final time point (p=0.314). However, there was a period during the early postoperative period (between 1-week and 1-month postoperative), where the bevacizumab-augmented group showed significant lower IOP values (p=0.015 and p=0.003, respectively) (figure 3).

Figure 3

IOP and IOP-lowering medication throughout the study visits. *P<0.05. IOP, intraocular pressure; Med, number of preoperative topical medication; MMC, mitomycin C; P25, 25th percentile; P75, 75th percentile.

Of note, nearly half of the MMC+bevacizumab group had IOP in the single-digit range (≤9 mm Hg) at the 12-month visit and 24-month visit (40.1% and 41.3%, respectively). At 12 months, nearly the entire MMC+bevacizumab group (98%, n=48) was drop-free, with a similar trend of drop-free patients at the 24-month visit (93%, n=43).

Surgical complications and postoperative interventions

No significant differences in rates of intraoperative or early postoperative complications were detected (table 4). Conversely, as expected with a higher IOP value in the early postoperative period, a higher rate of needlings was performed in the MMC group (19% vs 2%, p=0.005). No significant differences in cataract requiring surgery were found between the two groups (p=0.084). Additional glaucoma surgery was needed in seven patients in the MMC group, while no patient of the MMC+bevacizumab group required further interventions (p=0.014). Following the same trend as in preoperative visual acuity logMAR analysis (median 0.4 (0.1–1.0) in the MMC group and 0.1 (0.1–0.3) in the MMC+bevacizumab group, p<0.001), we found a significant difference in the last follow-up visual acuity logMAR evaluation (median 0.3 (0.1–1.0) in the MMC group and 0.1 (0.0–1.0) in the MMC+bevacizumab group, p<0.001). No other differences were detected during the postoperative management (table 5).

Table 4

Intraoperative and postoperative complications during follow-up

Table 5

Postoperative interventions


Our study describes a 24-month follow-up of a two-centre study comparing primary trabeculectomy with MMC with and without intraoperative bevacizumab intracamerular injection. Our data suggest a greater effectiveness of bevacizumab as an adjunct to MMC on improving the trabeculectomy outcomes in patients, particularly if the target IOP is in the low teens. Our 2-year data appear to confirm the 12-month data from an earlier prospective study using a similar methodology, where adding bevacizumab allowed for a less burdened postoperative care (ie, less needlings).10 Besides, our study confirms a ower number of antiglaucoma medication using bevacizumab as suggested in the diabetic population submitted to trabeculectomy with MMC subconjunctival and bevacizumab.16

Interestingly, while a higher success was detected throughout several success criteria, the added value of the bevacizumab effect only became clinically significant when a more ambitious criteria (IOP≤14 mm Hg) was used. This could partially explain the non-consensual data on this subject. Current literature does suggest that an MMC–trabeculectomy alone may be enough to achieve a high success rate if mid to high teens are needed.11 Accordingly, the added value of additional strategies would be only detected on the fringes of this IOP range (ie, in the ability to either reach or sustain a low-target IOP). Indeed, trabeculectomy is able to provide low-target IOPs, but usually not very low and often with a steady decline. Multicentric studies using modern MMC protocols place the range of success in the 60%–70% at 2 years, which is below our current study.17 Indeed, nearly 50% of the patients in our study had pressures below 10 mm Hg with bevacizumab+MMC at 24 months. This may explain the positive results of the only large prospective randomised trial on this subject. In Vandewalle et al’s study, a significant proportion of patients had normal-tension glaucoma and hence had a low baseline IOP (and accordingly a very low target IOP).10 Other studies with higher mean baseline IOP would have achieved their intended 30% reduction criteria despite being in the mid-teens/to high-teens range. In this range, the use of MMC alone would have been enough to achieve success.11 18

