To evaluate the efficacy and safety of intravitreal bevacizumab (IVB) in eyes with neovascular glaucoma (NVG) undergoing Ahmed glaucoma valve (AGV) implantation.
To evaluate the efficacy and safety of intravitreal bevacizumab (IVB) in eyes with neovascular glaucoma (NVG) undergoing Ahmed glaucoma valve (AGV) implantation.
This was a multicentre, prospective, randomized clinical trial that enrolled 40 patients with uncontrolled neovascular glaucoma that had undergone panretinal photocoagulation and required glaucoma drainage device implantation. Patients were randomized to receive IVB (1.25 mg) or not during Ahmed valve implant surgery. Injections were administered intra-operatively, and 4 and 8 weeks after surgery.
After a mean follow-up of 2.25 ± 0.67 years (range 1.5–3 years), both groups showed a significant decrease in IOP (p < 0.05). There was no difference in IOP between groups except at the 18-month interval, when IOP in IVB group was significantly lower (14.57 ± 1.72 mmHg vs. 18.37 ± 1.06 mmHg – p = 0.0002). There was no difference in survival success rates between groups. At 24 months, there was a trend to patients treated with IVB using less antiglaucoma medications than the control group (p = 0.0648). Complete regression of rubeosis iridis was significantly more frequent in the IVB group (80%) than in the control group (25%) (p = 0.0015).
Intravitreal bevacizumab may lead to regression of new vessels both in the iris and in the anterior chamber angle in patients with neovascular glaucoma undergoing Ahmed glaucoma valve implantation. There is a trend to slightly lower IOPs and number of medications with IVB use during AGV implantation for neovascular glaucoma.
Neovascular glaucoma (NVG) is a secondary glaucoma caused by abnormal fibrovascular tissue growth on the iris and drainage angle structures, including the trabecular meshwork (Sivak-Callcott et al. 2001; Hayreh 2007). Contraction of this tissue results in progressive angle closure and intraocular pressure (IOP) increase, eventually leading to a glaucoma that is poorly responsive to conventional treat-ment and has poor visual prognosis (Allingham et al. 2005). NVG may be caused by a number of ocular and systemic conditions, all of them leading to retinal ischaemia and subsequent release of angiogenic factors (Allingham et al. 2005; Hayreh 2007). It is now evident that several mediators are involved in pathologic neovascularization, the most important and well studied of which is the vascular endothelial growth factor type A (VEGF-A) (Casey & Li 1997; Tripathi et al. 1998; Avery et al. 2006). Bevacizumab (Avastin, Genentech, South San Francisco, CA, USA) is an anti-VEGF agent that has been used for the treatment of patients with NVG. Intravitreal bevacizumab (IVB) injection has been reported to reduce iris rubeosis and IOP (Costagliola et al. 2008; Eid et al. 2009; Moraczewski et al. 2009; Yazdani et al. 2009; Beutel et al. 2010).
Currently, the management of this condition is directed towards the treatment of retinal ischaemia, mostly by some form of retinal ablation to reduce the neovascular stimuli, and IOP reduction by means of medical and surgical therapy (Sivak-Callcott et al. 2001; Allingham et al. 2005; Hayreh 2007). Despite the reduction in retinal ischaemia and the use of antiglaucoma medication, NVG eyes frequently need to undergo surgery.
Considering the poor surgical success in the treatment of IOP in eyes with NVG (Sivak-Callcott et al. 2001; Yalvac et al. 2007), and the possibility that an anti-VEGF agent could be useful improving the success rate of surgical treatment of NVG, we conducted a study to evaluate the safety and efficacy of IVB in eyes with NVG undergoing Ahmed glaucoma valve (AGV) (New World Medical Inc, Rancho Cucamonga, CA, USA) implantation.
This was a multicentre, prospective, randomized clinical trial that enrolled 40 patients from 3 tertiary referral eye centres. Consecutive patients with NVG were screened and evaluated for eligibility. All patients with uncontrolled NVG who had undergone panretinal photocoagulation (PRP) and required glaucoma drainage device implantation were eligible.
