Intravitreal Injection of Bevacizumab in Primary Vitrectomy to Decrease the Rate of Retinal Redetachment: A Randomized Pilot Study

Abstract

Purpose:

To evaluate the effect of intravitreal bevacizumab (IVB) as a surgical adjunct in prevention of proliferative vitreoretinopathy (PVR) after retinal detachment surgery.

Methods:

In this controlled, randomized pilot study, 27 patients with primary retinal detachment undergoing pars plana deep vitrectomy were included. Of these, 12 received IVB at the end of procedure. The anatomic success and best corrected visual acuity (BCVA) were compared to the control group at months 3 and 6 postoperatively.

Results:

At three month follow-up, 3 of 11 eyes (27.3%) had detached retinas in the IVB group versus 6 of 12 (50.0%) in the control group (P = 0.40). At six-month follow-up, 3 of 10 eyes (30%) had detached retinas in the IVB group versus 3 in 8 (37.5%) in the control group (P > 0.99). Mean logMAR BCVA improved significantly in both groups relative to baseline, but did not show a significant difference at three-and six-month follow-ups between the two groups.

Conclusion:

Our preliminary results show neither a benefit nor any harm from intervention in both anatomic and visual outcomes. Our results support conducting additional studies to evaluate the effect of intravitreal bevacizumab on postoperative PVR.

INTRODUCTION

Rhegmatogenous retinal detachment is one of the most devastating sight-threatening conditions of posterior segment of the eye. Different treatment modalities like scleral buckling or pars plana vitrectomy are available with variable success rates. The most prominent process resulting in anatomic failure after initial successful detachment surgery is proliferative vitreoretinopathy (PVR), defined as the growth and contraction of cellular membranes within the vitreous cavity and on both retinal surfaces.[1] Despite improvements in surgical techniques and the use of intraocular tamponades, the rate of postoperative redetachment due to PVR is still 5-10%.[2] This process is influenced by different predisposing factors such as the grade of initial PVR, size and number of retinal breaks, extent of detachment, lens status, age, vitreous hemorrhage, and breakdown of blood retinal barrier.[3,4,5] Different adjunctive agents like 5-FU and heparin,[6] daunomycin,[7] colchicine,[8] and retinoids [9] have been tried intra- or post-operatively to prevent development of PVR with limited success, and the best way to prevent redetachment after an initial success is yet to be determined.

Vascular endothelial growth factor (VEGF) is one of the most important molecules in intraocular proliferative processes, and the possible role of its inhibition has been investigated extensively. Nowadays, these treatments are the first choice in the management of age-related macular degeneration [10] and play a major role in the management of diabetic macular edema and diabetic retinopathy.[11,12] Although primarily known as an important agent in vascular proliferative processes, studies have implied a role for VEGF in avascular processes like scar formation [13,14] as well. There are also reports of increased levels of VEGF in eyes with retinal detachment and the presence of higher levels of VEGF in eyes with higher grades of PVR.[15,16] However, reports about the relation of VEGF levels and redetachment rates are contradictory.[17,18,19] Given the favorable results of the use of anti-VEGF agents in intraocular proliferative processes and the dilemma about an adjunct to surgical repair of retinal detachment to prevent PVR, this trial was designed to find if it is useful to add an anti-VEGF agent to the surgical protocol. Herein, the primary results are reported.

METHODS

This prospective controlled, randomized pilot study was performed on patients recruited from September 2010. Results of cases eligible for report up to September 2012 are presented. The study adhered to tenets of the Declaration of Helsinki, and methods were approved by the Institutional Review Board of the Ophthalmic Research Center of Shahid Beheshti University of Medical Sciences, Tehran, Iran. Twenty seven eyes of 27 patients were primarily enrolled. Four were excluded as complete pre- or post-operative data were not available or patients were lost to follow-up. Inclusion criteria were the presence of primary rhegmatogenous RD with PVR no worse than grade B [20] and candidacy for pars plana deep vitrectomy on the discretion of the treating surgeon. Exclusion criteria were age <18 years, history of any kind of retinal detachment surgery in the enrolled eye, history of trauma or uveitis, or the presence of diabetic retinopathy, bleeding diathesis, hepatic or renal failure, age-related macular degeneration, giant retinal tear, or macular hole.

