To determine the percentage of ranibizumab-treated patients with retinal vein occlusion (RVO) who had resolution of edema for at least 6 months after the last injection, along with factors and outcomes that correlate with resolution.
Post hoc analysis of open-label clinical trial.
Twenty patients with branch RVO (BRVO) and 20 with central RVO (CRVO) received ranibizumab monthly for 3 months and as needed for recurrent/persistent macular edema, no more frequently than every 2 months. Patients still requiring injections after month 40 received scatter and grid laser photocoagulation to try to reduce the need for injections. Main outcome measures included the percentage of patients who had resolution of edema, change in best-corrected visual acuity (BCVA) from baseline, and change in area of retinal nonperfusion in central subfields.
Nine patients with BRVO (45%) had edema resolution from injections alone after a mean of 20.2 months, 4 resolved after addition of laser, 4 were unresolved through 72 months, and 3 exited prior to resolution. Five patients with CRVO (25%) resolved from injections alone after a mean of 14.0 months, 8 remained unresolved through 72 months despite addition of laser, and 7 exited prior to resolution. For BRVO or CRVO, there was a negative correlation between posterior retinal nonperfusion area and BCVA at months 18, 24, and 36 ( P < .05).
In patients with RVO, infrequent ranibizumab injections to control edema may not be sufficient to prevent progression of retinal nonperfusion, which may contribute to loss of visual gains.
Retinal vein occlusion (RVO) consists of central RVO (CRVO) and branch RVO (BRVO), which are prevalent retinal vascular diseases. In patients with CRVO, thrombosis of the main outflow vessel of the retina results in variable amounts of hemorrhage, edema, and retinal nonperfusion throughout the retina, whereas in patients with BRVO there is occlusion of a proximal branch of the central retinal vein that results in similar findings throughout about half of the retina. Hemorrhages and edema were assumed to be attributable to elevated intraluminal pressure, but the development of ranibizumab (Lucentis; Genentech, Inc, South San Francisco, California, USA), a Fab fragment that specifically binds all isoforms of vascular endothelial growth factor A (VEGF), made it possible to test the hypothesis that VEGF contributes to the edema, and the hypothesis was found to be correct. This has been confirmed by the Treatment of Macular Edema following Branch Retinal Vein Occlusion: Evaluation of Efficacy and Safety (BRAVO) and Treatment of Macular Edema following Central Retinal Vein Occlusion: Evaluation of Efficacy and Safety (CRUISE) trials. In BRAVO, monthly injections of 0.3 or 0.5 mg ranibizumab for 6 months resulted in gains in mean letter score from baseline BCVA of 16.6 and 18.3, compared to 7.3 in the sham injection group. In CRUISE, at the month 6 primary endpoint, there was a mean improvement from baseline in BCVA letter score of 12.7 (0.3 mg) and 14.9 (0.5 mg) vs 0.8 (sham). These gains were maintained between months 6 and 12 by intermittent injections of ranibizumab for recurrent/persistent edema. Thus, in patients with RVO, blockade of VEGF with ranibizumab reduces edema and improves vision. Monthly injections of ranibizumab also resulted in more rapid resolution of retinal hemorrhages than sham injections, indicating that high levels of VEGF contribute to retinal hemorrhages as well as macular edema.
These outcomes are outstanding, and it was hoped that they would be maintained and the need for injections would be eliminated as collaterals developed and circumvented the obstruction. However, 2-year follow-up of patients entered in the original ranibizumab study showed that many patients with BRVO or CRVO still required injections for chronic, recurrent macular edema and that injection only as frequently as every 2 months allowed maintenance of early visual gains in patients with BRVO, but was associated with some reduction in visual gains in patients with CRVO. Findings were similar in the much larger HORIZON trial, in which after 12 months of treatment with ranibizumab in BRAVO and CRUISE, patients with RVO were seen at least every 3 months and given an injection of ranibizumab for recurrent/persistent macular edema. These data suggest that many patients with RVO still require frequent anti-VEGF injections 24 months after the onset of treatment and that long-term outcomes are unknown. We now report long-term outcomes for patients with RVO enrolled in the original ranibizumab study.
