RANDOMIZED CONTROLLED STUDY OF INTRAVITREAL BEVACIZUMAB 0.16 MG INJECTED ONE DAY BEFORE SURGERY FOR PROLIFERATIVE DIABETIC RETINOPATHY AYUMU MANABE, MD, HIROYUKI SHIMADA, MD, PHD, TAKAYUKI HATTORI, MD, HIROYUKI NAKASHIZUKA, MD, PHD, MITSUKO YUZAWA, MD, PHD Purpose: To investigate the usefulness of 0.16 mg/0.05 mL intravitreal bevacizumab (IVB) injection 1 day before vitrectomy for proliferative diabetic retinopathy. Methods: Sixty-two patients with proliferative diabetic retinopathy (66 eyes) with an indication for primary vitrectomy were randomized to IVB group (34 eyes) or sham control group (32 eyes). Intravitreal bevacizumab group received intravitreal injection of 0.16 mg/ 0.05 mL bevacizumab, and sham control group received sham injection 1 day before vitrectomy. Vitreous fluid was sampled before vitrectomy was started. Results: Frequency of reoperation due to recurrent vitreous hemorrhage within 4 weeks after surgery was significantly lower (P = 0.033) in IVB group (3.1%, 1/32) than in sham control group (20.6%, 7/34). The number of intraoperative endodiathermy spots (0.63 ± 1.0 vs. 1.3 ± 1.4, P = 0.025) and frequency of postoperative recurrent vitreous hemorrhage (3.1%, 1/32 vs. 23.5%, 8/34, P = 0.017) were significantly lower in IVB group than in sham control group. Vitreous vascular endothelial growth factor concentrations were 1315.3 ± 1153.4 pg/mL in sham control group and 25.0 ± 13.6 pg/mL in IVB group (P , 0.0001). Conclusion: Intravitreal injection of 0.16 mg/0.05 mL bevacizumab 1 day before vitrectomy blocked vascular endothelial growth factor production in vitreous and significantly reduced the incidence of reoperation due to early postoperative recurrent vitreous hemorrhage. RETINA 35:1800–1807, 2015
V
vitrectomy to reduce neovascular activity has been reported to prevent recurrent hemorrhage after vitrectomy for PDR.5–14 However, animal studies have demonstrated disruption of mitochondria in the inner segments of the photoreceptor layers and changes in the choriocapillaris in eyes after IVB injection.15–17 The possibility that these adverse also occur in human eyes receiving IVB cannot be ruled out. Clinically, diverse ocular complications such as onset or aggravation of tractional retinal detachment (TRD) have been reported after IVB injection (1.0 mg–1.25 mg).14,18 Furthermore, systemic complications including death and cerebrovascular events have been documented after intravitreal injections of anti-VEGF drugs.19 We hypothesized that to reduce the ocular and systemic adverse effects of IVB, intravitreal injection of a low concentration of bevacizumab and conducting vitrectomy shortly after the injection is useful. In
itreous hemorrhage (VH) tends to occur after vitrectomy for proliferative diabetic retinopathy (PDR) because of various factors such as difficulty of hemostasis during surgery, neovascularization stimulated by sclerotomy, neovascularization caused by residual vitreous, and contraction of remaining vitreous.1–3 A study reported that VH recurred in 75% of eyes during the early period after vitreous surgery for PDR.4 Intravitreal injection of 1.25 mg/0.05 mL of bevacizumab, an inhibitor of vascular endothelial growth factor (VEGF), 1 days to 20 days before From the Department of Ophthalmology, School of Medicine, Nihon University, Chiyodaku, Japan. None of the authors have any financial/conflicting interests to disclose. Reprint requests: Hiroyuki Shimada, MD, PhD, Department of Ophthalmology, Surugadai Hospital of Nihon University, 1-8-13 Surugadai, Kanda, Chiyodaku, Tokyo 101-8309, Japan; e-mail: [email protected]
1800
IVB INJECTION MANABE ET AL
a pilot study, we found that intravitreal injection of bevacizumab at a minimum amount of 0.16 mg/0.05 mL before vitrectomy for PDR blocked VEGF in vitreous and reduced the frequency of using hemostatic procedures during surgery.20 In the present clinical study, we aimed to verify the usefulness of intravitreal injection of 0.16 mg/0.05 mL bevacizumab 1 day before conducting vitrectomy for PDR in reducing postoperative recurrent VH using the frequency of reoperation as the primary outcome measure. Patients and Methods Study Design The present study was a prospective, doublemasked, randomized, controlled clinical trial. The trial was registered at the ClinicalTrials.gov registry (NCT01854593) and the UMIN Clinical Trial Registry (UMIN000007482) before the study was started. This study was approved by the Ethical Committee of the Nihon University School of Medicine (Number 110903). The study adhered to the tenets of the Declaration of Helsinki. All subjects provided written informed consent after receiving full explanations of the study and the potential merits and risks. Patient Population Between June 2012 and August 2013 at Surugadai Hospital of Nihon University, we recruited 62 consecutive patients (66 eyes) diagnosed as having an indication for primary vitrectomy because of persistent VH for over 3 months and TRD caused by PDR; or fundus findings of proliferative changes or VH in patients with iris neovascularization or neovascular glaucoma (NVG). Patients who had undergone intraocular surgery or retinal photocoagulation within 3 months before the study and those who had undergone intravitreal injection of drugs or sub-Tenon injection of steroids within 3 months before the study were excluded. Patients who had cerebrovascular infarction or myocardial infarction within 3 months before the study were also excluded. In the IVB group, patients with bilateral PDR received IVB in one eye only, and the other eye was excluded from the study, considering that IVB injected in one eye may pass to the contralateral eye. In the control group, both eyes of bilateral cases were included in the study. Randomization and Masking Among the patients enrolled, 62 patients (66 eyes) who provided informed consent were randomized
1801
using the envelop method by 1 investigator (A.M.) to 2 groups: sham control group and IVB injection group. Both the investigators and patients were masked to the treatment assignment. One day before the scheduled surgery, the patients were admitted and received an intravitreal injection of 0.16 mg/0.05 mL of bevacizumab or sham injection. In the sham control group, a syringe with the needle removed was used to press on the injection site to mimic injection to ensure patient masking.21 One day after injection, three vitreoretinal surgeons (H.S., T.H., H.N.; excluding A.M. who performed randomization) who had no knowledge of the randomization conducted the vitrectomies (double masking). There were no differences in techniques, instruments used, and treatment methods among the three vitreoretinal surgeons. Bevacizumab (Avastin; Genentech, San Francisco, CA) for injection was prepared in the pharmacy department of our hospital, using a biologically clean room. Bevacizumab 100 mg/4 mL was diluted with physiological saline to a concentration of 0.16 mg/0.05 mL, and aliquots of 0.05 mL were dispensed in 1.0mL syringes. The aliquots were stored at −70°C and used within 2 months. Bacteriological culture was performed on the aliquots immediately after dispensing and 2 months later. Intravitreal Bevacizumab Injection and Vitrectomy Bevacizumab was injected by the following procedures. After administering oxybuprocaine hydrochloride ophthalmic solution (0.4% Benoxil, Santen Pharmaceutical Co., Osaka, Japan) as a local anesthetic, the conjunctiva was disinfected with 0.25% (1:40 dilution of neat solution) povidone-iodine (Isodine, Meiji Seika Pharma Co., Ltd., Tokyo, Japan).22 After placing the lid speculum, the conjunctiva was again disinfected with 0.25% povidone-iodine, and bevacizumab was injected intravitreally at a dose of 0.16 mg/0.05 mL. In the sham control group, after administering oxybuprocaine hydrochloride ophthalmic solution, a needleless syringe was pressed against the conjunctiva. On the next day after bevacizumab or sham injection, vitrectomy was conducted using the Constellation Visual System (Alcon Surgical, Fort Worth, TX) with a 25-G vitrectomy system (Alcon Surgical). A trocar was inserted at a site 4 mm from the corneal limbus. With infusion arrested, a 25-G vitreous cutter was inserted into the center of the vitreous body, and 0.5 mL of vitreous fluid was collected into a 2.5-mL syringe connected to the aspiration line. Thereafter, vitrectomy was conducted. During surgery, bleeding from the neovascular membrane was controlled using
1802 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
a 25-G straight bipolar pencil (Kirwan Surgical Products, Marshfield, MA) or a 25-G straight laser probe. The numbers of intraoperative endodiathermy spots and endolaser spots were recorded. Immediately after collection, the vitreous sample was transferred to a test tube and stored at −70°C. Two to three days later, the sample was thawed immediately before assay, and vitreous VEGF concentration was measured using an enzyme-linked immunosorbent assay for human VEGF (R&D Systems, Minneapolis, MN). Postoperative Follow-up Patients were examined daily from 1 to 7 days and at 2 weeks and 1 month after surgery. At each followup, all patients underwent a full ophthalmologic examination, intraocular pressure (IOP) measurement, slit-lamp biomicroscopy, and indirect ophthalmoscopy. All postoperative examinations were performed by three vitreoretinal surgeons (H.S., T.H., H.N.) who also were masked to the study group allocation. Intraocular pressure was measured at every follow-up examination. During the 4-week period, any increase of IOP to $25 mmHg was defined as elevated IOP. Corrected visual acuity was tested after 1 month. Visual acuity was measured by an optometrist who was masked to the randomization. Visual acuity was measured using the Landolt ring chart, and the result was expressed in Snellen visual acuity ratio and converted to logarithm of minimal angle of resolution (logMAR) scale for statistical analysis. Improvement of best-corrected visual acuity (BCVA) was defined as the difference between postoperative and preoperative BCVA and expressed in letters of improvement (Early Treatment Diabetic Retinopathy Study chart) as well as change in logMAR for statistical analysis. Postoperative recurrent VH was defined as bleeding that rendered the optic disk invisible. In eyes with silicon infusion, postoperative recurrent hemorrhage was defined as bleeding of 1 disk diameter or greater. At the end of vitrectomy, we confirmed that there was no retinal hemorrhage in all cases. Therefore, retinal hemorrhage occurring after vitrectomy was considered to be “recurrent VH” and not “persistent VH.” All patients with PDR were hospitalized for the primary vitrectomy and 1-week postoperative observation, or longer if necessary. To prevent hospitalization from becoming long term, we performed repeat vitrectomy in patients who developed severe recurrent VH with no fundus details for over 5 days and showing no tendency of resorption. We did not inject bevacizumab 0.16 mg for the purpose of promoting resorption of retinal hemorrhage and wait for spontaneous resorption to take place.