After enrollment, the following data were recorded: Patients' demographic data, lens status, duration of symptoms, extent and location of detached area, and best corrected visual acuity (BCVA, log MAR scale). Before surgery, each patient was given a sealed envelope containing a random assignment to a specific group. These assignments were generated randomly on the basis of a computer-generated list prepared by our statistician. Patients were prospectively divided into two groups: Intravitreal bevacizumab (IVB) group and control group.

All patients in both groups underwent a standard 3-port 20-gauge vitrectomy. Sampling of undiluted vitreous was performed by vitrectomy probe before opening the inflow port; volumes of the samples ranged from 0.5-1 ml and were immediately frozen and stored at −80°C. 360° vitreous base shaving under scleral depression was performed. No. 240 encircling band was placed based on decision of the surgeon. Perfluorocarbon liquid (DK-Line, Bausch and Lomb, Kingston, UK) was used in all cases followed by endolaser photocoagulation around breaks. Internal tamponade by air-sulfur hexafluoride (SF6) 20% mixture or silicone oil (Oxane 1300, Bausch and Lomb, Kingston, UK) was used in all eyes. In all surgeries, the retina was flat at the end of the procedure. In the IVB group, intravitreal injection of bevacizumab (1.25 mg, 25 mg/ml; Avastin; Roche Ltd. Mannheim, Germany) was performed at the end of the operation with a 30.. gauge needle through a sutured sclerotomy.

After surgery, additional data were recorded by the surgeon including size, number, location, and shape of breaks and specified intra-operative procedures.

Patients were scheduled for postoperative examinations on first and 7^th days and first, second, third, and sixth months after surgery. In the third and sixth months examinations, BCVA and the condition of the retina were recorded by observers masked to the type of intervention. The primary outcome measure was complete retinal reattachment at follow-up visits without any additional postoperative procedures like intravitreal gas injection or reoperation. The secondary outcome of the study was defined as the change in BCVA from baseline to 3- and 6-month follow-up visits. Eyes which needed re-intervention for retinal redetachment were considered as detached in analysis.

Statistical analysis was performed using SPSS statistical software (version 20, IBM Corp., USA). Snellen visual acuities were transformed into logarithm of minimum angle of resolution. For descriptive purposes, mean ± standard deviation, or median (range) were used. To compare data between groups, Chi-square, Fisher exact test, t-test, and Mann–Whitney test were used. For comparing VA with baseline values within each group, mixed model with Bonferroni method was used. In order to compare the results adjusted for the baseline status, we used analysis of covariance (ANCOVA). P value less than 0.05 was considered statistically significant.

RESULTS

Of the 27 enrolled patients, 23 and 18 subjects had completely recorded postoperative data after three and six months, respectively. Table 1 provides baseline characteristics of and intraoperative procedures for the patients. There were no significant differences in preoperative logMAR visual acuity, age, sex, lens status, and features related to detachment between the two groups. The rate of placement of encircling band and the type of intraocular tamponade was not significantly different in the two groups either. A noteworthy percentage of patients had more than one retinal break, and this percentage was the same in both groups.

Table 1.

Baseline characteristics and surgical procedures

All 23 eyes had an attached retina on postoperative day 1. In three eyes, redetachment evolved within three to five weeks, and these cases underwent successful reoperation. At three-month follow-up, 3 of 11 eyes (27.3%) had detached retinas in the IVB group versus 6 of 12 (50.0%) in the control group (P = 0.40 Fisher exact test). At six-month follow-up, 3 of 10 eyes (30%) had detached retinas in the IVB group versus 3 of 8 (37.5%) in the control group (P > 0.99) (The decrease in number is due to loss to follow-up).

