Article Text

Download PDFPDF

Baseline choroidal thickness as a short-term predictor of visual acuity improvement following antivascular endothelial growth factor therapy in branch retinal vein occlusion
  1. Nadim Rayess1,2,
  2. Ehsan Rahimy2,3,
  3. Gui-Shuang Ying4,
  4. Maria Pefkianaki2,
  5. Jason Franklin2,
  6. Carl D Regillo2,
  7. Allen C Ho2,
  8. Jason Hsu2
  1. 1 USC Roski Eye Institute, Keck School of Medicine, University of Southern California, Los Angeles, California, USA
  2. 2 Retina Service of Wills Eye Hospital, Mid Atlantic Retina, Philadelphia, Pennsylvania, USA
  3. 3 Department of Ophthalmology, Palo Alto Medical Foundation, Palo Alto, California, USA
  4. 4 Department of Ophthalmology, University of Pennsylvania, Philadelphia, Pennsylvania, USA
  1. Correspondence to Dr Jason Hsu, Retina Service of Wills Eye Hospital, Mid Atlantic Retina, Philadelphia, PA 19107, USA; jhsu{at}


Purpose To evaluate the association between subfoveal choroidal thickness (SFCT) and branch retinal vein occlusion (BRVO) eyes treated with antivascular endothelial growth factor (anti-VEGF) therapy.

Methods Retrospective cohort study of treatment naïve BRVO eyes treated with 3 monthly anti-VEGF injections. All patients received enhanced depth imaging spectral-domain optical coherence tomography scans to determine SFCT and central macular thickness (CMT). Baseline predictors (particularly SFCT) for functional response (best-corrected visual acuity (BCVA) gain ≥2 lines) were assessed at 3 months using univariate and multivariate analyses.

Results Forty eyes from 39 patients were included. Mean baseline SFCT was higher in functional responders (240.4±73.1 µm), compared with both non-responders (193.3±63.6 µm; p=0.036) and their corresponding fellow eye (202.2±67.1 µm; p=0.022). A higher baseline SFCT (for every 100 µm increase in SFCT) was found to be a positive predictor for functional response (regression coefficient: 1.1; p=0.03) on univariate analysis but not multivariate analysis. A worse baseline BCVA (for every 0.1 logMAR increase) was a positive predictor for visual improvement with an adjusted OR of 1.30 (95% CI 1.03 to 1.63; p=0.0009) on multivariate analysis.

Conclusions Patients with BRVO with a worse initial BCVA are most likely to achieve visual improvement following anti-VEGF therapy. Additionally, baseline SFCT may also help predict which patients with BRVO have favourable visual outcomes. Patients with an initial choroidal thickness thicker than their fellow eye are more likely to have short-term visual improvement following treatment.

  • retina
  • treatment medical
  • imaging

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Branch retinal vein occlusion (BRVO) is a prevalent retinal vascular disease in addition to diabetic retinopathy and central retinal vein occlusion. Macular oedema is one of the most common causes of visual acuity (VA) impairment in patients with BRVO. The pathophysiology of macular oedema is believed to be related to a rise in vascular endothelial growth factors (VEGF) levels which then contributes to increased exudation of fluid from retinal vasculature.1 As such, anti-VEGF injections have become the mainstay for treating macular oedema secondary to BRVO.2 3

The choroid is a highly vascularised network of tissue that provides the necessary source of oxygen to the outer third of the retina. Enhanced depth imaging optical coherence tomography (EDI-OCT) has provided the means for further studying the choroid, primarily by measuring choroidal thickness.4 In fact, studies have suggested that the choroid may play a role in the pathophysiology of various retinal diseases including: age-related macular degeneration, diabetic retinopathy, central serous chorioretinopathy and retinal vein occlusion (RVO).5–13