2015 VOLUME 35 NUMBER 9
Outcome Measures The following were evaluated during surgery: vitreous VEGF concentration, number of intraoperative endodiathermy spots, number of intraoperative endolaser spots, iatrogenic retinal tears, type of tamponade, and surgical time. The postoperative parameters evaluated included early postoperative (within 4 weeks) recurrent VH, reoperation due to recurrent VH, postoperative (within 4 weeks) IOP, persistence or progression of NVG, 1-month postoperative visual acuity, and degree of visual improvement. Primary outcome measure was the frequency of reoperation due to postoperative recurrent VH within 4 weeks after vitrectomy. Secondary outcome measures were numbers of intraoperative laser and endodiathermy spots, frequency of postoperative recurrent VH within 4 weeks after vitrectomy, and concentration of VEGF in vitreous at the beginning of surgery. Statistics Sample size was estimated using reported incidence of recurrent VH because we could not find reports that examined the rate of reoperation for recurrent VH, which was the primary outcome measure. With study power of 90%, significance level set at 0.05, and an assumed decrease in incidence of early postvitrectomy VH from 60% to 30% with preoperative IVB, a sample size of 30 eyes for each arm was calculated. Data are expressed as mean ± SD or percentage of patients. Chi-square test or Fisher’s exact test was used to analyze categorical variables. The frequencies of postoperative recurrent VH and reoperation due to postoperative recurrent VH were analyzed using the chi-square test to compare the differences between two groups. The frequency of intraoperative endodiathermy use, concentration of VEGF in vitreous, postoperative BCVA, and the improvement of postoperative BCVA were analyzed using the Mann– Whitney rank-sum test. Statistical analyses were performed using SPSS version 21 (SPSS, Inc., Chicago, IL). A P value less than 0.05 was considered statistically significant.
Results Participant Flow Patient flow is shown in Figure 1. Primary vitrectomy was conducted in a total of 68 eyes during the study period. Two patients (2 eyes) did not give informed consent and were excluded from the study.
1803
IVB INJECTION MANABE ET AL
Fig. 1. Patient flow.
Sixty-six eyes (62 patients) were randomized to IVB group (32 eyes of 32 patients) or sham control group (34 eyes of 30 patients). All patients completed follow-up for 1 month. Eventually, 32 eyes (32 patients) in IVB group and 34 eyes (30 patients) in sham control group were analyzed. Baseline Data Baseline patient data are shown in Table 1. The two groups did not differ significantly in age; BCVA; IOP;
history of hypertension, hyperlipidemia, or cerebral or myocardial infarction; glycated hemoglobin (HbA1c); use of anticoagulant; lens status; persistent VH; status of iris neovascularization or NVG; status of traction retinal detachment; and previous panretinal photocoagulation (PRP). Traction retinal detachment was defined as retinal detachment of 1 disk diameter or greater. Previous PRP was classified as “none” when the patient did not undergo PRP before surgery, and as “incomplete” when PRP conducted before vitrectomy was inadequate and additional PRP was given during
Table 1. Baseline Demographic and Clinical Data of All Patients
Age, mean ± SD (range), years Gender, male/female Preoperative BCVA, logMAR ± SD (Snellen visual acuity ratio) IOP, mean ± SD (range), mmHg Hypertension, n (%) Hyperlipidemia, n (%) Cerebral and myocardial infarction, n (%) HbA1c, mean ± SD, % Use of anticoagulants, n (%) Lens status, phakic/pseudophaki Preoperative complications, n (%) Persistent VH INV or NVG TRD Previous PRP, n (%) None Incomplete
IVB Group (n = 32)
Sham Control Group (n = 34)
P
59.9 ± 11.8 (20–79) 22/10 1.09 ± 0.60 (20/100)
59.2 ± 12.9 (35–82) 32/2 1.14 ± 0.58 (20/120)
0.667* 0.008† 0.756*
15.3 ± 2.69 (9–21) 14 (43.8) 5 (15.6) 5 (15.6) 8.0 ± 1.7 7 (21.9) 28/4
14.6 ± 3.42 (7–25) 16 (47.1) 7 (20.6) 6 (17.6) 7.6 ± 2.0 8 (23.5) 31/3
0.157* 0.787† 0.601† 0.826† 0.500* 0.873† 0.465†
29 (90.6) 2 (6.3) 5 (1.5)
30 (88.2) 2 (5.9) 4 (1.2)
0.535† 0.670† 0.460†
9 (28.1) 23 (71.9)
11 (32.4) 23 (67.6)
0.709† 0.709†
*Mann–Whitney rank-sum test. †Chi-square test. HbA1c, hemoglobin A1c; INV, iris neovascularization; logMAR, logarithm of minimal angle of resolution.
1804 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
vitrectomy. The only difference between two groups was a higher male/female ratio in the sham control group. Outcome Measures The results of intraoperative and postoperative parameters are summarized in Table 2. The frequency of reoperation due to postoperative (within 4 weeks) recurrent VH was 20.6% (7/34 eyes) in the sham control group and 3.1% (1/32 eyes) in the IVB group and was significantly lower in the IVB group (P = 0.033). The number of endolaser spots was not different between IVB group and sham control group (1260 ± 686 vs. 1335 ± 663, P = 0.667). The number of endodiathermy spots was 1.3 ± 1.4 in the sham control group and 0.63 ± 1.0 in the IVB group and was significantly lower in the IVB group (P = 0.025). This difference, while statistically significant, may not be clinically relevant because both represent one spot of diathermy. The incidence of postoperative (within 4 weeks) recurrent VH was 23.5% (8/34 eyes) in the sham control group and 3.1% (1/32 eyes) in the IVB group and was also significantly lower in the IVB group (P = 0.017). Among eyes that had silicon oil infusion (6 in sham control group and 4 in IVB group), 2 eyes in the sham control group (2/6, 33.3%) and 0 eye in the IVB group (0/4, 0%) had postoperative recurrent VH, with no significant difference between 2 groups (P = 0.467).