The preoperative mean VA was 2.04 ± 0.61 logMAR in the control group and 1.95 ± 0.79 logMAR in the IVB group. At three months, the mean VA was 1.26 ± 0.71 logMAR and 1.27 ± 0.71 logMAR in the control and IVB groups, respectively (P < 0.001 and 0.012, for comparison to the baseline values respectively). At six months postoperatively, the mean BCVA was 1.16 ± 0.74 logMAR and 0.96 ± 0.52 logMAR in the control and IVB groups, respectively (P = 0.001 and 0.007, for comparison to the baseline values respectively) [Figure 1]. The change in visual acuity from baseline to three and six months was statistically significant in both groups; however the difference in BCVA between the two groups at both time points was not significant [Figure 2]. There was no statistically significant difference between the two groups regarding status of the retina at months 3 and 6 (P = 0.400, P > 0.99, respectively) [Table 2].

Figure 1.

At three months, the mean VA was 1.26 ± 0.71 logMAR and 1.27 ± 0.71 logMAR in the control and IVB groups, respectively (P < 0.001 and 0.012, respectively). At six months postoperatively, the mean VA was 1.16 ± 0.74 logMAR and 0.96 ± 0.52 logMAR in the control and IVB groups, respectively (P = 0.001 and 0.007, respectively). BCVA, best-corrected visual acuity; IVB, intravitreal bevacizumab; logMAR, logarithm of minimum angle resolution

Figure 2.

The change in visual acuity from baseline to three and six months was statistically significant in both groups; however the difference in BCVA between the two groups in both intervals was not significant. BCVA, best-corrected visual acuity; IVB, intravitreal bevacizumab; logMAR, logarithm of minimum angle resolution

Table 2.

Postoperative anatomic and visual outcomes

DISCUSSION

In our preliminary results, we found neither a benefit nor any harm from intervention on both anatomic and visual outcomes; our following study will show if adding intravitreal injection of bevacizumab to surgical protocol is useful or not. Development of proliferative vitreoretinopathy is the main challenge for vitreoretinal surgeons after a primary successful surgery. Although there have been improvements in instruments and surgical techniques and the choice of procedure for reattachment has shifted from buckling towards vitrectomy,[21] the best means to prevent development of PVR still remains unclear. Rasieret al[15] compared VEGF levels of 22 eyes with RRD to 12 with no vitreoretinopathy and found significantly higher levels (P < 0.001) in eyes with RRD. In another study by Citirik et al,[16] the average vitreous levels of VEGF was 15.14 ± 5.22 pg/ml in eyes with grade B PVR compared to 99.15 ± 38.58 pg/ml in eyes with grade C PVR. Ricker et al[17] reported that the VEGF level of subretinal fluid in eyes that had successful buckling surgery was significantly lower than those that developed postoperative PVR. Although bevacizumab is a well-known agent in aborting vasoproliferative processes like PDR, reports like studies by Memarzadeh et al[22] and Pennock et al[23] show effectiveness of this molecule in slowing of an avascular process like scar formation or PVR in rabbits. These reports provide a rationale for the current study. To the best of our knowledge, no reports of clinical trials regarding advantages or disadvantages of the use of bevacizumab in the prevention of PVR in humans has been published yet.

The paradox of increased VEGF levels in an avascular process like PVR can be explained by observations by Ricker et al[24] and Perrinet al[25] which showed that antiangiogenic isoforms of VEGF (so-called VEGF_xxxb) account for the majority of the total VEGF. These isoforms are products of splicing in the C-terminal exon and inhibit proliferative and vasodilator effects of VEGF₁₆₅.[26,27] This may be a reason that there was no difference in the anatomic success rate between the two groups.

Based on the observed redetachment rate of 50% in the control group and 27.3% in the IVB group, our study has only 20% power to detect this difference as statistically significant. Our study revealed that at least 93 samples in each group are needed to have 80% power to detect a 20% improvement in the IVB group in comparison with the control group when type I error is only 5%.

One of the important factors in success after reattachment surgery is the number of retinal breaks, and the high rate of patients with more than one break (57% and 50% in the control and IVB groups, respectively) could be an explanation for the rather high rate of redetachment after initial success.

One limitation of the current study is that patients underwent surgery by different vitreoretinal fellows and senior surgeons, which was not considered in randomization; however, this is a common trend in a referral training center.[28]

In summary, our preliminary results support additional studies to find the effect of intravitreal anti-VEGF agents on the results of retinal detachment surgery.

Financial Support and Sponsorship

Nil.

Conflicts of Interest

There are no conflicts of interest.