Introduction

Central retinal vein occlusion (CRVO), a common retinal vascular disorder, remains an important cause of painless visual loss in patients of age more than 50 years.1 Macular edema is a common cause of visual loss in these patients. At this time, there is insufficient level I evidence to support any specific treatment to improve vision in CRVO.2 However, a variety of interventions are available such as systemic anticoagulation, panretinal photocoagulation (PRP), macular grid photocoagulation, hemodilution, laser chorioretinal venous anastomosis and radial optic neurotomy/laminar puncture, and so on.

Intravitreal levels of vascular endothelial growth factor (VEGF) in CRVO are the highest of those measured in retinal vascular disease.3 Various studies have confirmed that intravitreal injection of bevacizumab (a pan VEGF blocker) improves the visual acuity in these patients by virtue of reduction in macular edema and reversal of fundus changes.4, 5, 6, 7

Bevacizumab treatment early after the onset of CRVO was associated with statistically significant reduction in venous dilation, tortuosity, optic disc swelling and macular edema, and a dramatic improvement in visual acuity.8, 9 However, need of repeated injections at regular intervals to maintain the visual gain has been a major drawback. Although, central retinal vein occlusion study (CVOS) showed benefit of PRP for proliferative stage of CRVO, iris neovascularisation developed less frequently in prophylactically treated eyes as compared with untreated eyes, although statistically insignificant.10 Also, in CVOS although macular grid photocoagulation did not lead to significant improvement in visual acuity it significantly decreased angiographic macular edema at 12 month follow-up.11 Thus suggesting that retinal photocoagulation has a stabilizing role to play.

In this study, we combined the beneficial effect of early treatment with intravitreal bevacizumab and the stabilizing effect of laser to treat our patients in an attempt to obviate the need for repeated intravitreal injections.

Materials and methods

It was a prospective non-randomized clinical interventional study of nine eyes of nine consecutive patients with recent onset clinically diagnosed and angiographically proven CRVO with macular edema.

The patients were informed about the experimental nature of the treatment and had to sign an informed written consent form.

All patients of CRVO with macular edema and decreased visual acuity irrespective of degree of ischaemia were included in the study. Patients with more than 1 month duration of onset, history of previous treatment for CRVO, other retinal disease, such as diabetic retinopathy likely to affect visual acuity and less than 1 year follow-up, were excluded.

At presentation, patients underwent complete ophthalmic evaluation, which included Snellen's visual acuity assessment, slit-lamp biomicroscopy and indirect ophthalmoscopy. Digital fundus fluorescein angiography (FFA, Carl Zeiss Meditec AG, Jena, Germany) was done to confirm the diagnosis, classify the type of CRVO (according to the CRVO study criteria), and evaluate the presence of macular edema.

All the patients were treated with 2.5 mg (0.1 ml) of intravitreal injection of bevacizumab (Genentech, San Diego, CA, USA) on presentation. The injection was performed in the operation theatre under strict aseptic precautions.

The patients were examined on day 1, day 7, and 3 weeks postoperatively. At each visit routine evaluations, which included Snellen's best corrected visual acuity (BCVA), intraocular pressure measurement by applanation tonometry, slit-lamp biomicroscopy and indirect ophthalmoscopy, were done. The visual acuity measurements were converted to logarithm of the minimum angle of resolution (LOGMAR).

Panretinal laser photocoagulation with macular grid laser, (average 3600 shots over three sittings with 200–300 mW power, 200–300-μ spot size for panretinal laser and 100 mW power with 50-μ spot size for macular grid laser) was performed in all patients starting at 3 weeks after the injection, over three sittings, on alternate days in all cases.

Thereafter, all the patients were followed up on a monthly basis for 6 months and later on a 3-month basis for 1 year, or last follow-up visit whichever was later. Besides routine evaluation, FFA was to be done at each such visit if there was clinical suspicion of recurrence of macular edema or to detect any neovascular complication in the event of conversion from non-ischemic to ischaemic CRVO.

Main outcome measures included improvement in visual acuity and fundus appearance. Conversion from non-ischemic to ischaemic CRVO, recurrence of macular edema, disc collateral formation, and need for repeat injections were also studied.

Results

It was a prospective study of nine eyes of nine patients with CRVO. Four out of nine (44.44%) eyes had non-ischemic CRVO, 5/9 (55.55%) eyes had ischaemic CRVO. Overall, 8/9 (88.88%) patients were males and only 1/9 (11.11%) patients was female, with age ranging from 42 to 70 years (mean age 54 years). Among the systemic predisposing factors, hypertension alone was present in 4/9 patients (44.44%), pulmonary thromboembolism in 1/9 patients (11.11%) and hyperhomocysteinemia in 2/9 patients (22.22%). Whereas hypertension combined with diabetes was present in 1/9 (11.11%) and with hyperhomocysteinemia in 1/9 (11.11%) patients. All patients presented within 10 days of onset of symptoms (average 2.67 days).