2015 VOLUME 35 NUMBER 9
Vascular endothelial growth factor concentration in vitreous was 25.0 ± 13.6 pg/mL in the IVB group and 1315.3 ± 1153.4 pg/mL in the sham control group and was significantly lower (P , 0.0001) in the IVB group. Other Clinical Results There were no significant differences between 2 groups in other intraoperative parameters such as intraoperative laser treatment, the frequency of iatrogenic retinal breaks, the frequencies of using gas and silicon oil tamponade, and surgical time (Table 2). Postoperative 1-month BCVA (including reoperated eyes) improved significantly in both IVB group (logMAR: 1.09 ± 0.60 to 0.46 ± 0.54, P , 0.001; Snellen ratio: 20/100 to 20/35, Early Treatment Diabetic Retinopathy Study: 20 letters of improvement) and sham control group (logMAR: 1.14 ± 0.58 to 0.43 ± 0.48, P , 0.001; Snellen ratio: 20/120 to 20/35, Early Treatment Diabetic Retinopathy Study: 25 letters of improvement). However, postoperative BCVA and visual improvement were not significantly different between 2 groups (P = 0.929 and 0.445, respectively). There were no significant differences in the proportions of patients showing postoperative (within 4 weeks) IOP elevation and persistence or progression of NVG (Table 2). Intraoperative iatrogenic retinal breaks occurred in 5 of 32 eyes in the IVB group and 5 of 34 eyes in the sham control group. In all these
Table 2. Vitreous VEGF Level, Intra/Postoperative Outcomes and Adverse Events Examined
Intraoperative outcomes VEGF concentration in vitreous, mean ± SD, pg/mL Endolaser spots, mean ± SD, n Endodiathermy spots, mean ± SD, n Iatrogenic retinal breaks, n (%) Gas tamponade (air or long acting gas), n (%) Silicon oil tamponade, n (%) Surgical time, mean ± SD, minutes Postoperative outcomes Postoperative VH, n (%) Vitreoretinal reoperation due to recurrent VH, n (%) Elevation of IOP, n (%) Progressive or persistent NVG, n (%) Postoperative BCVA, logMAR ± SD (Snellen visual acuity ratio) Mean BCVA change, logMAR ± SD (ETDRS chart) Injection-related adverse events Systemic thromboembolic event Endophthalmitis
IVB Group (n = 32)
Sham Control Group (n = 34)
P
25.0 ± 13.6 1260 ± 686 0.63 ± 1.0 5 (15.6) 23 (71.9) 4 (12.5) 49 ± 20
1315.3 ± 1153.4 1335 ± 663 1.3 ± 1.4 5 (14.7) 24 (70.6) 6 (17.6) 56 ± 27
,0.0001* 0.667* 0.025* 0.593† 0.670† 0.378† 0.298*
1 (3.1) 1 (3.1) 2 (6.3) 0 (0) 0.46 ± 0.54 (20/35)
8 (23.5) 7 (20.6) 6 (17.6) 3 (8.8) 0.43 ± 0.48 (20/35)
0.017† 0.033† 0.149† 0.131‡ 0.929*
−0.63 ± 061 (20 letters)
−0.73 ± 0.80 (25 letters)
0.445*
0 0
0 0
*Mann–Whitney rank-sum test. †Chi-square test. ‡Fisher’s exact test. ETDRS, Early Treatment Diabetic Retinopathy Study; logMAR, logarithm of minimal angle of resolution; SO, silicone oil.
.0.999‡ .0.999‡
IVB INJECTION MANABE ET AL
eyes, the retinal breaks were treated by intraoperative endolaser photocoagulation, and retinal reattachment was confirmed during vitrectomy. No bacteria were isolated from the aliquoted bevacizumab. Injection-related Adverse Events No ocular and no systemic adverse events associated with IVB such as systemic thromboembolic event and endophthalmitis were observed in the present study (Table 2). Discussion In the present prospective, randomized, controlled clinical trial conducted in patients undergoing vitrectomy for PDR, we blocked VEGF in the vitreous by injecting bevacizumab at a dose (0.16 mg/0.05 mL) equivalent to one eighth of the usual dose 1 day before vitrectomy. Vitreous VEGF level was markedly reduced in patients given intravitreal bevacizumab. Since vitreous samples were frozen immediately after collection and stored at −70°C until assay, there is little possibility that residual bevacizumab in the sample binds with VEGF resulting in exaggerated reduction in active VEGF when measured using the enzymelinked immunosorbent assay. The primary outcome measure; frequency of vitreoretinal reoperation due to postoperative recurrent VH, and the secondary outcome measures; frequencies of intraoperative endodiathermy use and postoperative recurrent VH, clearly demonstrated a significantly beneficial effect of the IVB in preventing VH during and after vitrectomy. Two meta-analyses on preoperative IVB (1.25 mg/ 0.05 mL) for severe PDR have been reported.23,24 Both analyses led to a common conclusion that adjuvant IVB before vitrectomy in patients with PDR significantly decreased intraoperative bleeding, decreased the frequency of intraoperative endodiathermy, shortened the surgical time, and reduced postoperative (within 1 month) VH compared with controls.23,24 One of the two analyses also reported a lower frequency of intraoperative tear,24 shorter postoperative reabsorption time of vitreous bleeding, lower incidence of late (after 1 week) IOP elevation,23 and better final BCVA23 in the IVB group than in the control group. However, early postoperative (within 1 week) IOP elevation,23 BCVA at 3 months after surgery, late postoperative (after 1 month) VH, rate of vitreoretinal reoperation,24 rate of postoperative retinal detachment,23,24 and rate of NVG24 were not significantly different between two groups. Our study showed that intravitreal injection of bevacizumab at a dose of only 0.16 mg/0.05 mL
1805
reduced the frequency of reoperation due to postoperative recurrent hemorrhage. Although vitrectomy on the next day rapidly eliminates the bevacizumab injected into the vitreous, the low-dose (0.16 mg/0.05 mL) bevacizumab injected 1 day before vitrectomy probably penetrates into the retina to prevent postoperative recurrent hemorrhage. Indeed, previous reports indicate that intravitreally injected bevacizumab penetrates into the retina within 1 day25,26 and passes into systemic circulation.27,28 Avery et al29 injected bevacizumab at concentrations of 6.2 to 625 mg intravitreally in patients with retinal and iris neovascularization secondary to diabetes mellitus and observed reduction of fluorescein angiographic leakage 24 hours after injection. Therefore, we assumed that intravitreal injection of bevacizumab 0.16 mg results in penetration of the drug through the retina and exhibition of anti-VEGF effect after 24 hours. By reducing intraoperative bleeding and decreasing the frequency of hemostasis procedure during surgery, IVB injection before surgery facilitates the surgical procedures. Moreover, the decreases in frequency of postoperative recurrent VH and frequency of reoperation clearly indicate that this approach is beneficial to patients. In this study, visual acuity at 1 month after surgery improved markedly in both groups, but postoperative visual acuity and postoperative visual improvement did not differ between 2 groups. In patients with age-related macular degeneration, retinal vein occlusion, and diabetic macular edema, high doses of VEGF inhibitors are injected intravitreally to achieve a sustained therapeutic effect for 1 to 2 months. Serious adverse events potentially related to systemic VEGF inhibition, including cardiovascular events, cerebrovascular accidents, and death have been reported.19 Carneiro et al19 studied the risk of arterial thromboembolic events in 378 patients with exudative age-related macular degeneration treated with bevacizumab or ranibizumab. The incidence of arterial thromboembolic events was 12.4% (12/97 patients) with bevacizumab and 1.4% (3/219 patients) with ranibizumab, with odds ratio of 10.16. The methods reported in the literature for reducing the rates of early postvitrectomy recurrent hemorrhage involve intravitreal injection of 1.25 mg to 2.5 mg of the VEGF inhibitor bevacizumab 1 day to 20 days before vitrectomy (Table 3). Intravitreally injected bevacizumab at a dose of 1.25 mg has been reported to significantly reduce VEGF concentration in blood.30–32 Because of the vascular risk potentially associated with system VEFG inhibition, preoperative IVB injection was not indicated in patients with PDR with a history of cerebrovascular event or myocardial infarction.20 We used one eighth the normal
1806 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 9
Table 3. Summary of Studies of Vitrectomy With or Without Preoperative IVB Injection for Proliferative Diabetic Retinopathy
Eyes, n
Dose of IVB
Interval Between IVB Injection and Vitrectomy
Rizzo et al5 El-Batarny6
22 30
1.25 mg 1.25 mg/0.05 mL
5–7 days 5–7 days
Modarres et al7
40
2.5 mg/0.1 mL
3–5 days
Ahmadieh et al8 da R Lucena et al9
68 20
1.25 mg 1.5 mg/0.06 mL
1 week 2 weeks
Hernandez-Da Mota and NuñezSolorio10 di Lauro et al11
40
1.25 mg/0.05 mL
2 days
72
1.25 mg
7 or 20 days
107
1.25 mg/0.05 mL
1–14 days
Farahvash et al13
35
1.25 mg
7 days
This study
66
0.16 mg
1 day
Ahn et al12
dose of the VEGF inhibitor bevacizumab and removed residual VEGF inhibitor in the vitreous the next day by conducting vitrectomy. This method may reduce the amount of VEGF inhibitor in systemic circulation and may potentially be safer than the conventional methods. In this study, we included patients with cerebrovascular event or myocardial infarction more than 3 months before the study and observed no systemic adverse events. This study is significant from the viewpoint of expanding the indication of this approach. Regarding ocular complications, Oshima et al14 reported progression of TRD after IVB injection, and Arevalo et al18 reported onset of TRD in 11 of 211 eyes (5.2%) at 13 days on average after IVB injection. The risk of exacerbation or development of TRD may increase as the duration between injection and surgery increases. In this study, 90% of the patients had VH before vitrectomy, and assessment of TRD was difficult in these patients. As a limitation of this study, the observation period was short, and further investigation of long-term outcome is needed. Vitrectomies were done by three surgeons in this study. More reliable results may have been obtained with one surgeon performing all the vitrectomies. Moreover, the decision of reoperation was probably influenced by various factors such as the severity of PDR, compliance with positioning, and minor differences in techniques between surgeons.