The study was approved by the Institutional Review Board of University of Campinas (UNICAMP). Info-rmed consent was obtained from each patient. All study procedures adhered to the Declaration of Helsinki. The study was registered at Australian New Zealand Clinical Trial Registry (ANZCTR) – ACTRN12607000577415.
All patients older than 18 years, with uncontrolled NVG, followed at the Glaucoma Service of University of Campinas (UNICAMP), University of São Paulo – Ribeirão Preto (USP) and Federal University of Uberlândia (UFU), were candidates for inclusion in the study. All patients had to have undergone a PRP at least 2 weeks before enrolment. Uncontrolled NVG was defined as an eye with IOP above 22 mm Hg using maximum tolerated glaucoma medication. Patients were excluded from the study if any of the following criteria were present: (1) no light perception; (2) NVG secondary to intraocular tumours or uveitis; (3) unwilling or unable to return for follow-up; (4) pregnancy; (5) learning difficulties, mental illness or dementia; or (6) previous cyclodestructive procedure, scleral buc-kle procedure or silicone oil surgery.
Eligible patients with NVG were randomized to the following groups using a computer-generated randomization table: (1) study group – Ahmed valve implant with IVB and (2) control group – Ahmed valve implant without IVB. Surgeons were told to inject or withhold the IVB at the end of the surgical procedure. Ophthalmologists responsible for the patients' follow-up were masked to the use of IVB.
The surgical procedure consisted of a 1-stage AGV implantation using a standardized surgical technique. After administration of regional or general anaesthesia, a fornix-based conjunctival flap was fashioned in the superotemporal quadrant. The anterior edge of the plate was secured with 9-0 nylon sutures to the sclera at least 8 mm from the limbus. The tube tip was cut obliquely to protect the tube lumen from the iris. A 23-gauge needle tract was used to enter the anterior chamber 1 mm posterior to the limbus. The tube in the anterior chamber was positioned anterior to the iris and away from the corneal endothelium. A donor sclera graft was fashioned from banked tissue and secured with interrupted 10-0 nylon sutures over the exposed portion of the tube. Conjunctiva was sutured either with 8-0 polyglactin or 10-0 nylon sutures. The anterior chamber was reformed with BSS thro-ugh a paracentesis tract. Viscoelastic material was injected into the anterior chamber at the surgeon's discretion. In the IVB group, at the end of surgery, 1.25 mg of bevacizumab (0.05 ml of commercially available bevacizumab at a concentration of 25 mg/ml) was injected into the eye via pars plana (3.5 mm away from the limbus) with a 1.00-ml syringe attached to a 30-gauge needle. Injections were administered intra-operatively, and 4 and 8 weeks after surgery. The control group did not receive an intravitreal injection.
The post-operative regimen included topical antibiotic and cycloplegic for 2 to 4 weeks and topical steroids for approximately 2 months. Patients were examined 1 day, 1 week, 2 weeks, 1, 3, 6, 12, 18 and 24 months after surgery. Antiglaucoma medication was added as required to improve IOP control. The following preoperative variables were evaluated: patient's age, gender and race; lens status; gonioscopy; slit lamp biomicroscopy; previous surgical procedures; number of antiglaucoma medications; IOP; and best-corrected visual acuity (VA). Preoperatively and at all post-operative intervals, the presence of rubeosis iridis was evaluated at the slit lamp. Furthermore, the occurrence of new vessels (NV) and/or goniosynechiae at the anterior chamber angle were measured in clock hours with gonioscopy using a Posner lens. At each follow-up visit, the following variables were analysed: VA, IOP, gonioscopy, biomicroscopy, number of antiglaucoma medications and the presence of any complication. A decre-ase or an increase in VA > 2 Snellen lines or a change in category (i.e. count-ing fingers to hand movements) from the baseline preoperative value was considered to be a clinically significant change in VA.
Success was defined by the following criteria: (1) post-operative IOP level between 6 and 21 mmHg, with or without antiglaucoma medications, (2) IOP reduction of at least 30% relative to preoperative values. Eyes requiring addi-tional glaucoma surgery that developed phthisis or that showed loss of light perception were classified as failures.