Subjects and Methods
The focus of this report is long-term follow-up of a study that was approved by the institutional review board (IRB) of Johns Hopkins University School of Medicine. The study was registered on December 1, 2006 at www.clinicaltrials.gov ( NCT00407355 ) and conducted in compliance with the Declaration of Helsinki, US Code 21 of Federal Regulations, and the Harmonized Tripartite Guidelines for Good Clinical Practice (1996). All patients provided informed consent before participation in the study. Twenty patients with BRVO and 20 patients with CRVO were randomized 1:1 to receive either 0.3 mg or 0.5 mg ranibizumab monthly for 3 months. Patients were seen at months 4, 5, 6, 9, and 12, and an attempt was made to hold injections to determine whether a period of rebound edema would be followed by final resolution of edema, but if edema persisted for several months and there was concern that it was a threat to the patient’s long-term visual potential, an injection of 1.25 mg bevacizumab was allowed. Beginning at month 12, patients were seen every 2 months and given an injection of ranibizumab if time-domain optical coherence tomography (OCT) demonstrated recurrent edema involving the fovea. Three patients with CRVO reached the month 12 visit before the approval of the amendment allowing for ranibizumab injections and received an injection of bevacizumab at month 12. Beginning at month 40, in addition to being eligible for as-needed treatment with ranibizumab, patients with evidence of recurrent macular edema by OCT underwent ultra-widefield fluorescein angiography and received scatter laser photocoagulation to all areas of retinal nonperfusion outside the fovea, followed by an injection of ranibizumab. If edema persisted and required 2 consecutive injections after laser treatment, ultra-widefield fluorescein angiography was repeated and supplemental scatter laser photocoagulation was done along with grid laser to areas of leakage in the macula but outside the foveal avascular zone. This was also followed by an injection of ranibizumab.
At each study visit the patient had best-corrected visual acuity (BCVA) measured by an experienced examiner using the Early Treatment Diabetic Retinopathy Study (ETDRS) protocol, OCT performed by an experienced investigator with a Stratus OCT3 (Carl Zeiss Meditec, Dublin, California, USA) using the Fast Macular Scan protocol, and a complete eye examination. Serial 7-field fluorescein angiography (FA) and color fundus photographs using a 30-degree lens were performed at baseline and at months 1, 3, 4, 18, 24, 30, and 36. Ultra-widefield FAs using the Optos P200Tx (Optos, Dunfermline, Scotland) were performed at the first study visit after month 36, on each visit in which laser treatment was given, and at months 48 and 60.
Administration of Study Drug
Povidone-iodine was used to clean the lids and a lid speculum was inserted. Topical anesthesia was applied and the conjunctiva was irrigated with 5% povidone-iodine. A 30-gauge needle was inserted through the pars plana and 0.05 mL ranibizumab was injected into the vitreous cavity. Funduscopic examination was done to confirm retinal perfusion.
Areas of retinal nonperfusion were treated with 200- to 500-μm burns 1 burn width apart. It was also assumed that there was additional nonvisualized retinal nonperfusion in the far periphery and 5 rows of laser photocoagulation were given for 360 degrees for patients with CRVO and roughly 180-200 degrees for patients with BRVO (the 2 quadrants affected by the BRVO) starting just posterior to the aura serrata. An injection of ranibizumab was given after completion of the laser photocoagulation and the patient was seen every 2 months and treated with ranibizumab if there was recurrent/persistent macular edema. If a ranibizumab injection was required on 2 consecutive visits after laser, supplemental laser photocoagulation was given to provide complete scatter photocoagulation of the peripheral retina to within 2 disc diameters of the arcade vessels and grid laser photocoagulation (100-μm burns 1 burn width apart) was given to areas of diffuse leakage in the macula but outside the foveal avascular zone. An injection of ranibizumab was given after the laser. If injections of ranibizumab were needed on 2 consecutive visits after the second laser treatment, remaining retina outside the arcade vessels was treated, followed by an injection of ranibizumab. Patients were then seen every 2 months and treated with ranibizumab for recurrent edema.