Main Outcome Feasibility of surgery Feasibility of surgery, postoperative complications BCVA, number of endodiathermy applications, backflush needle applications, duration of surgery, and postoperative VH Incidence of early postvitrectomy hemorrhage Erythrocyte count in the fluid retrieved from the vitrectomy cassette BCVA, intraocular pressure, and fundus photographs Intraoperative management, safety, efficacy of IVB Incidence of early (#4 weeks) and late (.4 weeks) recurrent VH Facility of surgery (severity of intraoperative bleedings and intraoperative break formation) Vitreous VEGF concentration, intraoperative hemorrhage, recurrent postoperative VH, reoperation due to recurrent VH
In addition, the present results obtained from a single center have to be validated in multiple institutions. Further study is required to examine whether plasma VEGF concentration is reduced by intravitreal injection of 0.16 mg bevacizumab. Key words: bevacizumab, intravitreal injection, postoperative recurrent vitreous hemorrhage, proliferative diabetic retinopathy, prospective randomized controlled clinical trial, vascular endothelial growth factor. References 1. Novak MA, Rice TA, Michels RG, Auer C. Vitreous hemorrhage after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1485–1489. 2. Tolentino FI, Freeman HM, Tolentino FL. Closed vitrectomy in the management of diabetic traction retinal detachment. Ophthalmology 1980;87:1078–1089. 3. Lewis H, Abrams GW, Williams GA. Anterior hyaloidal fibrovascular proliferation after diabetic vitrectomy. Am J Ophthalmol 1987;104:607–613. 4. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Complications of vitreous surgery for diabetic retinopathy II. Postoperative complications. Ophthalmology 1983;90:522–530. 5. Rizzo S, Genovesi-Ebert F, Di Bartolo E, et al. Injection of intravitreal bevacizumab (Avastin) as a preoperative adjunct before vitrectomy surgery in the treatment of severe proliferative diabetic retinopathy (PDR). Graefes Arch Clin Exp Ophthalmol 2008;246:837–842. 6. El-batarny AM. Intravitreal bevacizumab as an adjunctive therapy before diabetic vitrectomy. Clin Ophthalmol 2008;2:709–716.
IVB INJECTION MANABE ET AL 7. Modarres M, Nazari H, Falavarjani KG, et al. Intravitreal injection of bevacizumab before vitrectomy for proliferative diabetic retinopathy. Eur J Ophthalmol 2009;19:848–852. 8. Ahmadieh H, Shoeibi N, Entezari M, Monshizadeh R. Intravitreal bevacizumab for prevention of early postvitrectomy hemorrhage in diabetic patients: a randomized clinical trial. Ophthalmology 2009;116:1943–1948. 9. da R Lucena D, Ribeiro JA, Costa RA, et al. Intraoperative bleeding during vitrectomy for diabetic tractional retinal detachment with versus without preoperative intravitreal bevacizumab (IBeTra study). Br J Ophthalmol 2009;93:688–691. 10. Hernández-Da Mota SE, Nuñez-Solorio SM. Experience with intravitreal bevacizumab as a preoperative adjunct in 23-G vitrectomy for advanced proliferative diabetic retinopathy. Eur J Ophthalmol 2010;20:1047–1052. 11. di Lauro R, De Ruggiero P, di Lauro R, et al. Intravitreal bevacizumab for surgical treatment of severe proliferative diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol 2010; 248:785–791. 12. Ahn J, Woo SJ, Chung H, Park KH. The effect of adjunct intravitreal bevacizumab for preventing postvitrectomy hemorrhage in proliferative diabetic retinopathy. Ophthalmology 2011;118:2218–2226. 13. Farahvash MS, Majidi AR, Roohipoor R, Ghassemi F. Preoperative injection of intravitreal bevacizumab in dense diabetic vitreous hemorrhage. Retina 2011;31:1254–1260. 14. Oshima Y, Shima C, Wakabayashi T, et al. Microincision vitrectomy surgery and intravitreal bevacizumab as a surgical adjunct to treat diabetic traction retinal detachment. Ophthalmology 2009;116:927–938. 15. Inan UU, Avci B, Kusbeci T, et al. Preclinical safety evaluation of intravitreal injection of full-length humanized vascular endothelial growth factor antibody in rabbit eyes. Invest Ophthalmol Vis Sci 2007;48:1773–1781. 16. Peters S, Heiduschka P, Julien S, et al. Ultrastructural findings in the primate eye after intravitreal injection of bevacizumab. Am J Ophthalmol 2007;143:995–1002. 17. Shimomura Y, Hirata A, Ishikawa S, Okinami S. Changes in choriocapillaris fenestration of rat eyes after intravitreal bevacizumab injection. Graefes Arch Clin Exp Ophthalmol 2009; 247:1089–1094. 18. Arevalo JF, Maia M, Flynn HW Jr., et al. Tractional retinal detachment following intravitreal bevacizumab (Avastin) in patients with severe proliferative diabetic retinopathy. Br J Ophthalmol 2008;92:213–216. 19. Carneiro AM, Barthelmes D, Falcão MS, et al. Arterial thromboembolic events in patients with exudative age-related macular degeneration treated with intravitreal bevacizumab or ranibizumab. Ophthalmologica 2011;225:211–221. 20. Hattori T, Shimada H, Nakashizuka H, et al. Dose of intravitreal bevacizumab (Avastin) used as preoperative adjunct
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
1807
therapy for proliferative diabetic retinopathy. Retina 2010;30: 761–764. Brown DM, Campochiaro PA, Singh RP, et al. Ranibizumab for macular edema following central retinal vein occlusion: sixmonth primary end point results of a phase III study. Ophthalmology 2010;117:1124–1133. Shimada H, Hattori T, Mori R, et al. Minimizing the endophthalmitis rate following intravitreal injections using 0.25% povidone-iodine irrigation and surgical mask. Graefes Arch Clin Exp Ophthalmol 2013;251:1185–1190. Zhao LQ, Zhu H, Zhao PQ, Hu YQ. A systematic review and meta-analysis of clinical outcomes of vitrectomy with or without intravitreal bevacizumab pretreatment for severe diabetic retinopathy. Br J Ophthalmol 2011;95:1216–1222. Zhang ZH, Liu HY, Hernandez-Da Mota SE, et al. Vitrectomy with or without preoperative intravitreal bevacizumab for proliferative diabetic retinopathy: a meta-analysis of randomized controlled trials. Am J Ophthalmol 2013;156:106–115. Heiduschka P, Fietz H, Hofmeister S, et al. Penetration of bevacizumab through the retina after intravitreal injection in the monkey. Invest Ophthalmol Vis Sci 2007;48: 2814–2823. Julien S, Heiduschka P, Hofmeister S, Schraermeyer U. Immunohistochemical localisation of intravitreally injected bevacizumab at the posterior pole of the primate eye: implication for the treatment of retinal vein occlusion. Br J Ophthalmol 2008; 92:1424–1428. Bakri SJ, Snyder MR, Reid JM, et al. Pharmacokinetics of intravitreal bevacizumab (Avastin). Ophthalmology 2007; 114:855–859. Matsuyama K, Ogata N, Matsuoka M, et al. Effects of intravitreally injected bevacizumab on vascular endothelial growth factor in fellow eyes. J Ocul Pharmacol Ther 2011;27: 379–383. Avery L, Pearlman J, Pieramici J, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology 2006;113:1695.e1-15. Zehetner C, Kirchmair R, Huber S, et al. Plasma levels of vascular endothelial growth factor before and after intravitreal injection of bevacizumab, ranibizumab and pegaptanib in patients with age-related macular degeneration, and in patients with diabetic macular oedema. Br J Ophthalmol 2013;97: 454–459. Carneiro AM, Costa R, Falcão MS, et al. Vascular endothelial growth factor plasma levels before and after treatment of neovascular age-related macular degeneration with bevacizumab or ranibizumab. Acta Ophthalmol 2012;90:e25–e30. Matsuyama K, Ogata N, Matsuoka M, et al. Plasma levels of vascular endothelial growth factor and pigment epitheliumderived factor before and after intravitreal injection of bevacizumab. Br J Ophthalmol 2010;94:1215–1258.
