Patterns and Predictors of Successful Treatment Discontinuation in Retinal Vein Occlusions With Macular Edema in the Real World

Abstract

BACKGROUND AND OBJECTIVE:

To identify factors associated with successful treatment discontinuation in eyes with retinal vein occlusions (RVOs) and macular edema (ME) in real-world settings.

PATIENTS AND METHODS:

Retrospective study of 214 eyes with RVO and ME with 24-month follow-up at five academic centers. Regression analyses identified factors associated with (1) successful treatment discontinuation for at least 6 months without fluid recurrence and (2) best-corrected visual acuity (BCVA) at 24 months.

RESULTS:

Forty percent of eyes with branch RVO and 35% with central RVO (CRVO) / hemi-retinal RVO (HRVO) successfully discontinued therapy without fluid recurrence, with median time to discontinuation of 6 and 7 months, respectively. Lower 6-month central subfield thickness was associated with greater likelihood of treatment discontinuation within 24 months for eyes with CRVO/HRVO (P = .001), whereas better 6-month BCVA was associated with better 24-month BCVA for all RVO subtypes (P < .001).

CONCLUSION:

Early anatomic response at 6 months is associated with greater likelihood of stopping treatments.

INTRODUCTION

Retinal vein occlusions (RVOs) affect approximately 16 million adults in the United States, Europe, Asia, and Australia.¹ The major cause of vision loss in patients with RVOs is macular edema, which can be treated with intravitreal injections of anti-vascular endothelial growth factor (VEGF) agents or corticosteroids, as well as laser photocoagulation.^2,3 Multiple prospective, randomized clinical trials have demonstrated the effectiveness of these treatments in improving vision and reducing the central subfield thickness (CST) in patients with RVOs and macular edema (ME).^4–14 However, patients in these trials often follow strict treatment protocols, in contrast to patients in “real-world” settings. Real-world patients are also often less adherent with follow-up visits and demonstrate inferior visual and anatomic outcomes than those in clinical trials.^15–18

Although intravitreal anti-VEGF therapy has become the first-line treatment for patients with ME secondary to RVOs,^4–6 frequent intravitreal injections are a clinical burden,¹⁹ with significant health care costs.²⁰ Risk of adverse events also increases with repeated treatments.^21,22 Whereas some eyes with RVO and ME may discontinue therapy without edema recurrence, few studies have evaluated the factors that are linked to successful discontinuation of treatments. In this study, we performed a multi-center, retrospective analysis of eyes with RVO and ME from five U.S. academic centers and examined associations between clinical and anatomic markers with sustained discontinuation of treatment for at least 6 months without ME recurrence and visual outcomes during a 2-year period.

PATIENTS AND METHODS

Study Design and Ethical Considerations

We performed a retrospective study of patients with treatment-naïve RVO and ME managed by retinal specialists at five academic centers in the United States between January 2011 and September 2015, including University of California Davis, University of Rochester Medical Center, Cleveland Clinic Foundation, Oregon Health and Science University, and University of California San Diego. Each center received study approval from their respective institutional review board, and all data were obtained and stored in accordance with the Declaration of Helsinki and the Health Insurance Portability and Accountability Act.

Patient Selection

Patients were included if they met the following criteria: 1) a new diagnosis of acute (symptoms < 3 months) central RVO (CRVO), hemi-RVO (HRVO), or branch RVO (BRVO), 2) presence of center-involving ME as determined by spectral-domain optical coherence tomography (SD-OCT), 3) follow-up for at least 24 months after their initial visit, and 4) age of 18 years or older. ME was defined by CST measurement on the Spectralis OCT device (Heidelberg Engineering, Heidelberg, Germany) (> 320 μm for men or > 305 μm for women) or the Cirrus OCT system (Zeiss, Oberkochen, Germany) (> 305 μm for men or > 290 μm for women). In patients who had two eligible eyes for the study, only the first affected eye was included in the study. Patients were excluded from the study if they had other concurrent retinal conditions (eg, diabetic retinopathy, diabetic ME, age-related macular degeneration, vitreomacular interface abnormalities, or retinal detachment), prior history of anti-VEGF or steroids injections, or a history of intraocular surgeries except for uncomplicated cataract extraction based on clinical documentation.