Measurement of Area of Posterior Retinal Nonperfusion
Areas of retinal nonperfusion were identified on FA by a distinctive darker appearance of the choroid and the “pruned” appearance indicating closure of retinal vessels in that region. The area of posterior retinal nonperfusion was measured on 7-field FA using the methodology previously described in the SCORE Study Report 9. A modified EDTRS grid measuring 16 disc areas and composed of 9 subfields (center, inner, and outer) was overlaid electronically on the center of the fovea. The area of retinal nonperfusion in the inner and outer subfields was measured using Image J Version 1.45s (National Institutes of Health, Bethesda, Maryland, USA; available at rsbweb.nih.gov ). The area in the center subfield was considered as the foveal avascular zone and was excluded from the measurement. In the SCORE Study Report 9, an FA was considered ungradable if intraretinal hemorrhage prevented the assessment of retinal nonperfusion within the grid. Similarly, an FA was considered ungradable if it was of poor quality (over-/underexposed or poorly focused) or if the amount of retinal hemorrhage was so great that it prevented any visualization of posterior retinal nonperfusion. If there was presence of intraretinal/retinal hemorrhage as well as visible retinal nonperfusion, the area under the hemorrhage was assumed not to have any retinal nonperfusion unless the hemorrhage was completely surrounded by an area of retinal nonperfusion. To differentiate hemorrhage from any adjacent retinal nonperfusion, an ETDRS grid was placed on the corresponding red-free image and hemorrhages were outlined. Hemorrhages were accurately identified (using the subfield grid lines and vessels as reference markers) and outlined on the FA, and posterior retinal nonperfusion was then measured. In addition to quantitative measurement of posterior retinal nonperfusion, a categorical assessment was done for all FAs from baseline to month 60 to take into consideration retinal nonperfusion outside the central subfields; retinal nonperfusion was graded as none, mild, moderate, severe, or ungradable. A poor-quality image or an image with severe hemorrhage was considered ungradable.
Statistical analyses were performed using the Statistical Package for the Social Sciences Version 19 (SPSS Inc, Chicago, Illinois, USA). The Mann-Whitney U test was performed to determine whether there were statistically significant differences between groups with or without severe retinal nonperfusion at year 5 and between groups with VA loss attributable or not attributable to retinal nonperfusion in several parameters: baseline BCVA, peak BCVA, year 5 BCVA, change in BCVA from baseline to peak, change in BCVA from baseline to year 5, and change in BCVA from peak to year 5. P values were calculated for 2-tailed significance. Relationship between BCVA and area of retinal nonperfusion were assessed using the Spearman rank order correlation (rho), followed by a linear mixed-effects model using area of retinal nonperfusion stratified by follow-up months.
Resolution of Macular Edema With Injection Independence
Patients were seen every 2 months between month 12 and month 72 and treated with ranibizumab as needed for recurrent/persistent edema, with a goal of achieving resolution of edema without need for ranibizumab injections. Patients were considered to reach this goal of “resolution” when they maintained a substantial amount of their peak improvement in vision and had no evidence of subretinal or intraretinal fluid for at least 6 months after their last injection. If a patient did not resolve by month 40, scatter photocoagulation was given in up to 3 sessions at least 3 months apart, and after the first session focal/grid laser was given to areas of leakage in the macula outside the foveal avascular zone. After resolution, patients were encouraged to remain in the trial, but some patients decided to exit the study after a prolonged period during which no treatment was required (these patients are classified as “reached endpoint and exited trial,” Tables 1 and 2 ). Some patients withdrew consent prior to resolution of edema; if this occurred prior to month 40 the patient was considered indeterminate with regard to edema resolution and if it occurred after month 40, the patient was considered unresolved.
