Hayreh SS, Zimmerman MB. Branch Retinal Vein OcclusionNatural History of Visual Outcome. JAMA Ophthalmol. 2014;132(1):13-22. doi:10.1001/jamaophthalmol.2013.5515
Understanding the natural history of visual outcome in branch retinal vein occlusion (BRVO) is fundamental to its management.
To investigate the natural history of visual outcome in major and macular BRVO.
Design, Setting, and Participants
Observational cohort study at a university-based referral practice from 1973 until 1999. The study comprised 216 consecutive eyes with BRVO (144 eyes with major and 72 eyes with macular BRVO) seen within 3 months of onset.
At first visit, all patients had a detailed ophthalmic and medical history and comprehensive ophthalmic evaluation. Ophthalmic evaluation at initial and follow-up visits included recording best-corrected visual acuity (VA) using the Snellen VA chart and visual fields with a Goldmann perimeter.
Main Outcomes and Measures
Best-corrected VA and visual field outcome on follow-up.
The median time to macular edema resolution was 21 months in those with major BRVO and 18 months in those with macular BRVO. Overall, for eyes with initial VA of 20/60 or better, VA improved or remained stable in 75% (95% CI, 63%-86%) for major BRVO and 86% (95% CI, 73%-95%) for macular BRVO. In those with initial VA of 20/70 or worse, VA improved in 69% (95% CI, 56%-80%) for major BRVO and in 53% (95% CI, 27%-79%) for macular BRVO, with median final VA of 20/60 for both BRVO types.
Conclusions and Relevance
Major and macular BRVOs are 2 distinct clinical entities, and initial visual status and final visual outcome in the 2 types are quite different. Overall, on resolution of macular edema, our study suggests that in both major and macular BRVO, VA and visual fields improved to a variable degree in the majority of eyes without any treatment.
Understanding the natural history of a disease is fundamental to its management so that treatments are not adopted with apparent benefits that in fact simply represent the natural history of a disease (they may even be harmful). Also, information about natural history acts as the basis to judge the extent of improvement by various therapies. Branch retinal vein occlusion (BRVO) is a common, visually disabling disease. Although BRVO has been known since 1896,1 there is still conflicting information on its visual outcome. The available data on the natural history of visual outcome in BRVO can be divided into 2 groups: (1) those in which no eye had any treatment2- 8 and (2) the control arm of randomized treatment studies on BRVO, with treated and untreated eyes.9- 19 All these studies have shown visual acuity (VA) improvement without any treatment; however, the incidence, criterion of VA improvement, and study designs vary widely among different studies so that it is impossible to compare their findings. Rogers et al20 reviewed the natural history of visual outcome in BRVO in articles published in English up to 2010 and concluded that VA generally improved in eyes with BRVO without intervention, although clinically significant improvement beyond 20/40 was uncommon.
Branch retinal vein occlusion actually consists of 2 distinct clinical entities: major BRVO and macular BRVO.21,22 Therefore, we investigated separately for major and macular BRVO the natural history of visual outcome in 216 consecutive untreated eyes (144 eyes with major BRVO and 72 eyes with macular BRVO) that fulfilled our inclusion and exclusion criteria.
We investigated various aspects of BRVO systematically in the Ocular Vascular Clinic at the tertiary care University of Iowa Hospitals and Clinics from 1973 until 1999 (when S.S.H. retired), as part of a large prospective study on ocular vascular occlusive disorders funded by the National Institutes of Health, approved by the University of Iowa Hospitals and Clinics institutional review board. In our study on BRVO, 214 patients (216 eyes: 144 eyes with major and 72 eyes with macular BRVO) fulfilled the following strict inclusion and exclusion criteria.
Only those patients who had a definite diagnosis of major or macular BRVO.
Only those patients who were seen within 3 months of onset of visual symptoms. With a longer interval between onset and evaluation of a patient, visual parameters may have already changed and do not provide valid information about the natural history of the disease.
We excluded all other retinopathies mimicking BRVO. All patients who had glaucoma and visual field loss, retinal or optic nerve lesion or any other factor (eg, cataract) that could influence visual outcome, unreliable visual fields, photocoagulation prior to the first visit, and inadequate information were excluded.
Natural history of visual outcome was documented by recording best-corrected VA and visual field defects. The data were collected prospectively and systematically. At the initial visit, all patients were seen by one of us (S.S.H.) and had detailed ocular and medical history as well as a comprehensive bilateral ophthalmic evaluation. This included (1) careful testing of the best-corrected VA using the Snellen VA chart, (2) visual field plotting with a Goldmann perimeter, (3) intraocular pressure recording with a Goldmann applanation tonometer, (4) relative afferent pupillary defect, (5) thorough anterior segment examination, including slitlamp examination of the anterior segment, lens, and vitreous, (6) meticulous fundus evaluation by direct and indirect ophthalmoscopy and, if required, by contact lens, and (7) stereoscopic color fundus photography and fluorescein fundus angiography (angiography only in the involved eye).
At each follow-up visit, the same ophthalmic evaluation and stereoscopic color fundus photography were performed, with fluorescein fundus angiography performed only when considered essential. Both eyes were examined at each visit.
Best-corrected VA was tested using the Snellen VA chart, under identical testing conditions, almost invariably by the same person (S.S.H.). Throughout this study, we used kinetic perimetry because automated perimetry did not exist when we started the study in 1973. Central visual field was also tested using the Amsler grid chart. It was most useful when kinetic perimetry showed no defect but the VA was not normal.
Each visual parameter was evaluated separately in a masked fashion so that the severity of one did not influence the evaluation of the other. Methods of evaluation were identical to those described in detail in the natural history of visual outcome in central retinal vein23 and hemicentral retinal vein24 occlusion. A change of at least 3 lines in the Snellen VA chart was considered a significant change, which is equivalent to a logMAR change of at least 0.30.
All patients who fulfilled the inclusion and exclusion criteria were followed up (by S.S.H.) according to the following protocol: at about 3-month intervals for 3 visits, then 6-month intervals for 4 visits, and then annually. Both eyes were examined at each visit. Since this was a natural history study, no intervention of any kind was undertaken in this cohort of patients.
Descriptive statistics (percentages, means, and medians) were computed for the demographic and clinical variables. Initial VA and visual field defect grade were compared between the sectors using the exact Wilcoxon rank sum test. A Kaplan-Meier curve was constructed to show the distribution of macula edema resolution over time. The association of VA improvement after macular edema had resolved with systemic conditions, smoking, and condition of capillary foveal arcade was tested using a Pearson χ2 test or Fisher exact test, with age by t test, and with degree of capillary obliteration by Cochran-Armitage trend. For data analysis, visual field defect grades were classified into the following 5 categories: minimal, mild, moderate, marked, and severe.
Quiz Ref IDComparison between major BRVO and macular BRVO showed significantly more arterial hypertension (66% vs 46%; P = .005), larger sector size (P < .001), worse initial VA (20/70 or worse in 47% vs 25%; P = .002), and worse initial visual field defect (moderate or worse visual field defect in 28% vs 2%; P < .001) in major BRVO. Therefore, the 2 types of BRVO were analyzed separately.
This study comprised 214 patients (216 eyes), 143 patients (144 eyes) with major BRVO and 71 (72 eyes) with macular BRVO. Table 1 gives their demographic characteristics. Quiz Ref IDThe sector involved for major BRVO was superior temporal in 65% of eyes and inferior temporal in 31%. For macular BRVO, the sector involved was superior in 81% of eyes and inferior in 19%. Severity of initial VA and visual field defects did not significantly differ between eyes with inferior sector and those with superior sector involvement (Table 2 and Table 3).