Data Collection

Medical records were reviewed to collect demographic and clinical characteristics including age, sex, and systemic comorbidities such as diabetes, hypertension, hypercoagulable states, and dyslipidemia. Due to variable documentation of clinical records and laboratory test results, the presence of these comorbidities were based on presence of clinical documentation. We reviewed ocular characteristics including lens status (phakic, pseudophakic, or aphakic), type of RVO (BRVO, CRVO, or HRVO), and documented history of glaucoma or diabetic retinopathy. Best-corrected visual acuity (BCVA) and CST measured using SD-OCT were recorded from visits at baseline, 6, 12, and 24 months. The cumulative number and type of intravitreal injections and laser procedures were also recorded at 6, 12, and 24 months. We carefully identified eyes that demonstrated successful treatment discontinuation, defined as macular edema resolution for at least 6 months without any intervention (intravitreal injection or laser), and with continuous follow-up as described previously. Participants who resumed treatments at a later time were still included.^4,23

Statistical Analysis

We compared demographic and clinical characteristics between eyes with BRVO and HRVO/CRVO using Student’s t-test for scale variables and Chi-square tests for binary variables. We then identified the factors associated with BCVA outcomes at 24 months using linear regression analysis, and those associated with treatment discontinuation using binary logistic regressions. The Snellen BCVAs were recorded and converted to logarithm of minimal angle of resolution (logMAR) equivalents, and BCVA of counting fingers, hand motion, and light perception were assigned logMAR values of 1.6, 2.0, and 2.5, respectively, for analysis. We analyzed our data using SPSS software version 25 (IBM Corp., Armonk, NY). P values less than .05 were considered statistically significant.

RESULTS

Patient and Ocular Characteristics

We included 214 eyes with newly diagnosed RVO and ME with 24-month follow-up. Mean age of patients was 69.8 ± 11.9 years, with more female than males (59.8% females). Of the eyes reviewed, 52.3% had CRVO, 9.3% had HRVO, and 38.3% had BRVO. Most eyes were phakic (73.8% phakic, 25.7% pseudophakic, 0.5% aphakic). For statistical analysis, the one aphakic eye was included with the pseudophakic eyes. Systemic and ocular comorbidities include diabetes (37.9%), hypertension (61.2%), dyslipidemia (41.1%), hypercoagulable states (11.2%), and glaucoma (23.8%). More patients with BRVO had hypertension compared to those with CRVO or HRVO (65.9% vs. 58.3%; P = .02). The proportion of eyes with glaucoma were greater in the CRVO / HRVO group, although this did not reach statistical significance (28.0% vs. 17.1%; P = .07). Other demographic, clinical, or ocular characteristics were not significantly different between the BRVO and HRVO/CRVO groups (Table 1).

TABLE 1.

Clinical Characteristics of Patients Who Were Managed for Retinal Vein Occlusions During a 24-Month Period

	All Patients (n = 214)	BRVO (n = 82)	CRVO/HRVO (n = 132)	P Value*
Patient Characteristics
Mean Age, years (SD)	69.8 (11.9)	69.6 (11.9)	70.0 (12)	.79
Female sex, No. (%)	128 (59.8%)	48 (58.5%)	80 (60.6%)	.76
Diabetes, No. (%)	81 (37.9%)	30 (36.6%)	51 (38.6%)	.67
Hypertension, No. (%)	131 (61.2%)	54 (65.9%)	77 (58.3%)	.02*
Dyslipidemia, No. (%)	88 (41.1%)	33 (40.2%)	55 (41.7%)	.54
Hypercoagulable state, No. (%)	24 (11.2%)	7 (8.5%)	17 (12.9%)	.47
Ocular Characteristics
Right eye, No. (%)	104 (48.6%)	38 (46.3%)	66 (50.0%)	.60
Glaucoma, No. (%)	51 (23.8%)	14 (17.1%)	37 (28.0%)	.07
Lens status (phakic / pseudophakic / aphakic), No.	158 / 55 / 1	58 / 24 / 0	100 / 31 / 1	.48

^*Statistically significant, P < .05.