More recently, choroidal thickness in patients with BRVO has been evaluated.5–8 These studies found a higher than normal choroidal thickness in patients with macular oedema secondary to BRVO. This suggests that the choroidal vasculature is also affected by BRVO and may have important clinical implications. Furthermore, two recent studies assessed the role of choroidal thickness in potentially predicting response to anti-VEGF therapy. In a study by Okamoto et al,6 patients with a thinner initial choroidal thickness had better anatomical outcomes following a single injection of ranibizumab, whereas Hasegawa and associates7 found that patients with a higher baseline choroidal thickness were more likely to have a reduction in central retinal thickness following anti-VEGF therapy. Since the choroid plays a primary role in perfusing the outer retina and photoreceptors, the primary objective of our study was to evaluate initial choroidal thickness as a predictor for visual recovery in patients with macular oedema secondary to BRVO.


This study was a retrospective consecutive case series of patients diagnosed with treatment naïve BRVO at Mid Atlantic Retina and the Retina Service of Wills Eye Hospital. Following approval by the Institutional Review Board (IRB), billing records from January 2010 through January 2014 were used to identify patients diagnosed with BRVO using the International Classification of Diseases, Ninth Revision (ICD-9) code 362.36.

Patients were included in this study if they were newly diagnosed with treatment naïve BRVO and were treated with 3 monthly anti-VEGF injections for macular oedema. The exclusion criteria of the study included patients who had received: focal/grid laser photocoagulation, pan-retinal photocoagulation (PRP) or intravitreal injections (eg, intravitreal corticosteroids). Patients were also excluded if they had any of the following concomitant ocular diseases: diabetic macular oedema, age-related macular degeneration, prior ocular surgery (except cataract surgery) and choroidal neovascularisation.

Patient charts were reviewed to collect the following data: age, gender, medical history (hypertension and diabetes mellitus), lens status, best-corrected Snellen visual acuity (BCVA) based on spectacle correction with pinhole and anti-VEGF injection dates as well as drug used.

Imaging protocol

Enhanced depth imaging optical coherence tomography (EDI-OCT) using the Heidelberg Spectralis (Heidelberg Engineering, Heidelberg, Germany) platform was used in all patients included. Subfoveal choroidal thickness (SFCT) was measured using the built-in calliper software at a single point below the fovea extending from the bottom of the hyper-reflective layer corresponding to Bruch’s membrane to the hyper-reflective layer at the sclerochoroidal junction. Choroidal thickness measurements were recorded at initial presentation and 3 months follow-up for both the BRVO eyes and fellow eyes. Two experienced reviewers (NR and MP) reviewed all EDI-OCT scans together and any disagreements on measuring end points were resolved by a third interpreter (ER) who were all blinded to patient clinical data. In such cases, the third reviewer’s measurement was used for data analysis. Furthermore, central macular thickness (CMT) was recorded for all patients using the automated software present in the 25-line raster scan pattern.

Statistical analysis

Data were analysed using SAS V.9.3. Paired two-tailed t-test was used to compare continuous data at baseline and 3-month follow-up. In particular, choroidal thickness in BRVO eyes was compared with the fellow eye at baseline and 3-month follow-up. Functional responders (BCVA gain ≥2 lines) and non-responders were compared using unpaired two-tailed t-test. Univariate followed by multivariate linear regression analyses modelling visual outcomes (BCVA change from baseline) and morphological outcomes (SFCT) as a continuous variable were performed to determine baseline predictors for functional response. Furthermore, we categorised the continuous measures into clinically meaningful binary outcomes, and performed the univariate and multivariate logistic regression analyses for categorical visual outcomes (≥2 line gain) and anatomical outcomes (SFCT ≤200 μm or >200 μm).


A total of 40 eyes from 39 patients met inclusion criteria for this study. Table 1 describes the baseline characteristics of the total patient population and compares the baseline characteristics of the responder and non-responder group. At baseline, mean SFCT was 220±72 µm, which was greater than the fellow eyes (199±64 µm; p=0.067). In addition, mean baseline CMT for BRVO eyes (496±156 µm) was found to be significantly higher than the fellow eyes (294±70 µm; p<0.001). The change in outcome measures of the study (choroidal thickness, CMT and BCVA) following 3 monthly anti-VEGF injections is summarised in table 2.