V
vitrectomy to reduce neovascular activity has been reported to prevent recurrent hemorrhage after vitrectomy for PDR.5–14 However, animal studies have demonstrated disruption of mitochondria in the inner segments of the photoreceptor layers and changes in the choriocapillaris in eyes after IVB injection.15–17 The possibility that these adverse also occur in human eyes receiving IVB cannot be ruled out. Clinically, diverse ocular complications such as onset or aggravation of tractional retinal detachment (TRD) have been reported after IVB injection (1.0 mg–1.25 mg).14,18 Furthermore, systemic complications including death and cerebrovascular events have been documented after intravitreal injections of anti-VEGF drugs.19 We hypothesized that to reduce the ocular and systemic adverse effects of IVB, intravitreal injection of a low concentration of bevacizumab and conducting vitrectomy shortly after the injection is useful. In
itreous hemorrhage (VH) tends to occur after vitrectomy for proliferative diabetic retinopathy (PDR) because of various factors such as difficulty of hemostasis during surgery, neovascularization stimulated by sclerotomy, neovascularization caused by residual vitreous, and contraction of remaining vitreous.1–3 A study reported that VH recurred in 75% of eyes during the early period after vitreous surgery for PDR.4 Intravitreal injection of 1.25 mg/0.05 mL of bevacizumab, an inhibitor of vascular endothelial growth factor (VEGF), 1 days to 20 days before From the Department of Ophthalmology, School of Medicine, Nihon University, Chiyodaku, Japan. None of the authors have any financial/conflicting interests to disclose. Reprint requests: Hiroyuki Shimada, MD, PhD, Department of Ophthalmology, Surugadai Hospital of Nihon University, 1-8-13 Surugadai, Kanda, Chiyodaku, Tokyo 101-8309, Japan; e-mail: [email protected]
1800
IVB INJECTION MANABE ET AL
a pilot study, we found that intravitreal injection of bevacizumab at a minimum amount of 0.16 mg/0.05 mL before vitrectomy for PDR blocked VEGF in vitreous and reduced the frequency of using hemostatic procedures during surgery.20 In the present clinical study, we aimed to verify the usefulness of intravitreal injection of 0.16 mg/0.05 mL bevacizumab 1 day before conducting vitrectomy for PDR in reducing postoperative recurrent VH using the frequency of reoperation as the primary outcome measure. Patients and Methods Study Design The present study was a prospective, doublemasked, randomized, controlled clinical trial. The trial was registered at the ClinicalTrials.gov registry (NCT01854593) and the UMIN Clinical Trial Registry (UMIN000007482) before the study was started. This study was approved by the Ethical Committee of the Nihon University School of Medicine (Number 110903). The study adhered to the tenets of the Declaration of Helsinki. All subjects provided written informed consent after receiving full explanations of the study and the potential merits and risks. Patient Population Between June 2012 and August 2013 at Surugadai Hospital of Nihon University, we recruited 62 consecutive patients (66 eyes) diagnosed as having an indication for primary vitrectomy because of persistent VH for over 3 months and TRD caused by PDR; or fundus findings of proliferative changes or VH in patients with iris neovascularization or neovascular glaucoma (NVG). Patients who had undergone intraocular surgery or retinal photocoagulation within 3 months before the study and those who had undergone intravitreal injection of drugs or sub-Tenon injection of steroids within 3 months before the study were excluded. Patients who had cerebrovascular infarction or myocardial infarction within 3 months before the study were also excluded. In the IVB group, patients with bilateral PDR received IVB in one eye only, and the other eye was excluded from the study, considering that IVB injected in one eye may pass to the contralateral eye. In the control group, both eyes of bilateral cases were included in the study. Randomization and Masking Among the patients enrolled, 62 patients (66 eyes) who provided informed consent were randomized
1801
using the envelop method by 1 investigator (A.M.) to 2 groups: sham control group and IVB injection group. Both the investigators and patients were masked to the treatment assignment. One day before the scheduled surgery, the patients were admitted and received an intravitreal injection of 0.16 mg/0.05 mL of bevacizumab or sham injection. In the sham control group, a syringe with the needle removed was used to press on the injection site to mimic injection to ensure patient masking.21 One day after injection, three vitreoretinal surgeons (H.S., T.H., H.N.; excluding A.M. who performed randomization) who had no knowledge of the randomization conducted the vitrectomies (double masking). There were no differences in techniques, instruments used, and treatment methods among the three vitreoretinal surgeons. Bevacizumab (Avastin; Genentech, San Francisco, CA) for injection was prepared in the pharmacy department of our hospital, using a biologically clean room. Bevacizumab 100 mg/4 mL was diluted with physiological saline to a concentration of 0.16 mg/0.05 mL, and aliquots of 0.05 mL were dispensed in 1.0mL syringes. The aliquots were stored at −70°C and used within 2 months. Bacteriological culture was performed on the aliquots immediately after dispensing and 2 months later. Intravitreal Bevacizumab Injection and Vitrectomy Bevacizumab was injected by the following procedures. After administering oxybuprocaine hydrochloride ophthalmic solution (0.4% Benoxil, Santen Pharmaceutical Co., Osaka, Japan) as a local anesthetic, the conjunctiva was disinfected with 0.25% (1:40 dilution of neat solution) povidone-iodine (Isodine, Meiji Seika Pharma Co., Ltd., Tokyo, Japan).22 After placing the lid speculum, the conjunctiva was again disinfected with 0.25% povidone-iodine, and bevacizumab was injected intravitreally at a dose of 0.16 mg/0.05 mL. In the sham control group, after administering oxybuprocaine hydrochloride ophthalmic solution, a needleless syringe was pressed against the conjunctiva. On the next day after bevacizumab or sham injection, vitrectomy was conducted using the Constellation Visual System (Alcon Surgical, Fort Worth, TX) with a 25-G vitrectomy system (Alcon Surgical). A trocar was inserted at a site 4 mm from the corneal limbus. With infusion arrested, a 25-G vitreous cutter was inserted into the center of the vitreous body, and 0.5 mL of vitreous fluid was collected into a 2.5-mL syringe connected to the aspiration line. Thereafter, vitrectomy was conducted. During surgery, bleeding from the neovascular membrane was controlled using
1802 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
a 25-G straight bipolar pencil (Kirwan Surgical Products, Marshfield, MA) or a 25-G straight laser probe. The numbers of intraoperative endodiathermy spots and endolaser spots were recorded. Immediately after collection, the vitreous sample was transferred to a test tube and stored at −70°C. Two to three days later, the sample was thawed immediately before assay, and vitreous VEGF concentration was measured using an enzyme-linked immunosorbent assay for human VEGF (R&D Systems, Minneapolis, MN). Postoperative Follow-up Patients were examined daily from 1 to 7 days and at 2 weeks and 1 month after surgery. At each followup, all patients underwent a full ophthalmologic examination, intraocular pressure (IOP) measurement, slit-lamp biomicroscopy, and indirect ophthalmoscopy. All postoperative examinations were performed by three vitreoretinal surgeons (H.S., T.H., H.N.) who also were masked to the study group allocation. Intraocular pressure was measured at every follow-up examination. During the 4-week period, any increase of IOP to $25 mmHg was defined as elevated IOP. Corrected visual acuity was tested after 1 month. Visual acuity was measured by an optometrist who was masked to the randomization. Visual acuity was measured using the Landolt ring chart, and the result was expressed in Snellen visual acuity ratio and converted to logarithm of minimal angle of resolution (logMAR) scale for statistical analysis. Improvement of best-corrected visual acuity (BCVA) was defined as the difference between postoperative and preoperative BCVA and expressed in letters of improvement (Early Treatment Diabetic Retinopathy Study chart) as well as change in logMAR for statistical analysis. Postoperative recurrent VH was defined as bleeding that rendered the optic disk invisible. In eyes with silicon infusion, postoperative recurrent hemorrhage was defined as bleeding of 1 disk diameter or greater. At the end of vitrectomy, we confirmed that there was no retinal hemorrhage in all cases. Therefore, retinal hemorrhage occurring after vitrectomy was considered to be “recurrent VH” and not “persistent VH.” All patients with PDR were hospitalized for the primary vitrectomy and 1-week postoperative observation, or longer if necessary. To prevent hospitalization from becoming long term, we performed repeat vitrectomy in patients who developed severe recurrent VH with no fundus details for over 5 days and showing no tendency of resorption. We did not inject bevacizumab 0.16 mg for the purpose of promoting resorption of retinal hemorrhage and wait for spontaneous resorption to take place.