One rationale behind the benefits of this association is that MMC upregulates proinflammatory, proangiogenic and proibrotic growth factors such as VEGF in the aqueous humour.19 An injection of anti-VEGF has been associated with blocking these MMC effects, thus with potential to less scarring bleb and less probability of surgery failure. The complementary effect of bevacizumab and MMC on bleb survival has been demonstrated in a mouse model for filtering surgery.19 Bevacizumab has been responsible for decreasing not only early postoperative angiogenesis (by the reduction in blood vessel density around the bleb area on day 14) but also late fibrosis (collagen deposition around 1 month after trabeculectomy).6 19 These observations are consistent with the significant difference between two groups in IOP reduction at 1 week and 1 month postoperative, with lower levels in the anti-VEGF group. Histological studies revealed that subconjunctival fibroblasts responsible for the bleb failure have the maximum proliferation between days 3 and 5 postoperatively with excessive scarring and subsequent encapsulation and bleb failure occurring after 2–6 weeks after surgery.8 20–23 A single bevacizumab injection is thought to be able to provide half-maximum inhibitory concentrations for up to 8 weeks, which could provide the scientific substrate to a larger bleb area, decrease in blood vessel density and collagen deposition, potentially with long-term implications (our results suggesting up to 24 months).

Several administration routes are available to use bevacizumab, and the most appropriate is still uncertain.24 The subconjunctival route was first used in clinical trials.8 9 It is thought to be less invasive and can be applied directly to the target site, with Nomoto et al suggesting that its longer half-time to be related to the scleral binding mechanism compared with intravitreal injection.21 23 However, it has the disadvantage of stimulating inflammation and fibroblast activation by subconjunctival manipulation.10 19 Intravitreal injection, despite the ‘depot effect’ of the drug in the vitreous cavity, can be associated with retinal complications (ie, retinal detachment), and bevacizumab has higher levels in plasma with more systemic absorption.10 19 The route used in our study was chosen over all alternatives not only because no additional incision is needed and it does not disturb conjunctiva integrity but also because VEGF will flow into the bleb, having its effect and decreasing the high levels in the anterior chamber of patients with glaucoma.6 10

Another advantage of adding bevacizumab was that 93% of the eyes no longer require use of topical medication (24 months’ visit). Saeed and AboulNasr confirmed the need of fewer antiglaucoma drugs at all postoperative visits when bevacizumab augmented MMC.25 This factor is relevant as it avoids issues associated with compliance, ocular surface changes, poor tolerance, and ophthalmic and systemic adverse effects. Importantly to clinical translation into a wider audience, this wound modulation strategy (MMC+bevacizumab) did not change the safety profile of the standard of care. Our results not only did not show an increase in intraoperative or postoperative complications, which appears consensual from the literature, but also had an acceptable safety profile.

The current study has some drawbacks. First, our findings are based on retrospective data from two centres and surgeries were performed by the same team. While this has the advantage of bringing real-world data into discussion, this may have generated an unintended bias such as patients allocated to the treatment group according to the hospital they belonged (no randomised and no masked treatments) and surgeries performed by only two surgeons (no masked). Furthermore, the retrospective nature of the study and the lack of prospective adverse event reporting could have underestimated a negative side effect in the bevacizumab group. Second, while our results suggest that a single bevacizumab intracamerular injection can improve surgical outcomes, the optimised dose, for example, is not yet stablished. Third, the study findings in our mostly low-risk population sample (mostly Caucasian and with no prior surgery) may not be applicable for higher-risk patients (such as uveitic or neovascular glaucomas). Finally, we have a high number of suture lysis treatments, which can be a particularity of our surgical approach. However, the standard MMC arm (which relates to a 90 s exposure) has overlapping results with most of the trabeculectomy literature, which suggests this strategy may not affect the outcome in this time frame.17

To conclude, intraoperative intracamerular bevacizumab enhances the effects of MMC in primary trabeculectomy, particularly when a very low IOP target is considered, with no influence on the safety profile.

Data availability statement

Data are available in a public, open access repository.

Ethics statements

Patient consent for publication

Ethics approval

This study adhered to tenets of the Declaration of Helsinki and was approved by the ethics committee from each centre in December 2019. All patients provided written informed consent for data retrieval as per General Data Protection Regulation.



  • Contributors Conception and design of the work: PJ, JPS and LAP. Acquisiton, analysis and interpretation of data: PJ, FJT, RB, DCS, REM, MA, ALP and LAP. Revision of the work: PJ, DCS, REM, ADDOB, CM-N, MA, ALP, IS and LAP. Final approval of the version to be published: PJ, FJT, RB, DCS, REM, ADDOB, CM-N, MA, ALP, IS, JPS and LAP.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • Highlights from this issue
    Frank Larkin