Table 1

Baseline demographics and ocular characteristics in BRVO eyes treated with anti-VEGF therapy (n=39 patients, 40 eyes)

Table 2

Visual and anatomical outcome measures at baseline and after 3 months of anti-VEGF

Functional responders and non-responders

Twenty-three BRVO eyes were functional responders (gain ≥2 Snellen lines) and 17 eyes were functional non-responders (gain <2 Snellen lines). Mean baseline logMAR BCVA was 1.15±0.55 (Snellen equivalent: 20/282), which was significantly worse compared with non-responders 0.57±0.42 (Snellen equivalent: 20/75; p<0.001). At 3 months follow-up, logMAR BCVA improved to 0.59±0.32 (Snellen equivalent: 20/78; p<0.001) for functional responders, whereas for non-responders logMAR BCVA was 0.58±0.54 (Snellen equivalent: 20/76; p=0.962). No significant difference was found in BCVA when comparing functional responders with non-responders at a 3-month follow-up (p=0.922).

Mean baseline SFCT was higher in functional responders (240.4±73.1 µm), compared with their corresponding fellow eyes (202.2±67.1 µm; p=0.022) and to the non-responders (193.3±63.6 µm; p=0.036). However, there was no difference in baseline choroidal thickness between the non-responder BRVO eyes and their fellow eye (195.5±62.0 µm; p=0.876). Following 3 monthly anti-VEGF injections, mean SFCT decreased in both the functional responder group (202.4±70.1 µm; p<0.001) and the non-responder group (154.6±59.0 µm; p<0.001). Conversely, choroidal thickness in the fellow eye at a 3-month follow-up remained similar at baseline for both functional responders (197.8±70.8 µm; p=0.30) and non-responders (188.9±66.2 µm; p=0.252).

Mean baseline CMT was greater in responders (533.7±171.9 µm) compared with non-responders (444.4±118.2 µm; p=0.059). At a 3-month follow-up, mean CMT decreased in both responders (323.4±92.3 µm; p<0.001) and non-responders (364.3±108.0 µm; p=0.009). Although baseline SFCT and CMT were both higher in responders compared with non-responders, the Spearman correlation coefficient (R=0.162; p=0.459) demonstrated that there was no correlation between the two variables.

Baseline predictors for functional response

Univariate analysis for predictors of functional response (gain ≥2 Snellen lines) after 3 monthly anti-VEGF injections is outlined in table 3. A worse baseline BCVA (logMAR >1.00) was a predictor for favourable functional outcomes (92.9% vs 38.5%, p=0.0009). Furthermore, having a higher baseline choroidal thickness (for every 100 µm increase in choroidal thickness) was also found to be a positive predictor of functional response in patients with BRVO (regression coefficient: 1.1; p=0.03). Of note, there was no difference in outcomes whether patients received ranibizumab, bevacizumab or a combination of the two medications. Multivariate regression analysis demonstrated that a worse baseline BCVA (for every 0.1 logMAR increase) was overall the strongest positive predictor with an adjusted OR of 1.30 (95% CI 1.03 to 1.63; p=0.0009).

Table 3

Univariate analysis for baseline predictors of functional response at 3 months in BRVO eyes treated with anti-VEGF


In this study evaluating baseline predictors for anti-VEGF treatment outcomes, a worse baseline visual acuity was the strongest independent predictor for functional response (gain ≥2 Snellen lines). In addition, a higher initial choroidal thickness was found to be a significant predictor by univariate analysis but not multivariate analysis. There was no difference in functional response according to anti-VEGF agent used, whether bevacizumab or ranibizumab. Furthermore, functional responders had a significantly thicker baseline choroidal thickness compared with both non-responders and their fellow eye. This contrasts with non-responders who had a similar baseline choroidal thickness as their corresponding fellow eye. These results indicate that comparing choroidal thickness in treatment naïve BRVO eyes with their fellow eye may potentially serve as an indicator for patients who are more likely to achieve short term visual gains following anti-VEGF therapy.