All data were captured on standardized data sheets and later transferred to CDC EpiInfo™ ver. 3.5.1 software for analysis (Centers for Disease Control and Prevention, Atlanta, GA, USA).Categorical variables were analysed using the Fisher's exact test or the chi-square test. We used the Bartlett's test to check the homogeneity of variances. Continuous variables were analysed using anova when the values were normally distributed. When values were not normally distributed and the variances were not homogeneous, we used the Kruskal–Wallis test, followed by the Wilcoxon two-sample test if statistically significant differences were found. Within-group changes from baseline were analysed using the paired t-test. Success rates in both groups were compared using Kaplan–Meier survival curves and the log-rank test. p Values of < 0.05 were considered statistically significant. With a sample size of 35 and an alpha error of 0.05, assuming a standard deviation of the IOP measurement of 20%, the power of the study was 0.8 to detect a 20% difference between the groups.
Forty patients (40 eyes) were included in the study. Randomization assigned 20 patients to the study group (Ahmed valve implant with IVB) and 20 to the control group (Ahmed valve implant without IVB).
Table 1 shows the baseline characteristics of both groups. No significant differences were found between the groups regarding clinical and demographic features. All patients showed rubeosis iridis at the baseline slit lamp examination.
|IVB (n = 20)||Control (n = 20)||p|
|Age (years)||59.25 ± 8.05||62.40 ± 11.78||0.3299|
|Follow-up (years)||2.15 ± 0.67||2.35 ± 0.67||0.3517|
The baseline and follow-up IOPs for both groups are listed in Table 2. Baseline IOPs did not differ between groups (p = 0.6454). There was no sign-ificant difference in mean IOP between groups until 12 months of follow-up (p > 0.2497). However, the study group had a significantly lower mean IOP than the control group at the 18th month post-operative follow-up visit (p =0.0002). After a mean follow-up of 2.25 ± 0.67 years (range 1.5–3 years), both groups showed a similar decrease in IOP (50.7% in the bevacizumab group and 61.5% in the control group, p = 0.7750).
|Baseline||40.10 ± 13.33 (n = 20)||38.35 ± 10.34 (n = 20)||0.6454|
|1 Day||10.68 ± 5.74 (n = 20)||10.85 ± 6.74 (n = 20)||0.9348|
|7 Days||10.35 ± 4.76 (n = 20)||11.45 ± 5.77 (n = 20)||0.5148|
|15 Days||14.00 ± 6.13 (n = 20)||16.50 ± 7.34 (n = 20)||0.2498|
|1 Month||17.45 ± 4.65 (n = 20)||19.05 ± 6.16 (n = 20)||0.3597|
|3 Months||18.30 ± 6.55 (n = 18)||18.33 ± 5.44 (n = 17)||0.9866|
|6 Months||16.78 ± 7.47 (n = 16)||16.33 ± 4.35 (n = 17)||0.3827|
|9 Months||18.31 ± 8.93 (n = 16)||16.17 ± 4.60 (n = 16)||0.8898|
|12 Months||17.40 ± 9.99 (n = 15)||16.00 ± 3.98 (n = 15)||0.4598|
|18 Months||14.57 ± 1.72 (n = 15)||18.37 ± 1.06 (n = 14)||0.0002b|
|24 Months||14.43 ± 0.53 (n = 14)||16.67 ± 4.40 (n = 12)||0.0526|
At baseline, the mean number of medications was 2.85 ± 1.18 in the IVB group and 2.80 ± 0.69 in the control group (p = 0.3912). Compared with baseline, both groups showed a statistically significant decrease in the number of antiglaucoma medications during all follow-up (p < 0.001) (Table 3). Until 18 months of follow-up, there was no statistically significant difference in the mean number of medications between the 2 groups (p > 0.116). Nevertheless, there was a trend for patients treated with bevacizumab using less medication than the control group 24 months after surgery (p = 0.0648).