For all eyes with major BRVO, median initial VA was 20/60, with no change in median VA at 3 and 6 months’ follow-up. At 9 and 15 months’ follow-up, there was a less than 0.1 median decrease in logMAR to a median VA of 20/50. For macular BRVO, median initial VA was 20/40, with a −0.03 to −0.10 median change in logMAR at 6 to 15 months’ follow-up to a median VA of 20/25. This initial analysis of VA gave the impression that in major and macular BRVO there was a small change in VA 15 months from onset of BRVO. However, this may be misleading since VA change may be dependent on the level of initial VA, such that those who start with good initial VA are expected to show small improvements but may possibly show more worsening, and those who start with worse VA are expected to have less worsening but may possibly show more improvement. Thus, to account for the effect of initial VA on change in visual outcome, VA acuity change was examined based on initial VA, with eyes grouped into those with initial VA 20/60 or better and those with initial VA 20/70 or worse.
As shown in Figure 1, major BRVO eyes with initial VA of 20/70 or worse improved from median VA of 20/200 (1.0 logMAR) at the initial visit to 20/100 (0.7 logMAR) at 3 months, 20/80 (0.6 logMAR) at 6 months, and 20/50 (0.4 logMAR) at 15 months (Figure 1). In contrast, those with initial VA of 20/60 or better showed a small change in VA. For macular BRVO, improvement was also observed in those with initial VA of 20/70 or worse, and in those with initial VA of 20/60 or better, median VA was 20/30 at baseline and 20/25 at 15 months (Figure 2).
Changes in VA and visual field defect at 3, 6, 9, and 15 months’ follow-up, by sector, are presented in Table 4 for major BRVO and Table 5 for macular BRVO. Overall, for major BRVO, 75% (95% CI, 63%-86%) had improved or stable VA in eyes with initial VA of 20/60 or better, and 69% (95% CI, 56%-80%) had improved VA in eyes with initial VA of 20/70 or worse (Table 6). Visual field defect improved or remained stable in 67% (95% CI, 56%-78%) of eyes with a minimal to mild initial defect and improved in 52% (95% CI, 34%-69%) of eyes with a moderate to severe initial defect (Table 7).
For macular BRVO, VA was stable or improved in 86% (95% CI, 73%-95%) of eyes with 20/60 or better initial VA and improved in 53% (95% CI, 27%-79%) of eyes with 20/70 or worse initial VA (Table 6). Visual field defect remained stable or improved in 85% (95% CI, 71%-94%) of eyes with a minimal to mild initial defect (Table 7).
The Amsler grid chart always showed a defect when there was a central defect on kinetic perimetry. In 20 eyes with macular and 7 with major BRVO, when perimetry showed no defect but the VA was abnormal, it commonly showed metamorphopsia.
The median time to macular edema resolution was 21 months for major BRVO (n = 122) and 18 months for macular BRVO (n = 60). The Kaplan-Meier curve showing the distribution of resolution of macular edema over time for major and macular BRVO is shown in Figure 3. Visual acuities after macular edema resolved, and at final follow-up for those in whom macular edema did not resolve and who were followed up for at least 6 months, are shown in Table 6, and visual field outcomes are shown in Table 7. In major BRVO eyes with initial VA of 20/70 or worse, final median VA after macular edema resolved was 20/60 (25th-75th percentile, 20/40-20/80), with VA improving in 76% (95% CI, 60%-87%). In those with unresolved macular edema, VA at last follow-up had improved in 8 of 16 eyes (50%; 95% CI, 25%-75%). For visual field change in major BRVO that presented with moderate to severe defect, there was overall improvement in 54% (95% CI, 34%-72%) after resolution of macular edema. In the 5 eyes where macular edema had not resolved, 2 (40%) had improved visual field.
For macular BRVO with initial VA of 20/70 or worse, median VA after macular edema resolved was 20/70 (25th-75th percentile, 20/30-20/400), and VA improved in 7 of 12 eyes (58%; 95% CI, 28%-85%). There were only 3 macular BRVO eyes with initial VA of 20/70 or worse where macular edema did not resolve, of which all had VA of 20/40 at last follow-up.