Presenting BCVA was <20/200 in 5/9 (55.55%) and 20/50 to 20/200 in 4/9 (44.44%) eyes. All patients were treated with intravitreal bevacizumab within <7 days of presentation (average 3.1 days). Postoperatively, none of the patients had any local or systemic side effects of injection such as endophthalmitis, clinically evident inflammation, retinal tears, retinal detachment, or thromboembolic events. Conjunctival hyperaemia was present in all and subconjunctival hemorrhage in 3/9 (33.33%) eyes.

All the patients showed rapid improvement in the form of rapid clearance of retinal hemorrhages, decreased optic disc swelling and decreased venous dilation, and tortuosity. This was clinically evident as early as at the 7 day after injection follow-up in all the patients. Figure 1a shows fluorescein angiogram on presentation in one of the study patients with CRVO showing severe dilation and tortuosity of venous circulation, blocked fluorescence due to superficial retinal hemorrhages, disc hyperfluorescence and macular leakage; Figure 1b shows the post treatment fluorescein angiogram in the same patient with complete resolution of macular edema. Scars of laser photocoagulation are seen.

Figure 1
figure 1

(a) Shows fluorescein angiogram on presentation in one of the study patient with CRVO showing severe dilation and tortuosity of venous circulation, blocked fluorescence due to superficial retinal hemorrhages, disc hyperfluorescence, and macular leakage; (b) shows Post-treatment fluorescein angiogram in the same patient with complete resolution of macular edema. Scars of laser photocoagulation are seen.

Full size image

Final post-operative vision was <20/200 in 2/9 (22.22%) eyes and >20/40 in 7/9 (77.78%) eyes. Mean baseline visual acuity (VA) improved from 20/320 (LOGMAR units, 1.2) to 20/63 (LOGMAR units, 0.5) at the final follow-up visit. (Figure 2 shows the plot of change in BCVA against time in all patients, till the last follow-up visit and Table 1 shows summary of patient data).

Figure 2
figure 2

Graph shows the plot of change in best corrected visual acuity (Log MAR) against time in all patients.

Full size image
Table 1 Summary of patient data
Full size table

Discussion

CRVO, a common retinal vascular disorder, remains an important cause of painless visual loss in patients more than 50 years of age.1 The clinical appearance typically demonstrates four quadrants of intraretinal hemorrhages with dilated and tortuous retinal veins. Macular edema, optic disc edema, and cotton-wool spots may be present to a variable degree. CRVO is broadly divided into two clinical subtypes, based on the degree of ischaemia: non-ischemic CRVO is typically associated with relatively better vision and a better prognosis for spontaneous visual improvement; ischaemic CRVO is typically associated with more profound visual loss on presentation, a relative afferent pupillary defect, and a relatively higher risk for neovascular glaucoma.

The most common cause of visual loss in patients with CRVO is macular edema. Other causes of visual loss include macular ischaemia and neovascular glaucoma.

A variety of interventions have been implicated to improve vision in CRVO.2 Especially, the off-label use of intravitreal injection of bevacizumab for macular edema has been studied extensively.4, 5, 6, 7

Bevacizumab is a full-length recombinant humanized antibody, active against VEGF, and approved for use in colorectal cancer.12 Off-label intravitreal injection of bevacizumab was first reported as a potential therapy for macular edema secondary to CRVO in 2005.13 In several case series, bevacizumab was reported to improve macular edema associated with CRVO, at least in the short term.4, 5, 6, 7 Patients experienced a dramatic improvement in the visual acuity and clinical fundus appearance, without collateral vessel formation.9

In addition, bevacizumab appears to have activity against anterior segment neovascularization. This approach is used most often in cases of non-ischemic CRVO, but bevacizumab may be effective to treat complications of ischaemic CRVO as well. The visual benefits of intravitreal bevacizumab for macular edema due to CRVO are apparent early but are not sustained without need for repeated injections. A larger clinical study and long-term follow-up will be necessary to better elicit the best regimen for this therapy.14

CVOS was a randomized clinical trial wherein eyes with at least 10 disc areas of retinal capillary non-perfusion, prophylactic scatter photocoagulation did not prevent the development of anterior segment neovascularization. The CVOS concluded that it was safe to observe these eyes until early neovascularization of the iris or angle developed. Neovascularization regressed after treatment in 56% of photocoagulation-naive eyes and 22% of eyes treated previously with scatter photocoagulation.10

Iris neovascularisation developed less frequently in prophylactically treated eyes as compared with untreated eyes, although statistically insignificant. Iris or angle neovascularisation developed in 117/714 (16%) of eyes. A total of 61/117 (52%) eyes that had iris or angle neovascularisation were initially categorized as perfused.15 Also one-third patients with non-ischemic CRVO converted into ischaemic CRVO. On this basis, all our patients received prophylactic PRP.