2015 VOLUME 35 NUMBER 9
Outcome Measures The following were evaluated during surgery: vitreous VEGF concentration, number of intraoperative endodiathermy spots, number of intraoperative endolaser spots, iatrogenic retinal tears, type of tamponade, and surgical time. The postoperative parameters evaluated included early postoperative (within 4 weeks) recurrent VH, reoperation due to recurrent VH, postoperative (within 4 weeks) IOP, persistence or progression of NVG, 1-month postoperative visual acuity, and degree of visual improvement. Primary outcome measure was the frequency of reoperation due to postoperative recurrent VH within 4 weeks after vitrectomy. Secondary outcome measures were numbers of intraoperative laser and endodiathermy spots, frequency of postoperative recurrent VH within 4 weeks after vitrectomy, and concentration of VEGF in vitreous at the beginning of surgery. Statistics Sample size was estimated using reported incidence of recurrent VH because we could not find reports that examined the rate of reoperation for recurrent VH, which was the primary outcome measure. With study power of 90%, significance level set at 0.05, and an assumed decrease in incidence of early postvitrectomy VH from 60% to 30% with preoperative IVB, a sample size of 30 eyes for each arm was calculated. Data are expressed as mean ± SD or percentage of patients. Chi-square test or Fisher’s exact test was used to analyze categorical variables. The frequencies of postoperative recurrent VH and reoperation due to postoperative recurrent VH were analyzed using the chi-square test to compare the differences between two groups. The frequency of intraoperative endodiathermy use, concentration of VEGF in vitreous, postoperative BCVA, and the improvement of postoperative BCVA were analyzed using the Mann– Whitney rank-sum test. Statistical analyses were performed using SPSS version 21 (SPSS, Inc., Chicago, IL). A P value less than 0.05 was considered statistically significant.
Results Participant Flow Patient flow is shown in Figure 1. Primary vitrectomy was conducted in a total of 68 eyes during the study period. Two patients (2 eyes) did not give informed consent and were excluded from the study.
1803
IVB INJECTION MANABE ET AL
Fig. 1. Patient flow.
Sixty-six eyes (62 patients) were randomized to IVB group (32 eyes of 32 patients) or sham control group (34 eyes of 30 patients). All patients completed follow-up for 1 month. Eventually, 32 eyes (32 patients) in IVB group and 34 eyes (30 patients) in sham control group were analyzed. Baseline Data Baseline patient data are shown in Table 1. The two groups did not differ significantly in age; BCVA; IOP;
history of hypertension, hyperlipidemia, or cerebral or myocardial infarction; glycated hemoglobin (HbA1c); use of anticoagulant; lens status; persistent VH; status of iris neovascularization or NVG; status of traction retinal detachment; and previous panretinal photocoagulation (PRP). Traction retinal detachment was defined as retinal detachment of 1 disk diameter or greater. Previous PRP was classified as “none” when the patient did not undergo PRP before surgery, and as “incomplete” when PRP conducted before vitrectomy was inadequate and additional PRP was given during
Table 1. Baseline Demographic and Clinical Data of All Patients
Age, mean ± SD (range), years Gender, male/female Preoperative BCVA, logMAR ± SD (Snellen visual acuity ratio) IOP, mean ± SD (range), mmHg Hypertension, n (%) Hyperlipidemia, n (%) Cerebral and myocardial infarction, n (%) HbA1c, mean ± SD, % Use of anticoagulants, n (%) Lens status, phakic/pseudophaki Preoperative complications, n (%) Persistent VH INV or NVG TRD Previous PRP, n (%) None Incomplete
IVB Group (n = 32)
Sham Control Group (n = 34)
P
59.9 ± 11.8 (20–79) 22/10 1.09 ± 0.60 (20/100)
59.2 ± 12.9 (35–82) 32/2 1.14 ± 0.58 (20/120)
0.667* 0.008† 0.756*
15.3 ± 2.69 (9–21) 14 (43.8) 5 (15.6) 5 (15.6) 8.0 ± 1.7 7 (21.9) 28/4
14.6 ± 3.42 (7–25) 16 (47.1) 7 (20.6) 6 (17.6) 7.6 ± 2.0 8 (23.5) 31/3
0.157* 0.787† 0.601† 0.826† 0.500* 0.873† 0.465†
29 (90.6) 2 (6.3) 5 (1.5)
30 (88.2) 2 (5.9) 4 (1.2)
0.535† 0.670† 0.460†
9 (28.1) 23 (71.9)
11 (32.4) 23 (67.6)
0.709† 0.709†
*Mann–Whitney rank-sum test. †Chi-square test. HbA1c, hemoglobin A1c; INV, iris neovascularization; logMAR, logarithm of minimal angle of resolution.
1804 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
vitrectomy. The only difference between two groups was a higher male/female ratio in the sham control group. Outcome Measures The results of intraoperative and postoperative parameters are summarized in Table 2. The frequency of reoperation due to postoperative (within 4 weeks) recurrent VH was 20.6% (7/34 eyes) in the sham control group and 3.1% (1/32 eyes) in the IVB group and was significantly lower in the IVB group (P = 0.033). The number of endolaser spots was not different between IVB group and sham control group (1260 ± 686 vs. 1335 ± 663, P = 0.667). The number of endodiathermy spots was 1.3 ± 1.4 in the sham control group and 0.63 ± 1.0 in the IVB group and was significantly lower in the IVB group (P = 0.025). This difference, while statistically significant, may not be clinically relevant because both represent one spot of diathermy. The incidence of postoperative (within 4 weeks) recurrent VH was 23.5% (8/34 eyes) in the sham control group and 3.1% (1/32 eyes) in the IVB group and was also significantly lower in the IVB group (P = 0.017). Among eyes that had silicon oil infusion (6 in sham control group and 4 in IVB group), 2 eyes in the sham control group (2/6, 33.3%) and 0 eye in the IVB group (0/4, 0%) had postoperative recurrent VH, with no significant difference between 2 groups (P = 0.467).