|Baseline||2.85 ± 1.18||2.80 ± 0.69||0.3912|
|1 Month||0.55 ± 0.89||0.70 ± 0.73||0.5633|
|3 Months||0.80 ± 0.89||1.26 ± 0.93||0.1221|
|6 Months||1.44 ± 1.25||1.17 ± 0.71||0.5117|
|9 Months||1.23 ± 1.30||1.28 ± 0.67||0.4212|
|12 Months||1.21 ± 1.12||1.18 ± 0.73||0.9106|
|18 Months||1.14 ± 0.69||1.67 ± 0.65||0.1162|
|24 Months||1.14 ± 0.69||1.75 ± 0.62||0.0648|
Post-operatively, there were no significant differences in visual acuity (LogMAR) (p > 0.1270) and extension of goniosynechiae (p > 0.6644) between the groups. However, at the end of follow-up, the decrease in the extension of new vessels in the anterior chamber angle in the bevacizumab group was significantly greater than in the control group (p = 0.0017). Although both groups showed a significant decrease in the extension of rubeosis iridis at the end of follow-up (p < 0.05), complete regression of rubeosis iridis was significantly more frequent in the bevacizumab group (80.0%) than in the control group (25%) (p = 0.0015).
No major intraoperative complication was reported in either group, and no statistically significant difference was found between them (Table 4). The most frequently occurring complication was hyphema.
|IVB (n = 20)||Control (n = 20)||p|
|Flat anterior chamber||2||1||1.0000|
At the end of follow-up, 13 (65%) patients in the IVB group and 12 (60%) in the control group had post-operative IOP levels between 6 and 21 mmHg, with or without antiglaucoma medications. Sixteen (80%) patients in the IVB group and 15 (75%) in the control group showed IOP reduction of at least 30% relative to preoperative values.
Kaplan–Meier survival analysis sho-wed that there was no significant difference in success rates between the two groups according to both criteria: (1) post-operative IOP level between 6 and 21 mmHg, with or without antiglaucoma medications (Fig. 1– p = 0.2838 by log-rank test), (2) IOP reduction of at least 30% relative to preoperative values (p = 0.3012 by log-rank test).
The management of NVG remains a challenge. The options include PRP, ocular antihypertensive medications, glaucoma drainage surgery and cyclodestructive procedures. However, adequate IOP control is sometimes difficult to be achieved by any of these means (Casson 2006).
Kjeka et al. (2013) evaluated the effects of early PRP in 36 patients with ischaemic central retinal vein occlusion (CRVO): one group (n = 18) was assigned to standard treatment that included regular examinations and PRP as soon as NV was found, whereas the other group received early PRP, before the development of NV. After a mean follow-up of 5 months, 12 patients in the standard treatment group developed neovascular glaucoma. In the early treatment group, only one patient developed subtle iris rubeosis 7 months after PRP. None of the remaining patients showed any signs of ocular NV or increased IOP during a mean follow-up of 41 months, suggesting that the development of ocular neovascularization can be prevented by early PRP.
Increased levels of VEGF-A have been identified in the aqueous humour of patients with rubeosis and NVG (Tripathi et al. 1998). Sone et al. 1996 found higher levels of VEGF in the aqueous and vitreous humour of patients with diabetic NVG compared with patients showing only proliferative diabetic retinopathy. These findings substantiate the importance of VEGF in the development of iris and angle NV (Sivak-Callcott et al. 2001).
The IOP-reducing effect of bevacizumab injection in eyes with NVG has been demonstrated in some case reports (Vatavuk et al. 2007; Yalvac et al. 2007; Batioglu et al. 2008; Martínez-Carpio et al. 2008; Moraczewski et al. 2009; Yazdani et al. 2009). The efficacy of IVB to improve the IOP control in patients with NVG depends on structural lesions of the trabecular meshwork and the extension of the fibrovascular membrane and goniosynechiae on the iridocorneal angle. IVB may eliminate the new vessels, but is not capable of extinguishing the synechia (Wakabayashi et al. 2008). In eyes with NVG and an open angle, IVB could be efficient in more than 70% of the cases, avoiding many cases of surgery. However, when the angle is closed, more than 90% of the cases need surgery (Wakabayashi et al. 2008). Ehlers et al. (2008) also confirmed the previous investigations comparing the advantages of PRP associated with IVB and PRP alone. After a 4-month follow-up, the neovascular regression was greater, and earlier, in the group treated with both methods, findings were confirmed in our series.