Recurrent macular edema may occur in patients with CRVO following treatment with bevacizumab; in some cases, the recurrent macular edema may be more severe than the pretreatment macular edema (‘rebound’ macular edema).16 Macular grid laser was given to all our patients to counter macula edema, which was the main cause of vision loss, and also the reason for reappearance of macular edema was the main indication for requirement of repeated injections. In CVOS, eyes with perfused macular edema on FFA and a visual acuity of 20/50 or worse were randomized to treatment with macular grid laser photocoagulation or observation only (no treatment). Although macular grid photocoagulation did not lead to statistically significant improvement in visual acuity, it significantly decreased angiographic macular edema. At the 12 month follow-up evaluation, all 72 untreated eyes (100%) had persistent angiographic leakage while 21/68 (31%) treated eyes had no quantifiable leakage (P<0.0001).11 In younger individuals, grid laser was suggested to be value in achieving better visual outcome. Our study patients were relatively younger with mean age of 54 years.

We combined the beneficial effect of intravitreal bevacizumab for vision gain with the stabilizing effect of laser to maintain the gain. Although, prophylactic PRP and macular grid photocoagulation did not show any statistically significant benefit in CRVO study, we included it in our protocol owing to its vision-stabilizing role especially to compensate for the short lasting effect of bevacizumab in the hope to reduce re-treatments for recurrent macular edema and to prevent neovascular complications.

According to CVOS, the BCVA at presentation was demonstrated to serve as an indicator of the final visual outcome.15 Although there is a vast difference between the sample size of CVOS and current study, a comparison of final visual outcome is notable. Among 209/714 eyes with the presenting BCVA of >20/40, 136 (65%) retained the vision, 52 (25%) deteriorated to 20/50 to 20/200, and 21 (10%) attained <20/200 BCVA. However, none of our patients had a presenting visual acuity >20/40.

In CVOS, of the 304/714 eyes with presenting BCVA of 20/50 to 20/200 58 (19%) eyes improved to >20/40, 133 (44%) eyes remained stable, and deterioration to <20/200 BCVA was seen in 113 (37%) eyes. In our study, all (100%) patients in this group improved and attained >20/40 BCVA.

In CVOS, of the 201/714 eyes with presenting BCVA <20/200 (most eyes), 564 (79%) eyes retained poor visual acuity, 39 (19%) eyes improved to 20/50 to 20/200 and only three (1%) eyes improved to >20/40. In our study, 3/5 (60%) developed >20/40 visual acuity and 2/5 (20%) patients remained at <20/200.

Mean baseline VA improved from 20/320 (LOGMAR units, 1.2) to 20/63 (LOGMAR units, 0.5) at the final follow-up visit. This was comparable or better than the results of other early intervention studies.8, 9

Although visual acuity did not improve significantly in all our patients, a significant reversal of anatomical features of CRVO was seen in all the patients. This suggests that besides the anti-permeability property of bevacizumab there exists an unknown mechanism, within the retinal venous system other than the thrombus, to be inhibited by bevacizumab.9

Collaterals are alternate pathways of venous outflow to compensate for poor venous outflow through the central vein. Collateral vessels at the optic nerve were seen to develop in ∼77% of CRVO cases in a study by Takahashi et al.17 None of our patient had visible collateral formation. Absence of collaterals in these patients implies that the treatment must have led to an improvement in the venous flow.9

In CVOS, at final follow-up visit, 34% of eyes showed conversion of non-ischemic to ischaemic CRVO.15 However, in our study none of the patients showed conversion.

This series, although prospective, has many drawbacks. The sample size was small. This was because it was a pilot study, also as the nature of treatment was prophylactic there was limited patient enrolment and it was a single center study. There was a lack of control group and masking of investigators and patients. Optical coherence tomography (OCT) was not done due to non-availability. However, FFA could not be deemed as ineffective for monitoring of macular edema, but for its invasive nature. The classification of patients into non-ischemic and ischaemic CRVO on the basis of FFA may be less reliable as in the acute stage extensive fundus hemorrhages may mask the actual areas of non-perfusion.

However, these drawbacks cannot undermine the results of this study especially when the early natural history of the disease that is, within 3 months has seldom being studied in detail. It suggests that early treatment with bevacizumab may have twofold effects in the form of early reversal of venous tortuosity and macular edema. Second, it improves the flow assuming that it inhibits VEGF which is known to cause decrease venous flow and retinal perfusion.8, 9

There are many salient features of this study. It includes early treatment of patients (within 10 days of presentation); the aim was to limit the irreversible damage by early intervention, no collateral formation, and better visual outcome compared with natural course. Most striking finding of the study was lack of necessity to re-inject bevacizumab. It also suggests a stabilizing role for PRP in an eye previously treated with an anti-VEGF agent. As there is no comparative study, the reason behind this cannot be ascertained. However, we propose that it would have had a sustained effect on the VEGF suppression induced by bevacizumab. It could have also prevented conversion of non-ischemic to ischaemic CRVO, occurrence of neovascular complications, and recurrence of macular edema.

In conclusion, the results of the present study suggests that early treatment with intravitreal bevacizumab followed by PRP may provide visually and anatomically favourable results in a case of CRVO and also obviate the need for repeated injection. This however, requires a randomized study with large sample size to substantiate the results.