In addition, baseline CMT was found to be thicker in functional responders compared with non-responders. While both initial CMT and SFCT were thicker in functional responders, there was no correlation between the two variables (Spearman correlation coefficient R=0.162). This indicates that SFCT does not appear to be a simple surrogate for CMT in separating responders from non-responders. Rather, CMT and SFCT may have different predictive values in determining response to anti-VEGF therapy. Similarly, baseline BCVA in responders was significantly worse than non-responders, which may indicate that a thicker choroidal thickness is associated with a worse initial BCVA. However, the correlation coefficient assessing the relationship between initial BCVA with baseline SFCT was very weak (R=0.085). Therefore, it is unlikely that choroidal thickness is simply a surrogate marker for BCVA. A more complex relationship may be at play between the two variables.

Prior studies have assessed baseline predictors for visual response in patients with BRVO. Kim et al reported that significant predictors of visual outcomes at 6 months included having a worse baseline BCVA, younger age and a shorter duration of symptoms prior to treatment.14 However, in this study, patients were treated with several regimens including: anti-VEGF therapy, triamcinolone, laser therapy or a combination of methods. This is in contrast to our study that only assessed outcomes in patients treated with a standard regimen of 3 monthly anti-VEGF injections. Another study by Farinha and associates assessed predictors of functional outcomes in patients with BRVO treated with ranibizumab. Their study demonstrated that a better initial BCVA correlated with a better final BCVA, while RPE disruption negatively correlated with final BCVA.15 These results appear to be in conflict with the visual acuity findings in our study and the study by Kim and colleagues. However, the discrepancy can largely be explained by differences in how visual outcomes were defined. In our study, functional response was defined by a gain of greater than two lines of BCVA, whereas the study by Farinha et al used final BCVA rather than gain in visual acuity to define treatment response.

Choroidal thickness in patients with BRVO has been evaluated in several recent studies.5–8 In all of these studies, patients with BRVO and macular oedema had a significantly greater mean choroidal thickness compared with the fellow eye. Furthermore, following anti-VEGF treatment, choroidal thickness significantly decreased.5–8 These findings are similar to our study. Two more recent studies assessed choroidal thickness as a predictor for anatomical outcomes and had conflicting results. In the study by Okamoto et al, patients with resolution of CME had a thinner choroidal thickness at baseline and follow-up compared with those who had recurrent macular oedema.6 In contrast, Hasegawa et al found that patients who had a thicker choroidal thickness at baseline compared with their fellow eye were more likely to have significant anatomical improvements compared with patients who had a thinner baseline choroidal thickness that was similar to the fellow eye.7 One explanation for the conflicting findings in the studies by Okamoto et al and Hasegawa et al is the difference in follow-up period. In the study by Okamoto et al, patients only received a single intravitreal injection of ranibizumab and outcomes were assessed at 1-month and 2-month follow-up, whereas patients received more injections and had longer follow-up (6 months) in the study by Hasegawa and colleagues.6 7 Furthermore, as Okamoto et al explained in their paper, a thicker choroidal thickness in patients who did not respond in their study may indicate a more VEGF-driven process. Since patients only received a single injection of ranibizumab, it is possible that with more injections, anatomical response would have been more sustained in the group with a thicker initial choroid.6 However, these studies are different from ours in that choroidal thickness was not evaluated as a predictor for functional response.

In our study, a significantly thicker choroidal thickness compared with the fellow eye correlated with improved functional outcomes, whereas BRVO eyes that had a choroidal thickness similar to their fellow eye did not demonstrate the same visual gains. The results of our study are similar to that by Hasegawa and associates in that patients with a higher baseline choroidal thickness were more likely to have improved treatment outcomes with anti-VEGF therapy. However, the two studies differ with regards to the primary outcome. Our study evaluated visual outcomes whereas the study by Hasegawa et al focused on anatomical response. Taken together, these two studies suggest that a higher initial choroidal thickness may predict both anatomical and functional response. Nevertheless, larger studies are necessary to further assess the role of initial choroidal thickness as a predictor of both visual and anatomical outcomes.