2015 VOLUME 35 NUMBER 9
Vascular endothelial growth factor concentration in vitreous was 25.0 ± 13.6 pg/mL in the IVB group and 1315.3 ± 1153.4 pg/mL in the sham control group and was significantly lower (P , 0.0001) in the IVB group. Other Clinical Results There were no significant differences between 2 groups in other intraoperative parameters such as intraoperative laser treatment, the frequency of iatrogenic retinal breaks, the frequencies of using gas and silicon oil tamponade, and surgical time (Table 2). Postoperative 1-month BCVA (including reoperated eyes) improved significantly in both IVB group (logMAR: 1.09 ± 0.60 to 0.46 ± 0.54, P , 0.001; Snellen ratio: 20/100 to 20/35, Early Treatment Diabetic Retinopathy Study: 20 letters of improvement) and sham control group (logMAR: 1.14 ± 0.58 to 0.43 ± 0.48, P , 0.001; Snellen ratio: 20/120 to 20/35, Early Treatment Diabetic Retinopathy Study: 25 letters of improvement). However, postoperative BCVA and visual improvement were not significantly different between 2 groups (P = 0.929 and 0.445, respectively). There were no significant differences in the proportions of patients showing postoperative (within 4 weeks) IOP elevation and persistence or progression of NVG (Table 2). Intraoperative iatrogenic retinal breaks occurred in 5 of 32 eyes in the IVB group and 5 of 34 eyes in the sham control group. In all these
Table 2. Vitreous VEGF Level, Intra/Postoperative Outcomes and Adverse Events Examined
Intraoperative outcomes VEGF concentration in vitreous, mean ± SD, pg/mL Endolaser spots, mean ± SD, n Endodiathermy spots, mean ± SD, n Iatrogenic retinal breaks, n (%) Gas tamponade (air or long acting gas), n (%) Silicon oil tamponade, n (%) Surgical time, mean ± SD, minutes Postoperative outcomes Postoperative VH, n (%) Vitreoretinal reoperation due to recurrent VH, n (%) Elevation of IOP, n (%) Progressive or persistent NVG, n (%) Postoperative BCVA, logMAR ± SD (Snellen visual acuity ratio) Mean BCVA change, logMAR ± SD (ETDRS chart) Injection-related adverse events Systemic thromboembolic event Endophthalmitis
IVB Group (n = 32)
Sham Control Group (n = 34)
P
25.0 ± 13.6 1260 ± 686 0.63 ± 1.0 5 (15.6) 23 (71.9) 4 (12.5) 49 ± 20
1315.3 ± 1153.4 1335 ± 663 1.3 ± 1.4 5 (14.7) 24 (70.6) 6 (17.6) 56 ± 27
,0.0001* 0.667* 0.025* 0.593† 0.670† 0.378† 0.298*
1 (3.1) 1 (3.1) 2 (6.3) 0 (0) 0.46 ± 0.54 (20/35)
8 (23.5) 7 (20.6) 6 (17.6) 3 (8.8) 0.43 ± 0.48 (20/35)
0.017† 0.033† 0.149† 0.131‡ 0.929*
−0.63 ± 061 (20 letters)
−0.73 ± 0.80 (25 letters)
0.445*
0 0
0 0
*Mann–Whitney rank-sum test. †Chi-square test. ‡Fisher’s exact test. ETDRS, Early Treatment Diabetic Retinopathy Study; logMAR, logarithm of minimal angle of resolution; SO, silicone oil.
.0.999‡ .0.999‡
IVB INJECTION MANABE ET AL
eyes, the retinal breaks were treated by intraoperative endolaser photocoagulation, and retinal reattachment was confirmed during vitrectomy. No bacteria were isolated from the aliquoted bevacizumab. Injection-related Adverse Events No ocular and no systemic adverse events associated with IVB such as systemic thromboembolic event and endophthalmitis were observed in the present study (Table 2). Discussion In the present prospective, randomized, controlled clinical trial conducted in patients undergoing vitrectomy for PDR, we blocked VEGF in the vitreous by injecting bevacizumab at a dose (0.16 mg/0.05 mL) equivalent to one eighth of the usual dose 1 day before vitrectomy. Vitreous VEGF level was markedly reduced in patients given intravitreal bevacizumab. Since vitreous samples were frozen immediately after collection and stored at −70°C until assay, there is little possibility that residual bevacizumab in the sample binds with VEGF resulting in exaggerated reduction in active VEGF when measured using the enzymelinked immunosorbent assay. The primary outcome measure; frequency of vitreoretinal reoperation due to postoperative recurrent VH, and the secondary outcome measures; frequencies of intraoperative endodiathermy use and postoperative recurrent VH, clearly demonstrated a significantly beneficial effect of the IVB in preventing VH during and after vitrectomy. Two meta-analyses on preoperative IVB (1.25 mg/ 0.05 mL) for severe PDR have been reported.23,24 Both analyses led to a common conclusion that adjuvant IVB before vitrectomy in patients with PDR significantly decreased intraoperative bleeding, decreased the frequency of intraoperative endodiathermy, shortened the surgical time, and reduced postoperative (within 1 month) VH compared with controls.23,24 One of the two analyses also reported a lower frequency of intraoperative tear,24 shorter postoperative reabsorption time of vitreous bleeding, lower incidence of late (after 1 week) IOP elevation,23 and better final BCVA23 in the IVB group than in the control group. However, early postoperative (within 1 week) IOP elevation,23 BCVA at 3 months after surgery, late postoperative (after 1 month) VH, rate of vitreoretinal reoperation,24 rate of postoperative retinal detachment,23,24 and rate of NVG24 were not significantly different between two groups. Our study showed that intravitreal injection of bevacizumab at a dose of only 0.16 mg/0.05 mL
1805
reduced the frequency of reoperation due to postoperative recurrent hemorrhage. Although vitrectomy on the next day rapidly eliminates the bevacizumab injected into the vitreous, the low-dose (0.16 mg/0.05 mL) bevacizumab injected 1 day before vitrectomy probably penetrates into the retina to prevent postoperative recurrent hemorrhage. Indeed, previous reports indicate that intravitreally injected bevacizumab penetrates into the retina within 1 day25,26 and passes into systemic circulation.27,28 Avery et al29 injected bevacizumab at concentrations of 6.2 to 625 mg intravitreally in patients with retinal and iris neovascularization secondary to diabetes mellitus and observed reduction of fluorescein angiographic leakage 24 hours after injection. Therefore, we assumed that intravitreal injection of bevacizumab 0.16 mg results in penetration of the drug through the retina and exhibition of anti-VEGF effect after 24 hours. By reducing intraoperative bleeding and decreasing the frequency of hemostasis procedure during surgery, IVB injection before surgery facilitates the surgical procedures. Moreover, the decreases in frequency of postoperative recurrent VH and frequency of reoperation clearly indicate that this approach is beneficial to patients. In this study, visual acuity at 1 month after surgery improved markedly in both groups, but postoperative visual acuity and postoperative visual improvement did not differ between 2 groups. In patients with age-related macular degeneration, retinal vein occlusion, and diabetic macular edema, high doses of VEGF inhibitors are injected intravitreally to achieve a sustained therapeutic effect for 1 to 2 months. Serious adverse events potentially related to systemic VEGF inhibition, including cardiovascular events, cerebrovascular accidents, and death have been reported.19 Carneiro et al19 studied the risk of arterial thromboembolic events in 378 patients with exudative age-related macular degeneration treated with bevacizumab or ranibizumab. The incidence of arterial thromboembolic events was 12.4% (12/97 patients) with bevacizumab and 1.4% (3/219 patients) with ranibizumab, with odds ratio of 10.16. The methods reported in the literature for reducing the rates of early postvitrectomy recurrent hemorrhage involve intravitreal injection of 1.25 mg to 2.5 mg of the VEGF inhibitor bevacizumab 1 day to 20 days before vitrectomy (Table 3). Intravitreally injected bevacizumab at a dose of 1.25 mg has been reported to significantly reduce VEGF concentration in blood.30–32 Because of the vascular risk potentially associated with system VEFG inhibition, preoperative IVB injection was not indicated in patients with PDR with a history of cerebrovascular event or myocardial infarction.20 We used one eighth the normal
1806 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
2015 VOLUME 35 NUMBER 9
Table 3. Summary of Studies of Vitrectomy With or Without Preoperative IVB Injection for Proliferative Diabetic Retinopathy
Eyes, n
Dose of IVB
Interval Between IVB Injection and Vitrectomy
Rizzo et al5 El-Batarny6
22 30
1.25 mg 1.25 mg/0.05 mL
5–7 days 5–7 days
Modarres et al7
40
2.5 mg/0.1 mL
3–5 days
Ahmadieh et al8 da R Lucena et al9
68 20
1.25 mg 1.5 mg/0.06 mL
1 week 2 weeks
Hernandez-Da Mota and NuñezSolorio10 di Lauro et al11
40
1.25 mg/0.05 mL
2 days
72
1.25 mg
7 or 20 days
107
1.25 mg/0.05 mL
1–14 days
Farahvash et al13
35
1.25 mg
7 days
This study
66
0.16 mg
1 day
Ahn et al12
dose of the VEGF inhibitor bevacizumab and removed residual VEGF inhibitor in the vitreous the next day by conducting vitrectomy. This method may reduce the amount of VEGF inhibitor in systemic circulation and may potentially be safer than the conventional methods. In this study, we included patients with cerebrovascular event or myocardial infarction more than 3 months before the study and observed no systemic adverse events. This study is significant from the viewpoint of expanding the indication of this approach. Regarding ocular complications, Oshima et al14 reported progression of TRD after IVB injection, and Arevalo et al18 reported onset of TRD in 11 of 211 eyes (5.2%) at 13 days on average after IVB injection. The risk of exacerbation or development of TRD may increase as the duration between injection and surgery increases. In this study, 90% of the patients had VH before vitrectomy, and assessment of TRD was difficult in these patients. As a limitation of this study, the observation period was short, and further investigation of long-term outcome is needed. Vitrectomies were done by three surgeons in this study. More reliable results may have been obtained with one surgeon performing all the vitrectomies. Moreover, the decision of reoperation was probably influenced by various factors such as the severity of PDR, compliance with positioning, and minor differences in techniques between surgeons.