|Patient ID||BCVA at Baseline (ETDRS Letter Score(Snellen Equivalent))||BCVA at Last Visit (ETDRS Letter Score(Snellen Equivalent))||CST at Last Visit (μm)||Follow-up Period (Months)||Last Injection (Month)||Time for Assessing Stability (Months)||Stability Assessment||Total Number of Anti-VEGF Injections||Status|
|B01||32 (20/250)||70 (20/40)||197||48||2||46||S||3||Reached endpoint and exited trial|
|B02||30 (20/250)||65 (20/50)||168||12||2||10||CD||3||Withdrew consent|
|B03||51 (20/100)||60 (20/63)||207||58||34||24||S||5||Developed NVAMD at M58|
|B04||66 (20/50)||67 (20/50)||284||48||48||0||NS||12||Withdrew consent – laser M48|
|B05||35 (20/200)||73 (20/32)||168||38||2||36||S||3||Reached endpoint and exited trial|
|B06||49 (20/100)||75 (20/32)||229||72||63||9||SL||27||Laser M48, M52, M60|
|B07||61 (20/63)||68 (20/50)||416||9||2||7||CD||3||Withdrew consent|
|B08||67 (20/50)||63 (20/50)||233||48||20||28||S||5||Reached endpoint and exited trial|
|B09||62 (20/63)||83 (20/20)||220||60||32||28||S||13||Reached endpoint and exited trial|
|B10||61 (20/63)||80 (20/25)||180||72||66||6||SL||29||Laser M48, M56|
|B11||39 (20/160)||73 (20/32)||229||72||70||2||NS||30||Laser M48|
|B12||50 (20/100)||61 (20/63)||184||72||22||28||S||6||Reached endpoint and exited trial|
|B13||56 (20/80)||64 (20/50)||240||72||34||38||S||15||Stable|
|B14||34 (20/200)||38 (20/160)||573||72||72||0||NS||32||Laser M4, M46|
|B15||65 (20/50)||65 (20/50)||215||63||60||3||NS||18||Laser M52, M56, M60|
|B16||65 (20/50)||79 (20/25)||232||42||34||8||S||5||Deceased|
|B17||71 (20/40)||89 (20/16)||180||72||2||70||S||3||Stable|
|B18||39 (20/160)||53 (20/125)||193||72||54||18||SL||18||Lost vision due to RNP, laser M44|
|B19||58 (20/80)||69 (20/40)||215||4||2||2||CD||3||Withdrew consent|
|B20||60 (20/63)||68 (20/40)||255||72||60||12||SL||9||Severe RNP, laser M60|
|Patient ID||BCVA at Baseline (ETDRS Letter Score(Snellen Equivalent))||BCVA at Last Visit (ETDRS Letter Score(Snellen Equivalent))||CST at Last Visit (μm)||Follow up Period (Months)||Last Injection (Month)||Time for Assessing Stability (Months)||Stability Assessment||Total Number of Anti-VEGF Injections||Status|
|C01||38 (20/160)||47 (20/125)||390||22||20||2||CD||6||Lost to follow-up|
|C02||32 (20/250)||50 (20/100)||172||2||2||0||CD||3||Withdrew consent|
|C03||33 (20/200)||43 (20/125)||278||6||6||0||CD||4||Deceased|
|C04||70 (20/40)||85 (20/20)||241||12||1||11||S||2||Reached endpoint and exited trial|
|C05||34 (20/200)||38 (20/160)||348||72||72||0||NS||29||Laser M48, M54, M64|
|C06||48 (20/100)||89 (20/16)||199||44||6||38||S||4||Reached endpoint and exited trial|
|C07||50 (20/100)||26 (20/320)||203||30||24||6||CD||10||Deceased- lost vision due to RNP|
|C08||58 (20/80)||67 (20/50)||280||72||72||0||NS||33||Laser M48, M54, M67|
|C09||19 (20/400)||36 (20/200)||152||72||43||29||NS||13||Laser M32, vitrectomy and laser M52|
|C10||37 (20/200)||30 (20/250)||489||72||72||0||NS||39||Laser M48, M54|
|C11||54 (20/80)||71 (20/40)||352||72||72||0||NS||33||Laser M48, M54, M63|
|C12||25 (20/320)||42 (20/160)||565||4||4||0||CD||5||Withdrew consent|
|C13||62 (20/63)||72 (20/32)||175||50||34||16||S||14||Reached endpoint and exited trial|
|C14||69 (20/40)||61 (20/63)||433||6||2||4||CD||2||Withdrew consent|
|C15||53 (20/80)||75 (20/32)||176||72||72||0||NS||26||Laser M46, M66|
|C16||49 (20/100)||64 (20/50)||146||72||26||46||S||12||Stable|
|C17||64 (20/50)||74 (20/32)||243||72||69||3||NS||20||Laser M46, M54|
|C18||62 (20/63)||62 (20/63)||271||72||72||0||NS||33||Laser M42, M46|
|C19||66 (20/50)||90 (20/16)||202||72||3||69||S||4||Stable|
|C20||48 (20/125)||60 (20/63)||193||6||2||4||CD||3||Withdrew consent|
In patients with BRVO, 8 out of 20 patients completed 6 years of follow-up and 11 completed 5 years; of the latter, 4 had resolution of edema from injections alone, 4 resolved after the addition of laser, and 3 remained unresolved. Of the remaining 9 patients who did not have 5 years of follow-up, 5 resolved before exiting the trial, 1 is unresolved because of withdrawal at month 48 while still requiring injections, and 3 are indeterminate because of withdrawal before month 40. Thus, of the 20 BRVO patients enrolled in the trial, injections alone resulted in resolution in 9 (1 of whom had had grid laser therapy prior to entering the trial) and failed to cause resolution in 8, and 3 were indeterminate ( Table 1 ). In the 9 patients who resolved with injections alone, the mean time from baseline to the last anti-VEGF injection was 20.2 months, the mean number of injections was 6.4, and baseline characteristics showed a mean age, BCVA, and duration of disease of 67.0 years, 24.3 letters, and 16.4 months vs 67.2 years, 21.6 letters, and 14.1 months in the 8 patients who did not resolve from injections alone. Four of the 8 patients who did not resolve from injections alone resolved after scatter and grid laser photocoagulation and an average of 20.8 anti-VEGF injections.