Alterations in choroidal thickness as seen in BRVO with macular oedema suggest that it may play a role in the pathophysiology. As seen in our study, patients with a higher initial choroidal thickness were more likely to experience visual gains following anti-VEGF therapy compared with patients with a thinner choroid. Similarly, improved anatomical outcomes have been demonstrated in patients with a thicker baseline choroidal thickness.7 One hypothesis is that patients with a greater choroidal thickness are more likely to have intact perfusion to the outer retinal layers and therefore have the greater potential for visual gains following treatment of macular oedema. Conversely, a thinner baseline choroidal thickness may be indicative of a more ischaemic process and poorer perfusion to the outer retinal layers with less potential for visual recovery. An alternative hypothesis is that a thicker choroid may correlate with higher VEGF levels indicating a primarily VEGF-driven aetiology for the macular oedema, making these patients more likely to respond to anti-VEGF therapy.16–18 In contrast, a thinner choroid may suggest that the macular oedema is not entirely due to VEGF production but other proinflammatory mediators, which may partly explain the lower likelihood of responding to anti-VEGF therapy.

One of the limitations of this study is that it assessed short-term outcomes for patients treated with anti-VEGF therapy and may not be an indicator of long-term outcomes. However, implementing a standard treatment algorithm of 3 monthly anti-VEGF injections is one of the strengths of our study as it prevents the treatment algorithm from confounding visual outcomes. Following the initial 3 monthly anti-VEGF injections, most patients were treated with variable algorithms (eg, pro re nata or treat and extend), which would have influenced treatment outcomes had we included a more extended follow-up period. Given the strict inclusion criteria, the sample size of the study was relatively small. As such, baseline choroidal thickness was found to be a predictor for functional outcomes only on univariate analysis and not multivariate regression analysis. Furthermore, given the small sample size, analysing SFCT and BCVA as a dichotomous variable may have led to loss of statistical power. Larger studies are needed to help further elucidate the role of choroidal thickness. Another potential limitation is including patients who received either bevacizumab or ranibizumab injections. However, there are no studies to date that demonstrate a difference in visual outcomes between the drugs when they are administered as monthly injections.19 In addition, this was a retrospective study and, therefore, patients’ visual acuity was measured using Snellen charts rather than standard ETDRS letters commonly used in prospective studies. A final limitation relates to choroidal thickness measurements which were not standardised for diurnal variations and were made at only one location (subfoveally) which may have missed regional variations in thickness.

In conclusion, treatment-naïve patients with BRVO who have a thicker baseline choroidal thickness compared with their fellow eye were more likely to have improved short-term visual outcomes following anti-VEGF therapy compared with those with an initial choroidal thickness that is similar to their fellow eye. In addition, a worse initial BCVA was found to be strongly correlated with the likelihood of gaining 2 lines of visual acuity. Larger, prospective studies are needed to further evaluate the role of choroidal thickness as a predictor of visual outcomes in BRVO.


J Arch McNamara Memorial Fund for Retina Research through Wills Eye Hospital was used to fund statistical analysis.



  • Contributors NR, ER and JH: concept and design. NR, ER, G-sY, MP, JF, CDR, ACH and JH: data collection, analysis and interpretation of the data; preparation, review or approval of the article.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Disclaimer The sponsor or funding organization had no role in the design or conduct of this research.

  • Competing interests None declared.

  • Patient consent Detail has been removed from this case description/these case descriptions to ensure anonymity. The editors and reviewers have seen the detailed information available and are satisfied that the information backs up the case the authors are making.

  • Ethics approval Institutional Review Board.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas