Early Vitreous Hemorrhage after Vitrectomy with Preoperative Intravitreal Bevacizumab for Proliferative Diabetic Retinopathy

Abstract

Purpose:

To evaluate the effect of preoperative intravitreal bevacizumab (IVB) on surgical procedures, visual prognosis, and postoperative complications, especially postoperative vitreous hemorrhage, in cases with proliferative diabetic retinopathy (PDR).

Materials and Methods:

Seventy-one eyes of 54 consecutive patients (23 eyes of 18 women, 48 eyes of 36 men) were investigated in this study. Twenty-five eyes received IVB one to 30 days before the vitrectomy (Bevacizumab Group) and the other 46 eyes had the vitrectomy alone (Control Group). The surgical procedures, best-corrected visual acuities at baseline, 1, 3, and 6 months after the vitrectomy, and postoperative complications in the Bevacizumab Group were compared to the Control Group.

Results:

The patients were significantly younger in the Bevacizumab Group compared to the Control Group (P = 0.008). The incidence of preoperative vitreous hemorrhage, tractional retinal detachment, and iris neovascularization was significantly higher in the Bevacizumab Group than in the Control Group (P = 0.017, 0.041, and 0.018, respectively). The surgical procedures performed and the visual acuity at all time points was not significantly different between groups (P > 0.05, all comparisons). The incidence of early (≤4 weeks) postoperative vitreous hemorrhage was significantly higher in the Bevacizumab Group (27%) than in the Control Group (7%; P = 0.027) although the rate of late (>4 weeks) postoperative vitreous hemorrhage was not significantly different between groups (P > 0.05).

Conclusion:

Vitrectomy with preoperative IVB may have no detrimental effect on surgical procedures and achieves the surgical outcomes for repair of PDR equal to vitrectomy alone despite the obvious selection bias of the patients in this study. However, special monitoring is highly recommended for early postoperative vitreous hemorrhage because bevacizumab in the vitreous may be washed out during vitrectomy.

INTRODUCTION

Proliferative diabetic retinopathy (PDR) is a representative ocular disorder with retinal neovascularization. Panretinal laser photocoagulation is the primary treatment for PDR,¹ and vitrectomy is performed for more complicated cases of PDR, such as, vitreous hemorrhage and tractional retinal detachment.²,³ Vascular endothelial growth factor (VEGF) is a key molecule that plays a role in the development of retinal neovascularization.⁴,⁵ In eyes with PDR, the VEGF level in the vitreous is correlated with the severity of PDR, and the level is reduced after panretinal laser photocoagulation.⁶

Bevacizumab (Avastin^®; F. Hoffmann-La Roche, Ltd., Basel, Switzerland) is a full-length humanized monoclonal antibody that is directed against all isoforms of VEGF. Although it is approved for intravenous use for the treatment of metastatic colon cancer,⁷ bevacizumab has been widely used in off-label treatments for ocular disorders with neovascularization or macular edema, e.g., age-related macular degeneration,⁸ retinal vein occlusion,⁹,¹⁰ and retinopathy of prematurity.¹¹ A new pharmacotherapeutic strategy has been used to treat PDR involving the use of bevacizumab as an adjunct for diabetic vitrectomy.^12–16

Intravitreal bevacizumab (IVB) before vitrectomy can reduce intraoperative bleeding and facilitate vitrectomy.¹²,¹³,¹⁶ Additionally, a randomized clinical trial by Ahmadieh et al.¹⁴ demonstrated that IVB one week prior to vitrectomy seems to reduce the incidence of early postvitrectomy hemorrhage in PDR cases. However, some retrospective series report that IVB pretreatment for PDR vitrectomy does not influence the rates of postoperative vitreous hemorrhage.¹⁵,¹⁶ The discrepancy in the effect of IVB on postvitrectomy hemorrhage among the previous reports^14–16 led the authors to compare the incidence of postoperative vitreous hemorrhage in PDR cases with or without preoperative IVB.

This retrospective study was performed to evaluate the effect of preoperative IVB on surgical procedures, visual prognosis, and postoperative complications, especially postoperative vitreous hemorrhage in PDR cases.

MATERIALS AND METHODS

The procedures used in this study conformed to the tenets of the Declaration of Helsinki, and were approved by the Institutional Review Board of Osaka Rosai Hospital. All patients were informed of the nature and possible consequences of the procedures, and a written informed consent was obtained from each subject.

Patients

The patients who had the primary vitrectomy for PDR from January 2006 through April 2008 at Osaka Rosai Hospital were investigated in this study. All the vitrectomies were performed by one experienced surgeon. The indications for vitrectomy were; PDR with non-clearing vitreous hemorrhage, active fibrovascular proliferation, and traction of retinal detachment involving or threatening the macula. Exclusion criteria included eyes that had undergone intraocular surgery within 6 months prior to vitrectomy, eyes that had undergone IVB more than one month prior to the vitrectomy, and eyes that had less than 6 months of follow-up after vitrectomy.