Sothornwit (2008) reported the short-term efficacy and safety of IVB injection in conjunction with implantation of AGV in a patient with refractory NVG caused by proliferative diabetic retinopathy. At 48 hrs post-operatively, IOP was markedly decre-ased and iris neovascularization was not visible. At 6-week follow-up, the IOP was 8 mmHg. At 8 weeks post-operatively, there was small recurrent iris neovascularization without any rise in IOP.
Most of the comparative studies investigating the efficacy of IVB injections in conjunction with AGV implantation suggest that lower IOPs are achieved with the injections, although long-term success rates may not differ. Zhou et al. (2013) evaluated the efficacy and safety of adjunctive IVB in conjunction with AGV implantation in the management of NVG in 53 eyes (IVB group, n = 25 eyes). The mean follow-up was 15 ± 5 months. Iris neo-vascularization had completely regre-ssed within 2–8 days after IVB in 22 eyes (88.0%). The success rates were 84.0% (IVB group) and 75.0% (control group) at 12 months and 84.0% (IVB group) and 71.4% (control group) at 18 months (p = 0.344, log-rank test). Significantly less antiglaucoma medication was required after surgery in the IVB group than in the control group at all follow-up time-points.
Another study to assess the efficacy of preoperative IVB before AGV implantation in the treatment of NVG was published by Sevim et al. (2013). The authors retrospectively compared the results of 19 eyes (19 patients) that underwent AGV implantation and had received IVB 7–14 days before surgery to a control group of 22 patients who did not receive IVB prior to AGV implantation. The success rate in the study group (79%) was better than in the control group (64%), but this difference was not statistically significant (p = 0.28).
Although IVB injection before Ahmed valve implantation showed a beneficial effect (Ehlers et al. 2008; Sothornwit 2008; Sevim et al. 2013; Zhou et al. 2013), IVB injections during valve impl-antation showed no statistical effect on the success rate in a study conducted by Yang et al. (2012). According to them (Yang et al. 2012), although the success rate in the IVB group (70.0%) was higher than that in the control group (62.5%) 1 year after the procedure, the differences were not statistically significant (p = 0.828, log-rank test). Notwithstanding, the mean IOPs in the IVB group were significantly lower than those of the control group at 12 and 15 months (p = 0.03).
Mahdy et al. (2012) evaluated the efficacy of bevacizumab with PRP followed by AGV implantation in controlling the IOP in cases of NVG. Twenty eyes underwent AGV implantation with IVB (1.25 mg in 0.5 ml) and PRP, and 20 eyes were managed by AGV and PRP only (control group). After a mean follow-up of 18 months, a highly significant decrease in IOP occurred in both groups (p < 0.001), but the IOP was significantly higher IOP in the control group in comparison with the IVB group at 18 months (p < 0.01). However, different from our study and the previous reports, a 95% success rate was achieved in the IVB group, compared with a 50% success rate in the control group (p < 0.05).
Kang et al. (2013) evaluated the effect of IVB before AGV implantation for the treatment of NVG in a retrospective, comparative, consecutive series of 27 eyes of 26 patients. Although IVB provided better visual outcome in the early post-operative period and reduced the incidence of hyphema, it did not significantly improve mean IOP, number of antiglaucoma medications, or success rate (p > 0.05). Recently, Simha et al. (2013) attempted to perform a meta-analysis to evaluate the efficacy and safety of intraocular anti-VEGF agents as an adjunct to existing modalities for the treatment of NVG. Unfortunately, as no randomized controlled trial met their inclusion criteria, no assessment of risk or meta-analysis was undertaken.
According to our findings, IVB may lead to regression of new vessels both in the iris and in the anterior chamber angle in patients with NVG undergoing AGV implantation. Although these findings were not associated with better IOP control after a short-term follow-up, at long-term follow-up patients who received IVB showed lower IOPs and tended to need less additional medication to achieve IOP control.