SD = standard deviation; RVO = retinal vein occlusion; BRVO = branched retinal vein occlusion; CRVO = central retinal vein occlusion; HRVO = hemi-retinal vein occlusion

Real-World Management of RVO With Macular Edema

Eyes with RVO and ME were most commonly treated with intravitreal anti-VEGF injections, including bevacizumab (Avastin; Genentech, South San Francisco, CA), ranibizumab (Lucentis; Genentech, South San Francisco, CA), and aflibercept (Eylea; Regeneron, Tarrytown, NY), with a cumulative mean of 4.1 ± 1.7 injections at 6 months, 6.6 ± 3.1 injections at 12 months, and 10.2 ± 5.5 injections at 24 months (Figure 1A). Bevacizumab was the most commonly given anti-VEGF agent at all time points, with increasing proportions of aflibercept and ranibizumab given later in the course of treatment (Figure 1A). Some eyes also received intravitreal steroids, including triamcinolone and dexamethasone implants, with similar proportions of the two types of intravitreal steroids used at each time point (Figure 1B). Of the eyes that received intravitreal triamcinolone, the cumulative mean was 1.25 ± 0.50, 1.67 ± 0.52, and 2.00 ± 1.05 injections at Months 6, 12, and 24, respectively. Of those that received dexamethasone implants, the cumulative mean was 1.50 ± 0.58, 1.83 ± 0.98, and 2.10 ± 1.51 injections at Months 6, 12, and 24. Some eyes also received focal laser treatments (Figure 1C), and a small proportion of eyes received peripheral scatter laser photocoagulation as well, with 3.3%, 6.5%, and 9.3% of eyes treated by Months 6, 12, and 24, respectively.

Figure 1.

Real-world treatment patterns of eyes with retinal vein occlusion (RVO) and macular edema (ME). Bar graphs showing mean cumulative number of (A) intravitreal anti-vascular endothelial growth factor injections, (B) intravitreal steroids, and (C) focal laser treatments for eyes with RVO and ME at different time points. In (A) and (B), the shaded regions represent the proportion of bevacizumab, ranibizumab, and aflibercept in (A), and the proportion of triamcinolone and dexamethasone implants in (B).

Treatment Discontinuation in RVO With Macular Edema

A total of 79 (36.9%) patients demonstrated treatment discontinuation for at least 6 months during the 2-year study period, with a slightly greater proportion of eyes with BRVO than CRVO / HRVO (40% vs. 35%; P = 0.427). Among these eyes, the median time to treatment discontinuation was 6 and 7 months for BRVO and CRVO/HRVO, respectively (Figure 2). Treatment discontinuation was associated with lower CST at 6 months (P = .001) in eyes with CRVO/HRVO during a 24-month period. For eyes with BRVO, none of the clinical and anatomic factors were significantly associated with treatment discontinuation (Table 2).

Figure 2.

Rates of successful treatment discontinuation in eyes with retinal vein occlusion (RVO) and macular edema (ME). Kaplan-Meier curves of the proportion of eyes with (A) branch RVO (BRVO) or (B) central RVO (CRVO) / hemi-RVO (HRVO) and ME that demonstrated successful treatment discontinuation during the 24-month study period. The solid line shows the Kaplan-Meier estimate of the proportion of treated subjects, and the dashed line show 95% confidence interval.

TABLE 2.

Factors Associated With Discontinuation of Treatments for 6 Months or More in Retinal Vein Occlusions

Univariate Regression
		BRVO		CRVO / HRVO
	Category or Increment	Odds Ratio (95% CI)	P Value	Odds Ratio (95% CI)	P Value
Gender	Male vs. female	1.569 (0.639–3.853)	.326	0.741 (0.352–1.557)	.428
Age	1 year	1.012 (0.974–1.051)	.541	0.980 (0.951–1.010)	.197
Diabetes	Present vs. absent	0.379 (0.142–1.013)	.379	0.537 (0.249–1.157)	.113
Hypertension	Present vs. absent	0.825 (0.224–3.043)	.773	0.473 (0.212–1.056)	.068
Dyslipidemia	Present vs. absent	1.250 (0.436–3.584)	.678	0.818 (0.371–1.803)	.619
Hypercoagulable	Present vs. absent	0.181 (0.020–1.621)	.127	1.213 (0.418–3.521)	.722
Laterality	Right vs. left	1.210 (0.497–2.945)	.674	0.765 (0.373–1.569)	.465
Glaucoma	Present vs. absent	0.486 (0.137–1.717)	.262	0.996 (0.446–2.226)	.992
Lens status	Pseudophakic/aphakic vs. phakic	0.704 (0.258–1.920)	.493	0.973 (0.421–2.247)	.949
BCVA at baseline	1 logMAR	1.699 (0.591–4.882)	.325	1.443 (0.703–2.960)	.317
BCVA at 6 months	1 logMAR	1.856 (0.458–7.517)	.386	0.526 (0.220–1.258)	.149
CST at baseline	1 μm	0.999 (0.995–1.002)	.516	1.000 (0.998–1.002)	.890
CST at 6 months	1 μm	0.996 (0.990–1.001)	.114	0.993 (0.990–0.997)	.001*

^*Statistically significant, P < .05.

BCVA = best-corrected visual acuity; logMAR = logarithm of minimum angle of resolution; CST = central subfield thickness; RVO = retinal vein occlusion; BRVO = branch retinal vein occlusion; CRVO = central retinal vein occlusion; HRVO = hemi-retinal vein occlusion; OR = odds ratio; CI = confidence interval

Real-World Outcomes of RVO With Macular Edema

After 24 months, visual acuity (VA) improved in most eyes with RVO and ME, including those that discontinued treatment, from mean logMAR 0.66 ± 0.48 (Snellen 20/91) at baseline to logMAR 0.46 ± 0.42 (Snellen 20/58) at Month 6, logMAR 0.42 ± 0.39 (Snellen 20/52) at Month 12, and logMAR 0.46 ± 0.48 (Snellen 20/58) at Month 24. Eyes with BRVO exhibited better VA at baseline (logMAR 0.54 ± 0.42, Snellen 20/69) compared with those with CRVO / HRVO (logMAR 0.74 ± 0.50, Snellen 20/110), and remained statistically superior across all time points (P = .001) (Figure 3A). At 24 months, eyes with BRVO showed a mean BCVA of logMAR 0.31 ± 0.32 (Snellen 20/41), whereas those with CRVO / HRVO were logMAR 0.55 ± 0.53 (Snellen 20/71).

Figure 3.

Real-world visual and anatomic outcomes of eyes with retinal vein occlusion (RVO) and macular edema (ME). Line graphs of the mean (A) logMAR best-corrected visual acuity (BCVA) and (B) central subfield thickness of eyes with central RVO (CRVO), hemi-RVO (HRVO; solid line), or branch RVO (BRVO; dashed line) and ME at different time points. Error bars represent standard error. * Statistically significant, P < .05.

Mean CST decreased for most patients with RVO-related ME from 486 ± 174 μm at baseline to 354 ± 164 μm at 6 months, 325 ± 11 μm at 12 months, and 324 ± 114 μm at 24 months. Eyes with BRVO had lower CST compared to eyes with CRVO / HRVO at baseline (433 ± 133 μm vs. 519 ± 188 μm; P < .001), but were not significantly different at 24 months (326 ± 98 μm vs. 321 ± 122 μm; P = .777) (Figure 3B).

On univariate regression, better BCVA at 24 months was associated with younger age (BRVO: P = .016; CRVO/HRVO: P = .047), phakic lens status (BRVO: P = .015), better BCVA at baseline (BRVO: P = .025; CRVO/HRVO: P < .001) and at 6 months (BRVO and CRVO/HRVO: P < .001), and lower CST at baseline (BRVO: P = .001; CRVO/HRVO: P = .045) and at 6 months (BRVO: P = .001). However, multivariate regression using these factors showed that only pseudophakic eyes (BRVO: P = .018), BCVA at 6 months (BRVO and CRVO/HRVO: P < .001), and CST at 6 months (BRVO: P = .038) independently predicted visual outcomes at 24 months (Table 3).

TABLE 3.