After the informed consent procedure, they chose to undergo either vitrectomy with preoperative IVB (Bevacizumab Group) or without preoperative IVB (Control Group).

The surgical procedures, best-corrected visual acuity (BCVA) in logarithm of the minimum angle resolution (logMAR) at baseline, 1, 3, and 6 months after the vitrectomy, and postoperative complications were compared between groups.

Intravitreal injection of bevacizumab

Bevacizumab (1.25 mg/0.05 ml) was injected into the vitreous in a surgical room in the outpatient clinic. Prior to the injection, 4% lidocaine hydrochloride and 0.625% povidone-iodine solutions¹⁷,¹⁸ were instilled in the eye. A lid speculum was inserted, and the injection of bevacizumab was performed with a 29-gauge needle inserted through the superior pars plana 3-4 mm from the limbus. The patient’s vision was tested after the injection to ensure that the retinal perfusion was intact. After the injection, the patients received topical antibiotic treatment with levofloxacin (Cravit^®; Santen Pharmaceutical Co., Ltd., Osaka, Japan) 4 times/day.

Vitrectomy

Standard three port vitrectomy was performed using 20- or 25-gauge systems under local anesthesia. The Accurus^® (Alcon Laboratories, Inc., Fort Worth, TX) vitrectomy surgical system was used. Phacoemulsification and intraocular lens implantation was also performed in phakic eyes.

Core vitrectomy was performed with a high-speed (2500 cycles/minute) vitreous cutter and a conventional halogen light source. Peripheral vitrectomy and vitreous base shaving were performed with scleral indentation using the halogen light source in the 20-gauge system or a xenon chandelier illumination fiber in the 25-gauge system. Fibrovascular membranes were initially cut and when needed, removed by a bimanual technique with cutters, scissors, and forceps. Intraoperative bleeding was controlled by temporarily raising the intraocular pressure or endocautery, or both. Panretinal laser photocoagulation was performed up to the peripheral retina. All retinal breaks were treated by photocoagulation. At the end of surgery, the intraocular pressure was decreased to identify and treat any bleeding from vessels. Fluid-gas exchange was performed for eyes with retinal breaks and rhegmatogenous retinal detachment, and if, required, long-standing gas was placed in the vitreous cavity.

Statistical analyses

Statistical analyses were performed with the SPSS program (Sigma Stat; Systat Software, Inc., San Jose, CA). Data are presented as the averages and standard deviations. If the data were normally and equally distributed, t tests were used to compare the two groups. If the data were not normally or equally distributed, the Mann-Whitney rank sum test was used to compare the two groups. A significant difference of the ratio between the two groups was determined by the Chi-square or Fisher exact test. A P value less than 0.05 was considered statistically significant.

RESULTS

The demographics of the patients enrolled in this study are presented in Table 1. Seventy-one eyes of 54 consecutive patients (23 eyes of 18 women; 48 eyes of 36 men) were studied. Twenty-five (35%) eyes were placed in the Bevacizumab Group and 46 (65%) eyes were placed in the Control Group. In the Bevacizumab Group, the mean interval between the IVB and vitrectomy was 11.6 days (range: 1-30 days). The preoperative BCVA, intraocular pressure, the ratio of phakic to pseudophakic eyes, and the incidence of prior panretinal laser photocoagulation were not statistically different between the two groups (P > 0.05, all comparisons). The patients were significantly younger in the Bevacizumab Group than in Control Group (P = 0.008). The incidence of preoperative complications, such as vitreous hemorrhage, tractional retinal detachment, and iris neovascularization, was significantly higher in the Bevacizumab Group than in the Control Group (P = 0.017, 0.041, and 0.018, respectively). The difference in the number of eyes in which the vitrectomy was performed with 20- or 25-gauge instrument was not significant between groups (P > 0.05).

Table 1.

Patient demographics and preoperative ocular and systemic status

The surgical procedures, postoperative BCVA and postoperative complications are summarized in Table 2. No significant ocular (specifically, the remarkable fibrovascular contraction leading to the aggravation of tractional retinal detachment) or systemic problems were observed after IVB in the Bevacizumab Group. The surgical procedures and postoperative BCVAs were not significantly different between groups. The incidence of late (>4 weeks) postoperative vitreous hemorrhage, progressive neovascular glaucoma, and a recurrent retinal detachment were not significantly different between groups. Whereas, the incidence of early (≤4 weeks) postoperative vitreous hemorrhage was significantly higher in the Bevacizumab Group (27%) than in the Control Group (7%; P = 0.027).

Table 2.

Surgical procedures, outcomes, and postoperative complications

The grading³ of the postoperative vitreous hemorrhage is shown in Table 3. Among the cases with vitreous hemorrhage, two cases with severe vitreous hemorrhage required surgery.

Table 3.