Quiz Ref IDThe association of age, systemic conditions, and smoking with improvement in VA after macular edema had resolved was examined in eyes with initial VA of 20/70 or worse. Of these factors, only age showed a significant association with VA improvement, with those who improved being significantly younger than those with no improvement (mean [SD] age, 60.9 [8.8] vs 69.9 [10.8] years; P = .002). The association of capillary obliteration in the macular region and condition of the capillary foveal arcade (intact or broken) on fluorescein angiography (n = 50 eyes) with improvement in VA after macular edema had resolved was also evaluated in eyes with initial VA of 20/70 or worse. Of these eyes, capillary foveal arcade was intact in 31 and broken in 19, with VA improved in 81% of intact and 58% of broken arcades (P = .08). There was no significant association of VA improvement with severity of capillary obliteration in the macular region (P = .74). Visual acuity improved in 65% of eyes with extensive obliteration and in 69% of those with no capillary obliteration.
The median time to resolution of BRVO based on Kaplan-Meier analysis was 4 years (interquartile range, 2.2-9.8 years) for major BRVO (n = 135) and 1.5 years (interquartile range, 1.0-6.0 years) for macular BRVO (n = 66). Criteria for resolution of BRVO retinopathy were resolution of both macular edema and retinal hemorrhages.
Quiz Ref IDThis study showed that VA improved (≥0.3 logMAR) in the majority of eyes with both major and macular BRVO without any treatment. After macular edema resolved, VA had improved in 76% of eyes with major BRVO and 58% of eyes with macular BRVO, and it was worse in 9% and 42%, respectively; this shows that eyes with macular BRVO do not improve to the same extent as those with major BRVO. In eyes that presented with 20/60 or better VA, deterioration was minimal (12% of those with major BRVO and 3% of those with macular BRVO).
On fluorescein fundus angiography, Clemett et al7 found that at 1 year, in eyes with initial VA of 20/60 or worse, 67% of those with an intact foveal capillary arcade had final VA of 20/40 or better, while those with a broken arcade showed no improvement. There was a similar finding, with a smaller difference, in our study, with VA improvement in 80% of those with an intact arcade and 58% of those with a broken arcade (P = .08). Finkelstein25 showed that eyes with capillary obliteration in the macular region had greater frequency of improvement (91%) in VA than those without capillary obliteration (29%). In our study, there was no significant association of VA improvement with severity of capillary obliteration in the macular region (P = .74).
Previous studies on the natural history of visual outcome in BRVO can be divided into 2 groups: (1) those in which no eye had any treatment2- 8 and (2) the control arm of randomized treatment studies, with treated and untreated eyes.9- 19 In the second group, intravitreal steroid or anti–vascular endothelial growth factor therapies14- 19 do improve VA, but they require frequent intravitreal injections to maintain that improvement, and the therapies can have complications.26- 33
The findings from these studies cannot be compared with those of our study because of the different study designs, for example:
In our study, visual outcome was stratified by initial VA, and we used the currently accepted criterion, a change of at least 3 lines (0.3 logMAR) in the Snellen VA chart, for a significant change—a criterion used in all our previous studies on natural history of visual outcome.23,24,34 The Branch Vein Occlusion Study Group,11 by contrast, defined 2 lines in the Snellen VA chart as a significant change. In more recent studies, Early Treatment Diabetic Retinopathy Study letter score has been used to evaluate change in VA.
Our studies21,22 have shown that major and macular BRVOs are 2 distinct clinical entities, as is clear from the comparison of various parameters between the 2 types of BRVO and the different initial visual status and final visual outcomes of the 2 types (Tables 2, 3, 4, and 5). Previous studies did not make that distinction.
Also, in previous studies, visual outcome was based on VA only, with little information about visual fields. But VA provides information only about the function of the foveal region, not of the entire retina, whereas the visual field provides information about the entire retina. Major BRVO involves not only the fovea but also the entire sector of the retina up to the periphery drained by the occluded vein. Thus, for complete evaluation of visual loss in BRVO, information about the visual field loss is essential.
It could be argued that our study has some limitations. This requires explanation.