In patients with CRVO, 2 of 10 patients who completed 6 years of follow-up resolved from injections alone and 8 patients failed to resolve, including a patient who no longer needed injections after vitrectomy and laser for recurrent vitreous hemorrhages at month 52 but had a poor visual outcome from severe retinal nonperfusion and was therefore classified as unresolved. None of the other 7 patients resolved after scatter and grid laser therapy. Of the 10 patients who did not complete 6 years of follow-up, 3 resolved before exiting the trial and 7 were indeterminate because they failed to resolve before exiting prior to month 40. Thus, of the 20 CRVO patients enrolled in the trial, injections alone resulted in edema resolution in 5 and failed to cause resolution in 8, and 7 were indeterminate ( Table 2 ). In the 5 patients who resolved, the mean time from baseline to the last anti-VEGF injection was 14.0 months, the mean number of injections was 7.2, and baseline characteristics showed a mean age, BCVA, and duration of disease of 52.8 years, 29 letters, and 4.4 months vs 71.6 years, 19 letters, and 14.4 months in the 8 patients who did not resolve.
Progression of Retinal Nonperfusion in Patients With Retinal Vein Occlusion
Measurement of the area of retinal nonperfusion in the central subfields in patients with BRVO who had at least 3 years of follow-up showed that 9 of 14 patients had an increase between their first and last measurement ( Figure 1 , Upper left). The mean and median area of retinal nonperfusion over time for all patients with BRVO who had gradable FAs at baseline showed little change over the first 3 months, a sharp increase between 3 and 4 months, and then a gradual increase over time ( Figure 1 , Upper right). Changes in the area of posterior retinal nonperfusion over time were smaller in magnitude in CRVO patients compared with BRVO patients, but like patients with BRVO, the majority (7 of 11) showed an increase between their first and last measurement ( Figure 1 , Lower left). The mean and median area of retinal nonperfusion over time for CRVO patients who had a gradable FA at baseline showed a decrease between baseline and month 3, a sharp increase between months 3 and 4, and then a slight increase thereafter ( Figure 1 , Lower right).
Supplemental Figure 1 (available at AJO.com ) shows an image from the baseline FA of a patient with a BRVO illustrating retinal nonperfusion superior to the fovea and temporal to the macula. At month 3, there was no change in the retinal nonperfusion superior to the fovea but some reperfusion of previously nonperfused retina temporal to the fovea ( Supplemental Figure 2 , available at AJO.com ). At month 4, 2 months after the last ranibizumab injection, there was an increase in retinal nonperfusion temporal to the fovea ( Supplemental Figure 3 , available at AJO.com ). The patient received 4 anti-VEGF injections between months 4 and 18, trying to minimize injections while maintaining control of edema, and at month 18 there was further worsening of retinal nonperfusion temporal to the fovea ( Supplemental Figure 4 , available at AJO.com ). The patient received 8 injections of ranibizumab between months 18 and 36 with little change in nonperfusion ( Supplemental Figure 5 , available at AJO.com ). Wide-angle fluorescein angiography at month 48 showed very extensive retinal nonperfusion throughout the superior temporal quadrant ( Supplemental Figure 6 , available at AJO.com ) and scatter photocoagulation was done to try to reduce the need for ranibizumab injections. Injections of ranibizumab were still required at every visit (about every 2 months) and supplemental scatter and grid laser were given at month 56. At month 66, there was less leakage but edema was still present ( Supplemental Figure 7 , available at AJO.com ). Despite the many episodes of recurrent edema and lack of stability at M72, visual outcome was excellent ( Supplemental Figure 8 , available at AJO.com ).