Incidence of postoperative hemorrhage

Discussion

We performed vitrectomy with or without preoperative IVB for eyes with PDR. The Bevacizumab Group were significantly younger and had more complications, such as vitreous hemorrhage, tractional retinal detachment, and iris neovascularization, than the cases in the Control Group (P < 0.05, all comparisons). Thus, it was expected that more frequent and complex surgical maneuvers and more repeat surgeries would be required in the Bevacizumab Group, which would indicate a poorer visual prognosis.

However, our results showed that the operating time and the necessity of a gas tamponade were not significantly different between groups. The incidence of reoperation was also not significantly different (P > 0.05). In addition, no significant difference in the postoperative visual acuity was observed at any time. These results suggest that preoperative IVB has no detrimental effect on surgical procedures and visual prognosis because no significant differences in surgical maneuvers and postoperative visual acuity were observed despite the obvious selection bias in the patients.

With regard to postoperative complications, the incidence of early postoperative vitreous hemorrhage was significantly higher in the Bevacizumab Group (27%) than in the Control Group (7%), while the rate of late postoperative vitreous hemorrhage was not significantly different between groups. Lo et al.¹⁵ demonstrated that IVB pretreatment for diabetic vitrectomy does not influence the rates of postoperative vitreous hemorrhage. Oshima et al.¹⁶ also showed that, although intraoperative bleeding cases achieved hemostasis much easier, preoperative IVB failed to prevent postoperative hemorrhage. Alternately, a randomized clinical study reported that IVB one week prior to vitrectomy was effective in reducing early postoperative vitreous hemorrhage in PDR cases.¹⁴

The reason for these contradictory results remains unclear. However, one possibility for the discrepancy may be the selection bias in this study. In fact, Lo et al.¹⁵ commented that, because the treated group with preoperative IVB clustered toward younger, more poorly managed diabetic mellitus, and greater severity of retinopathy, their treated groups were at a much higher risk of postoperative vitreous hemorrhage than their untreated group. The patients in this study were also younger and had more complicated retinopathy in the Bevacizumab Group than the Control Group.

Another difference between Ahmadieh et al.¹⁴ prospective study and the other studies¹⁵,¹⁶ was the interval between IVB and vitrectomy. Ahmadieh et al.¹⁴ performed IVB 1 week before vitrectomy whereas, in the other retrospective series including this study, the injection was performed within 30 days (mean, 7.6- 11.6 days) preoperatively.¹⁵,¹⁶ On this point; the pharmacokinetics of IVB in a rabbit model demonstrated that concentrations of >10 μg/ml bevacizumab were maintained in the vitreous humor for 30 days.¹⁹ Thus, it is believed that IVB reduced the vascular activity during the vitrectomy in our cases with PDR. In fact, there were no significant differences in the surgical maneuvers between the Bevacizumab Group and the Control Group despite the obvious selection bias. Oshima et al.¹⁶ also reported that intraoperative bleeding hemostasis was achieved much easier during intraoperative bleeding in cases with preoperative IVB.

Etiological studies have demonstrated that the causes of the early postoperative vitreous hemorrhage include dispersion of residual blood from the peripheral vitreous skirt or retinal surface into the vitreous cavity and from dissected fibrovascular tissue.¹⁵ Thus, in this study, peripheral vitrectomy and vitreous base shaving were performed with scleral indentation to prevent dispersion of residual blood, and hemostasis was confirmed by decreasing the intraocular pressure at the end of surgery. Nevertheless, early vitreous hemorrhage occurred at a higher rate in the Bevacizumab Group than the Control Group. One of the reasons may be that bevacizumab was likely to be washed out during vitrectomy and, thus unavailable to inhibit the growth of remaining or future fibrovascular tissues. As a result, there is a possibility that a recurrence of neovascularization and vitreous hemorrhage may appear soon after surgery before the intraoperative panretinal laser photocoagulation becomes effective. Itakura et al. collected fluid samples of cases with PDR from 5 to 36 days postoperatively and reported that a high VEGF level was maintained in the vitreous cavity after vitrectomy for PDR.²⁰ These results are in good agreement with our observations; postoperative vitreous hemorrhage within 4 weeks after surgery developed in PDR cases with preoperative IVB at a higher frequency. A prospective randomized trial by Ahn et al.²¹ demonstrated that the incidence of early vitreous hemorrhage was significantly decreased in the cases with IVB at the end of vitrectomy than in the cases without IVB. Reinjection of bevacizumab may be considered at the end of surgery to prevent postoperative vitreous hemorrhage.

The limitations of the current study are its retrospective nature, the lack of randomization, the limited number of patients and the potential selection bias. Further prospective randomized studies on a larger number of cases are needed to determine the effect of preoperative IVB on postoperative vitreous hemorrhage.

In conclusion, vitrectomy with preoperative IVB may have no detrimental effect on surgical procedures and achieve the surgical outcomes for repair of PDR equal to vitrectomy alone despite the obvious selection bias of the patients in this study. However, special monitoring must be done for early postoperative vitreous hemorrhage because bevacizumab in the vitreous may be washed out during vitrectomy.