Factors Associated With Best-Corrected Visual Acuity (logMAR) at 24 Months in Retinal Vein Occlusions

Univariate Regression
		BRVO		CRVO/HRVO
	Category or Increment	β Estimate (95% CI)	P Value	β Estimate (95% CI)	P Value
Gender	Male vs. female	−0.081 (−0.225 to 0.064)	.270	−0.070 (−0.259 to 0.119)	.466
Age	1 year	0.007 (0.001 to 0.013)	.016*	0.008 (0.000 to 0.015)	.047*
Diabetes	Present vs. absent	0.061 (−0.087 to 0.209)	.417	−0.042 (−0.237 to 0.154)	.674
Hypertension	Present vs. absent	0.060 (−0.172 to 0.292)	.605	0.138 (−0.082 to 0.359)	.216
Dyslipidemia	Present vs. absent	0.055 (−0.126 to 0.237)	.543	−0.133 (−0.354 to 0.088)	.236
Hypercoagulable	Present vs. absent	−0.140 (−0.421 to 0.140)	.321	−0.076 (−0.386 to 0.235)	.629
Laterality	Right vs. left	0.066 (−0.077 to 0.208)	.361	0.783 (−0.159 to 0.211)	.783
Glaucoma	Present vs. absent	0.081 (−0.099 to 0.261)	.373	0.099 (−0.112 to 0.309)	.354
Lens status	Pseudophakic/aphakic vs. phakic	0.189 (0.038 to 0.339)	.015*	0.172 (−0.041 to 0.385)	.113
BCVA at baseline	1 logMAR	0.188 (0.024 to 0.351)	.025*	0.513 (0.349 to 0.677)	< .001*
BCVA at 6 months	1 logMAR	0.575 (0.390 to 0.761)	< .001*	0.850 (0.707 to 0.993)	< .001*
CST at baseline	1 μm	9.956E-6 (−0.001 to 0.001)	.001*	0.001 (0.000 to 0.001)	.045*
CST at 6 months	1 μm	2.073E-5 (−0.001 to 0.001)	.001*	0.000 (−0.001 to 0.000)	.463
Multivariate Regression
Age	1 year	−0.001 (−0.007 to 0.005)	.687	0.005 (−0.001 to 0.010)	.113
Lens status	Pseudophakic/aphakic vs. phakic	0.198 (0.035 to 0.361)	.018*
BCVA (baseline)	1 logMAR	0.034 (−0.148 to 0.215)	.714	0.070 (−0.108 to 0.248)	.437
BCVA (6 month)	1 logMAR	0.589 (0.369 to 0.809)	< .001*	0.804 (0.616 to 0.992)	.001*
CST (baseline)	1 μm	−2.474E-5 (−0.001 to 0.000)	.920	0.000 (−0.001 to 0.000)	.499
CST (6 month)	1 μm	−0.001 (−0.001 to 0.000)	.038*

^*Statistically significant, P < .05.

BCVA = best-corrected visual acuity; logMAR = logarithm of minimum angle of resolution; CST = central subfield thickness; RVO = retinal vein occlusion; CRVO = central retinal vein occlusion; HRVO = hemi-retinal vein occlusion; OR = odds ratio; CI = confidence interval

DISCUSSION

Multiple prospective studies have demonstrated the effectiveness of intravitreal anti-VEGF, intravitreal corticosteroids, and focal laser for treating RVO-related ME,^4–14 but patients enrolled in clinical trials are often managed with more regular treatment regimens than those in the real world. Most studies of real-world experience and outcomes have been published from small cohorts of eyes with RVO-related ME, ranging from 28 to 66 patients,^4,17,23 with few that assess factors associated with treatment discontinuation.¹⁸ Given the dearth of prospective clinical trial evidence for an exit strategy, ophthalmologists have little guidance for when, how, or in whom treatment may be discontinued. In this multicenter study, we evaluated the treatment pattern and clinical outcomes of 214 patients with treatment-naïve RVO-related macular edema, and identified factors associated with successful treatment discontinuation and good visual outcome at 2 years.

In our study, patients received approximately four anti-VEGF injections during the first 6 months, and on average only six more injections from Months 6 to 24, similar to other real-world studies.^17,18 The majority of treatments given were bevacizumab, followed by ranibizumab and aflibercept, although the proportion of the on-label agents increased with time, likely due to the increasing proportion of incomplete responders and broadening of insurance coverage after initiating treatment. Focal laser was given to 9% of patients, and intravitreal steroids were also given to 9% of patients. We could not determine the specific treatment strategy for each patient, such as injection pattern including monthly, pro re nata (PRN), or “treat-and-extend,” the use of loading doses, or the criteria for switching between therapies, which are limitations of this study.