Not all patients in our study were followed up for 15 months, when the final data analysis was done, and that might have influenced the results of visual outcome in our study. To determine the effect of the dropout of patients before 15 months on the reported visual outcome, we did the following data analysis.
We compared the VA and presence of macular edema at the last visit of those with less than 15 months’ follow-up with those who were followed up for 15 months or longer. For major BRVO, those who dropped out before 15 months of follow-up had a median VA at last follow-up of 20/55, which was not significantly different when compared with those who were followed up for at least 15 months, with median VA at 9 months of 20/55 (P = .99) and at 15 months of 20/50 (P = .78). Macular edema had resolved in 42% of those who dropped out before 15 months, which was similar to those with at least 15 months of follow-up, with macular edema resolved in 36% (P = .51) at 9 months and 47% (P = .64) at 15 months. For macular BRVO, median VA was 20/30 and 39% had macular edema resolved at last follow-up for those with less than 15 months’ follow-up compared with those followed up at least 15 months, with median VA at 9 months of 20/30 (P = .91) and at 15 months of 20/25 (P = .86), with macular edema resolved in 47% (P = .59) at 9 months and 59% (P = .14) at 15 months.
Among the patients who dropped out before 15 months, macular edema had resolved in 42% of major and 39% of macular BRVO. Once macular edema had resolved, we found that there was little likelihood of any further visual improvement. That is why some patients with resolved macular edema might have decided not to come back for follow-up. Thus, patient dropout before 15 months could not have influenced the final results of visual outcome in our study.
Some may argue that we used kinetic perimetry for visual field evaluation, while automated perimetry is currently commonly used. Both types of perimetry have their advantages and disadvantages, discussed elsewhere.35,36 Automated static threshold perimetry (Humphrey 30-2 or 24-2 Swedish interactive thresholding algorithm) provides no information beyond about 24° to 30° in the periphery. Kinetic perimetry, by contrast, provides peripheral visual field information all the way to about 80° to 90° temporally, 70° inferiorly, 60° to 70° nasally, and 50° to 60° superiorly. In major BRVO, there is usually a segmental visual field loss in the entire involved region. To evaluate functional disability produced by the visual field defect in these eyes, it is essential to obtain information about the peripheral visual field loss, particularly the inferior visual loss. Kinetic perimetry, not automated perimetry, provides that information.
Some may contend that our evaluation of macular edema by contact lens examination does not provide satisfactory information about mild macular edema compared with optical coherence tomography.37 One of us (S.S.H.) has used contact lens examination for evaluation of macular edema for about 50 years. With that experience comes the ability to evaluate even mild macular edema by contact lens.
Although a majority of BRVO eyes had variable amounts of VA improvement without treatment, there was lack of improvement in some eyes that may be due to the same factors as those seen in ischemic central retinal vein occlusion,23 since most BRVO cases (particularly major BRVO) are ischemic in nature. These factors may include (1) ischemic damage to macular retinal ganglion cells, which are most vulnerable to ischemic damage,38 and (2) pigmentary degeneration and/or epiretinal membrane that may develop in the foveal region secondary to prolonged macular edema. Identifying the factors that play a role in the poor VA recovery in some eyes with BRVO may help guide clinical practice to target patients who will derive the most benefit from treatment.
Corresponding Author: Sohan Singh Hayreh, MD, PhD, DSc, FRCS, FRCOphth(Hon), Department of Ophthalmology and Visual Sciences, University Hospitals and Clinics, 200 Hawkins Dr, Iowa City, IA 52242-1091 (firstname.lastname@example.org).
Submitted for Publication: March 23, 2013; final revision received May 29, 2013; accepted May 31, 2013.
Published Online: October 24, 2013. doi:10.1001/jamaophthalmol.2013.5515.
Author Contributions: Dr Hayreh had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Hayreh.
Acquisition of data: Hayreh.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: All authors.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Zimmerman.
Administrative, technical, or material support: Hayreh.
Study supervision: Hayreh.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by grant EY-1151 from the National Institutes of Health and in part by an unrestricted grant from Research to Prevent Blindness Inc.
Role of the Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.