Supplemental Figure 9 (available at AJO.com ) shows an image from the baseline FA of a 66-year-old patient with an 18-month history of CRVO. There is widening of the foveal avascular zone and areas of retinal nonperfusion throughout the posterior pole. After 3 injections of ranibizumab, there was less retinal nonperfusion ( Supplemental Figure 10 , available at AJO.com ). Ten anti-VEGF injections were given between months 4 and 36 to try and manage edema while minimizing the number of injections and at month 36 there was a substantial increase in posterior retinal nonperfusion ( Supplemental Figure 11 , available at AJO.com ). A wide-angle FA at month 46 showed extensive retinal nonperfusion throughout the periphery ( Supplemental Figure 12 , available at AJO.com ) and scatter photocoagulation was done throughout the periphery anterior to the equator ( Supplemental Figure 13 , available at AJO.com ), after which recurrent edema seemed less severe, although injections of ranibizumab were still required at most visits through month 72 ( Supplemental Figure 14 , available at AJO.com ).
Patients with BRVO who did not have at least 5 years of follow-up provided little information regarding change in retinal nonperfusion over time because of insufficient assessments, but their gradings are shown for completeness ( Figure 2 , Top). One had severe retinal nonperfusion at baseline that decreased to moderate at months 3 and 4 and then increased to severe at months 24 and 36. In the 11 patients with BRVO who completed at least 5 years of follow-up, posterior retinal nonperfusion at baseline was severe in 3, moderate in 3, mild in 2, none in 1, and not gradable in 2 ( Figure 2 , Bottom). The 2 patients with ungradable FAs at baseline had no retinal nonperfusion on their first gradable FA. At month 60, all patients who had severe or moderate retinal nonperfusion and 1 who had mild retinal nonperfusion at baseline had severe retinal nonperfusion, while the 4 remaining patients had no or mild retinal nonperfusion on their first gradable FA and at month 60. One patient who had severe retinal nonperfusion at baseline showed additional worsening of retinal nonperfusion between baseline and month 60 and 1 patient with no retinal nonperfusion on the first gradable FA had mild retinal nonperfusion at month 60. Thus, 54.5% of patients showed an increase in retinal nonperfusion category over the course of 5 years. Patients who had little or no retinal nonperfusion at baseline appeared less likely to progress, but the number of patients evaluated was too small to make any firm conclusions.
At baseline in the 10 patients with CRVO who completed at least 5 years of follow-up, retinal nonperfusion was moderate in 2, mild in 4, none in 1, and not gradable in 3 ( Figure 3 ). The patients who were not gradable at baseline had mild retinal nonperfusion on their first gradable FA. The 2 patients with moderate retinal nonperfusion at baseline and 1 with mild retinal nonperfusion on the first gradable FA progressed to severe retinal nonperfusion by month 60; the remaining 7 patients had mild or no retinal nonperfusion at month 60. In addition to the 3 patients who changed categories, 2 patients showed a definite increase in retinal nonperfusion but still fell within the mild retinal nonperfusion category at month 60, and therefore 50% of patients had worsening of retinal nonperfusion.