We found that 40% of BRVO and 35% of CRVO/HRVO eyes demonstrated ME resolution, defined as discontinuation of all therapies based on physician discretion without fluid recurrence for at least 6 months during the 24-months study period. A duration of 6 months was chosen based on published studies evaluating ME resolution,^4,23 and to ensure all eyes had the same observation period. Our rate of ME resolution was slightly lower than the rates reported in the prospective RETAIN study (50% resolution in BRVO, 44% resolution in CRVO),²³ and a smaller real-world study of ranibizumab-treated RVO eyes (69% resolution in BRVO, 48% resolution in CRVO).²⁴ However, both studies only included ranibizumab-treated eyes and had longer mean follow-up periods of 49 and 48 months, respectively. Our study showed that CST but not BCVA at 6 months was associated with successful treatment discontinuation. This finding is limited by dependence on physician interpretation rather than a centralized reading center, use of different SD-OCT devices across different institutions, and lack of quantitative analysis of different types of retinal fluid. We chose the 6-month time point because most prospective RVO trials required at least 6 monthly loading doses of anti-VEGF treatments prior to PRN or interval extension,^{4–6,11,25,26} and most visual and anatomic improvements occur during this period (Figure 3). However, we cannot exclude the possibility of stopping due to futility rather than treatment success, although our study defined successful discontinuation in the absence of fluid recurrence to minimize this possibility.

Eyes with BRVO demonstrated better VA than those with CRVO or HRVO at baseline and maintained better vision throughout the study, with 24-month Snellen-equivalent BCVA of 20/40 versus 20/71, respectively. These values are slightly inferior to real-world data reported by others, although the baseline VA in those studies were better and eyes that received intraocular steroids were not included, which may have resulted in selection bias for eyes with less severe pathology.^17,18,24 Anatomically, we found that although eyes with CRVO / HRVO had worse baseline CST than BRVO, both groups experienced substantial anatomic improvements that stabilized after Month 12, consistent with published studies,²⁷ supporting the effectiveness of current treatment paradigms for reducing ME independent of visual outcomes.

Previous studies have identified factors associated with visual outcomes in treated RVO-related ME. In the SCORE2 study comparing monthly aflibercept and bevacizumab, baseline CST was associated with 6-month BCVA on univariate regression, but only age and baseline BCVA were independent predictors on multivariate regression.²⁸ A comprehensive analysis of SD-OCT biomarkers from the SHORE study comparing monthly versus PRN ranibizumab also found several anatomic features such as size of intraretinal cysts and disorganization of retinal inner and outer layers to predict visual gains on univariate analyses, but again, only age and baseline BCVA were independently associated with 6-month visual improvements in multivariate regression.²⁹ Similarly, although we found multiple factors such as age, pseudophakia, and baseline or 6 month BCVA/CST associated with BCVA at 2 years on univariate regressions, only 6-month BCVA was associated with visual outcomes in multivariate models for all RVO subtypes, likely because initial responses to treatment provide much stronger prognostic value than baseline characteristics. For eyes with BRVO, phakic lens status and lower CST at 6 months were also independently associated with BCVA at 24 months, suggesting that these additional factors may help predict visual prognosis in eyes with BRVO.

This study is limited by the retrospective nature of the analysis and the inclusion of multiple academic centers where inherent practice patterns may differ. Documentation of fluorescein angiography data or laboratory results were not required for inclusion in our study, and the diagnosis and management of each patient may vary with physician discretion. Although RVO management strategies may not have been uniform across all sites, our study provides important, real-world prognostic indicators for functional and anatomic success using current treatment paradigms for the management of RVO-related ME. We found that while lower 6-month CST is associated with successful treatment discontinuation for eyes with CRVO/HRVO only, better 6-month BCVA is associated with better 24-month visual outcomes in all RVO subtypes. Future studies with longer duration of follow-up may help determine the durability of the ME resolution in these patients.