Main Outcome Feasibility of surgery Feasibility of surgery, postoperative complications BCVA, number of endodiathermy applications, backflush needle applications, duration of surgery, and postoperative VH Incidence of early postvitrectomy hemorrhage Erythrocyte count in the fluid retrieved from the vitrectomy cassette BCVA, intraocular pressure, and fundus photographs Intraoperative management, safety, efficacy of IVB Incidence of early (#4 weeks) and late (.4 weeks) recurrent VH Facility of surgery (severity of intraoperative bleedings and intraoperative break formation) Vitreous VEGF concentration, intraoperative hemorrhage, recurrent postoperative VH, reoperation due to recurrent VH
In addition, the present results obtained from a single center have to be validated in multiple institutions. Further study is required to examine whether plasma VEGF concentration is reduced by intravitreal injection of 0.16 mg bevacizumab. Key words: bevacizumab, intravitreal injection, postoperative recurrent vitreous hemorrhage, proliferative diabetic retinopathy, prospective randomized controlled clinical trial, vascular endothelial growth factor. References 1. Novak MA, Rice TA, Michels RG, Auer C. Vitreous hemorrhage after vitrectomy for diabetic retinopathy. Ophthalmology 1984;91:1485–1489. 2. Tolentino FI, Freeman HM, Tolentino FL. Closed vitrectomy in the management of diabetic traction retinal detachment. Ophthalmology 1980;87:1078–1089. 3. Lewis H, Abrams GW, Williams GA. Anterior hyaloidal fibrovascular proliferation after diabetic vitrectomy. Am J Ophthalmol 1987;104:607–613. 4. Schachat AP, Oyakawa RT, Michels RG, Rice TA. Complications of vitreous surgery for diabetic retinopathy II. Postoperative complications. Ophthalmology 1983;90:522–530. 5. Rizzo S, Genovesi-Ebert F, Di Bartolo E, et al. Injection of intravitreal bevacizumab (Avastin) as a preoperative adjunct before vitrectomy surgery in the treatment of severe proliferative diabetic retinopathy (PDR). Graefes Arch Clin Exp Ophthalmol 2008;246:837–842. 6. El-batarny AM. Intravitreal bevacizumab as an adjunctive therapy before diabetic vitrectomy. Clin Ophthalmol 2008;2:709–716.
IVB INJECTION MANABE ET AL 7. Modarres M, Nazari H, Falavarjani KG, et al. Intravitreal injection of bevacizumab before vitrectomy for proliferative diabetic retinopathy. Eur J Ophthalmol 2009;19:848–852. 8. Ahmadieh H, Shoeibi N, Entezari M, Monshizadeh R. Intravitreal bevacizumab for prevention of early postvitrectomy hemorrhage in diabetic patients: a randomized clinical trial. Ophthalmology 2009;116:1943–1948. 9. da R Lucena D, Ribeiro JA, Costa RA, et al. Intraoperative bleeding during vitrectomy for diabetic tractional retinal detachment with versus without preoperative intravitreal bevacizumab (IBeTra study). Br J Ophthalmol 2009;93:688–691. 10. Hernández-Da Mota SE, Nuñez-Solorio SM. Experience with intravitreal bevacizumab as a preoperative adjunct in 23-G vitrectomy for advanced proliferative diabetic retinopathy. Eur J Ophthalmol 2010;20:1047–1052. 11. di Lauro R, De Ruggiero P, di Lauro R, et al. Intravitreal bevacizumab for surgical treatment of severe proliferative diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol 2010; 248:785–791. 12. Ahn J, Woo SJ, Chung H, Park KH. The effect of adjunct intravitreal bevacizumab for preventing postvitrectomy hemorrhage in proliferative diabetic retinopathy. Ophthalmology 2011;118:2218–2226. 13. Farahvash MS, Majidi AR, Roohipoor R, Ghassemi F. Preoperative injection of intravitreal bevacizumab in dense diabetic vitreous hemorrhage. Retina 2011;31:1254–1260. 14. Oshima Y, Shima C, Wakabayashi T, et al. Microincision vitrectomy surgery and intravitreal bevacizumab as a surgical adjunct to treat diabetic traction retinal detachment. Ophthalmology 2009;116:927–938. 15. Inan UU, Avci B, Kusbeci T, et al. Preclinical safety evaluation of intravitreal injection of full-length humanized vascular endothelial growth factor antibody in rabbit eyes. Invest Ophthalmol Vis Sci 2007;48:1773–1781. 16. Peters S, Heiduschka P, Julien S, et al. Ultrastructural findings in the primate eye after intravitreal injection of bevacizumab. Am J Ophthalmol 2007;143:995–1002. 17. Shimomura Y, Hirata A, Ishikawa S, Okinami S. Changes in choriocapillaris fenestration of rat eyes after intravitreal bevacizumab injection. Graefes Arch Clin Exp Ophthalmol 2009; 247:1089–1094. 18. Arevalo JF, Maia M, Flynn HW Jr., et al. Tractional retinal detachment following intravitreal bevacizumab (Avastin) in patients with severe proliferative diabetic retinopathy. Br J Ophthalmol 2008;92:213–216. 19. Carneiro AM, Barthelmes D, Falcão MS, et al. Arterial thromboembolic events in patients with exudative age-related macular degeneration treated with intravitreal bevacizumab or ranibizumab. Ophthalmologica 2011;225:211–221. 20. Hattori T, Shimada H, Nakashizuka H, et al. Dose of intravitreal bevacizumab (Avastin) used as preoperative adjunct
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
1807
therapy for proliferative diabetic retinopathy. Retina 2010;30: 761–764. Brown DM, Campochiaro PA, Singh RP, et al. Ranibizumab for macular edema following central retinal vein occlusion: sixmonth primary end point results of a phase III study. Ophthalmology 2010;117:1124–1133. Shimada H, Hattori T, Mori R, et al. Minimizing the endophthalmitis rate following intravitreal injections using 0.25% povidone-iodine irrigation and surgical mask. Graefes Arch Clin Exp Ophthalmol 2013;251:1185–1190. Zhao LQ, Zhu H, Zhao PQ, Hu YQ. A systematic review and meta-analysis of clinical outcomes of vitrectomy with or without intravitreal bevacizumab pretreatment for severe diabetic retinopathy. Br J Ophthalmol 2011;95:1216–1222. Zhang ZH, Liu HY, Hernandez-Da Mota SE, et al. Vitrectomy with or without preoperative intravitreal bevacizumab for proliferative diabetic retinopathy: a meta-analysis of randomized controlled trials. Am J Ophthalmol 2013;156:106–115. Heiduschka P, Fietz H, Hofmeister S, et al. Penetration of bevacizumab through the retina after intravitreal injection in the monkey. Invest Ophthalmol Vis Sci 2007;48: 2814–2823. Julien S, Heiduschka P, Hofmeister S, Schraermeyer U. Immunohistochemical localisation of intravitreally injected bevacizumab at the posterior pole of the primate eye: implication for the treatment of retinal vein occlusion. Br J Ophthalmol 2008; 92:1424–1428. Bakri SJ, Snyder MR, Reid JM, et al. Pharmacokinetics of intravitreal bevacizumab (Avastin). Ophthalmology 2007; 114:855–859. Matsuyama K, Ogata N, Matsuoka M, et al. Effects of intravitreally injected bevacizumab on vascular endothelial growth factor in fellow eyes. J Ocul Pharmacol Ther 2011;27: 379–383. Avery L, Pearlman J, Pieramici J, et al. Intravitreal bevacizumab (Avastin) in the treatment of proliferative diabetic retinopathy. Ophthalmology 2006;113:1695.e1-15. Zehetner C, Kirchmair R, Huber S, et al. Plasma levels of vascular endothelial growth factor before and after intravitreal injection of bevacizumab, ranibizumab and pegaptanib in patients with age-related macular degeneration, and in patients with diabetic macular oedema. Br J Ophthalmol 2013;97: 454–459. Carneiro AM, Costa R, Falcão MS, et al. Vascular endothelial growth factor plasma levels before and after treatment of neovascular age-related macular degeneration with bevacizumab or ranibizumab. Acta Ophthalmol 2012;90:e25–e30. Matsuyama K, Ogata N, Matsuoka M, et al. Plasma levels of vascular endothelial growth factor and pigment epitheliumderived factor before and after intravitreal injection of bevacizumab. Br J Ophthalmol 2010;94:1215–1258.