Effect of Worsening Retinal Nonperfusion on Visual Acuity
Among patients who completed at least 5 years of follow-up, 3 with BRVO and 2 with CRVO had a reduction in BCVA that was directly attributable to worsening of retinal nonperfusion (a third patient with CRVO who died at month 38 also experienced severe loss of vision from progression of retinal nonperfusion, Table 2 ). Supplemental Figure 15 (available at AJO.com ) shows an image from the baseline FA of an 82-year-old man with BRVO for 15 months. There was substantial retinal nonperfusion temporal to the fovea. After 3 injections of ranibizumab, there was improvement in retinal nonperfusion ( Supplemental Figure 16 , available at AJO.com ), but at month 4, 2 months after the last ranibizumab injection, the previously reperfused vessels were closed ( Supplemental Figure 17 , available at AJO.com ). Between months 4 and 18, BCVA dropped with recurrent edema and improved after injection of ranibizumab, but FA at month 18 showed nasal extension of retinal nonperfusion to involve the entire perifoveal capillary bed ( Supplemental Figure 18 , available at AJO.com ), and after that point BCVA did not improve to the former levels when macular edema was controlled. At month 36, there was further progression of retinal nonperfusion with closure of some large vessels ( Supplemental Figure 19 , available at AJO.com ). Supplemental Figure 20 (available at AJO.com ) shows a frame from a wide-angle FA after scatter photocoagulation at months 42 and 46, which did not eliminate the need for ranibizumab injections at each study visit through month 72 ( Supplemental Figure 21 , available at AJO.com ). Supplemental Figure 22 (available at AJO.com ) shows FA frames from a 78-year-old woman with CRVO after 3 injections of ranibizumab; there are hemorrhages, but no definite retinal nonperfusion. The next FA at month 18 shows severe retinal nonperfusion ( Supplemental Figure 23 , available at AJO.com ) with further worsening at month 36 ( Supplemental Figure 24 , available at AJO.com ). A plot of the patient’s BCVA over time shows that the BCVA dropped precipitously between months 9 and 12 and never recovered ( Supplemental Figure 25 , available at AJO.com ), suggesting that ischemic damage to the macula contributed to the permanent reduction in BCVA.
Table 3 shows baseline, peak, and month 60 BCVA in BRVO with at least 5 years of follow-up. Peak improvements from baseline ranged from 11-37 letters with a mean of 22.1 ( Table 4 ) and improvement ≥20 letters in 8 of 11 patients ( Table 3 ). At month 60, change from baseline ranged from −20 to +30 letters with a mean of 12.5. The mean change from baseline to peak visual acuity was similar in the 3 patients that ultimately lost vision because of retinal nonperfusion and the other 8 patients (19.7 vs 23.0 letters), but at month 60 they showed worsening from baseline (−1.3 letters) compared to mean improvement of 17.6 letters ( P < .05) in the other 8 patients. The Spearman correlation coefficient between BCVA and area of posterior retinal nonperfusion was −0.667 ( P < .05) at baseline, −0.564 ( P < .05) at months 18 and 24, and −0.693 ( P < .02) at month 36 ( Table 5 ). The correlation coefficient by linear mixed-effects model was not significant.
|ID||BL BCVA||Peak BCVA||Year 5 BCVA||ΔBCVA-BL to Peak||ΔBCVA- BL to Year 5||ΔBCVA- Peak to Year 5||Decrease in BCVA attributed to RNP||Number of Anti-VEGF Injections given|
|BL BCVA||Peak BCVA||Year 5 BCVA||ΔBCVA-BL to Peak||ΔBCVA-BL to Year 5||ΔBCVA- Peak to Year 5|
|All patients (n=11)||53.3||75.4||65.7||22.1||12.5||−9.6|
|Patients with severe RNP at Year 5 (n=7)||51.1||72.0||60.1||20.9||9.0||−11.9|
|Patients without severe RNP at Year 5 (n=4)||57.0||81.3||75.5||24.3||18.5||−5.8|
|Patients whose VA loss is attributed to RNP progression (n=3)||49.7||69.3||48.3 ∗||19.7||−1.3 ∗||−21. 0 ∗|
|Patients whose VA loss is not attributed to RNP progression (n=8)||54.6||77.6||72.3 ∗||23.0||17.6 ∗||−5.4 ∗|
|Baseline||Month 3||Month 4||Month 18||Month 24||Month 36|
|Spearman Correlation Coefficient||−.667 ∗||−.537||−.511||−.564||−.564 ∗||−.693 ∗|
|Adjusted Coefficient controlling for central subfield thickness||−0.827||−1.029||−0.427||−1.293||−0.275||−0.641|