Outcomes of Anti–Vascular Endothelial Growth Factor Treatment for Choroidal Neovascularization in Fellow Eyes of Previously Treated Patients With Neovascular Age-Related Macular Degeneration

Key Points

Question

Is there a difference in visual outcomes between eyes with established neovascular age-related macular degeneration (index eyes) and fellow eyes that more recently developed choroidal neovascularization?

Findings

In this single-center case series of 264 patients, fellow eyes presented with better visual acuity (VA; mean VA, 20/50) than index eyes (mean VA, 20/90) at the time of conversion to neovascular age-related macular degeneration. Fellow eyes maintained better VA than index eyes after equivalent follow-up (20/50 vs 20/70).

Meaning

Fellow eyes of patients with established neovascular age-related macular degeneration achieved better VA than index eyes after an equivalent follow-up period, which may be attributed to earlier diagnosis and treatment of choroidal neovascularization in fellow eyes compared with index eyes.

This single-center case series compares the visual outcomes of eyes with established, active neovascular age-related macular degeneration in index eyes with outcomes of fellow eyes that subsequently developed choroidal neovascularization during the management protocol.

Abstract

Importance

Neovascular age-related macular degeneration (nvAMD) is a leading cause of vision loss. The optimal screening protocol to detect choroidal neovascularization (CNV) in fellow eyes of patients undergoing treatment for unilateral CNV has not been determined.

Objective

To compare the visual outcomes of eyes with established, active nvAMD in index eyes with outcomes of fellow eyes that subsequently developed CNV during the management protocol.

Design, Setting, and Participants

In this retrospective single-center case series conducted at a private vitreoretinal practice, data were collected for all patients treated for bilateral nvAMD between October 1, 2015, and October 1, 2016, for whom we could determine the date of index eye and fellow eye conversion to nvAMD (n = 1600). Per institutional protocol, patients were screened for new CNV in the fellow eye at every office visit. Patients were excluded if they had a condition that could result in marked asymmetric vision loss.

Exposures

Development of nvAMD.

Main Outcomes and Measures

Visual acuity (VA) at the time of diagnosis of nvAMD and at equivalent time points following conversion to nvAMD for both index eyes and fellow eyes.

Results

A total of 264 patients met the inclusion criteria; 197 (74.6%) were women and 253 (95.8%) were white, and the mean (SD) age was 79.1 (8.2) years at time of index eye conversion to nvAMD and 80.6 (8.2) years at time of fellow eye conversion to nvAMD. Fellow eyes presented with better VA (mean VA, 20/50 [0.40 logMAR]) compared with index eyes (mean VA, 20/90 [0.67 logMAR]) at the time of conversion (difference, 14 letters [0.27 logMAR]; 95% CI, 10-17 [0.20-0.34]; P < .001). Index eyes did not achieve the same level of VA as fellow eyes after an equivalent postconversion follow-up of approximately 20 months (mean VA: index eye; 20/70 [0.56 logMAR]; fellow eye, 20/50 [0.40 logMAR]; difference, 8 letters [0.15 logMAR]; 95% CI, 4-11 [0.08-0.22]; P < .001). No difference was detected between the mean number of anti–vascular endothelial growth factor injections received by fellow eyes and index eyes (9.7 vs 10.0 injections, respectively).

Conclusions and Relevance

This retrospective study suggests that fellow eyes of previously treated patients with nvAMD may achieve better VA than their index eye counterparts after an equivalent amount of follow-up. This may be because the CNV was detected and treated earlier and at a better level of VA, although it is unknown whether the frequent office visits, VA measurements, or optical coherence tomography testing was responsible for the detection at a better level of VA.

Introduction

Age-related macular degeneration (AMD) is the most common cause of vision loss among elderly Americans.¹ While anti–vascular endothelial growth factor (VEGF) agents have revolutionized the treatment of neovascular AMD (nvAMD),^2,3 an effective means of treating nonneovascular AMD has eluded the ophthalmic community. Thus, patients with nvAMD in one eye and nonneovascular AMD in the fellow eye must wait until they develop signs of nvAMD in the fellow eye before definitive treatment can be initiated.

The time interval at which to screen patients who are being actively treated for nvAMD in one eye but who have nonneovascular AMD in the fellow eye is somewhat controversial. Some physicians examine and obtain optical coherence tomography (OCT) images of both eyes at every office visit to check for evidence of conversion from nonneovascular AMD to nvAMD. Other physicians treat the eye with nvAMD on a fixed, as required, or treat-and-extend protocol and only examine the fellow eye at some other interval (eg, every 3, 6, or 12 months).

Data from the ForeseeHome trial,⁴ the Submacular Surgery Trial,⁵ the MARINA study,⁶ and others⁷ suggest that timely identification and treatment of choroidal neovascularization (CNV) leads to better visual outcomes for patients with nvAMD. Several investigations have shown that better baseline visual acuity (VA) is a predictor of better posttreatment VA, which again highlights the importance of early detection and treatment of CNV.^8,9 In our practice, we routinely examine and obtain OCT imaging of fellow eyes with nonneovascular AMD in patients who are being actively treated for nvAMD in the other (index) eye. The purpose of the present analysis was to determine if fellow eyes fared better in terms of posttreatment VA after conversion to nvAMD than their index eye counterparts. We hypothesized that visual outcomes would be superior in fellow eyes compared with index eyes because of earlier detection and treatment of CNV in the fellow eyes.

Methods

A retrospective medical record review was performed to identify patients with nvAMD who received a same-day bilateral intravitreal injection of an anti-VEGF agent between October 1, 2015, and October 1, 2016, at Associated Retinal Consultants, William Beaumont Hospital, Royal Oak, Michigan. Patients were identified based on the International Classification of Diseases terminology codes for nvAMD (ICD-9-CM 365.52 and ICD-10-CM H35.32) and the Current Procedural Terminology code for bilateral injection (67028.50). From an initial list of 1600 patients with bilateral nvAMD who received at least 1 bilateral injection during the study period, we then reviewed medical records to identify individuals who had (1) previously been referred to our practice with a diagnosis of new nvAMD in one eye and nonneovascular AMD in the other eye or (2) converted to nvAMD in both eyes while under our care. From the resultant 480 patients, we excluded anyone with a history of an ophthalmic condition that could result in marked asymmetric vision loss, such as endophthalmitis, retinal vein or artery occlusion, amblyopia, macular hole, optic neuropathy, or retinal detachment. The medical records of the remaining 264 patients were then reviewed to identify the date of conversion from nonneovascular to nvAMD in the first (index) eye and subsequently the date of conversion in the second (fellow) eye for each patient. Demographic and clinical data from patients’ office visits were also recorded. The study was approved by the Western Institutional Review Board, and informed consent was waived because the study was retrospective and did not include an intervention. Paired t tests were used for comparisons between continuous variables, and χ² tests were used to evaluate differences among categorical variables. For all 2-tailed statistical tests, a P value less than .05 was considered statistically significant. Data analysis was performed using Stata version 14.2 (StataCorp).

Results

There were 264 patients who met the study eligibility criteria. Of these, 197 (74.6%) were women and 253 (95.8%) were white, and the mean (SD) age was 79.1 (8.2) years at the time of index eye conversion to nvAMD and 80.6 (8.2) years at the time of fellow eye conversion to nvAMD. The mean (SD) time interval between index eye conversion and fellow eye conversion was 1.6 (1.9) years, with a range of 1 month to 8 years.

Ocular characteristics of index and fellow eyes are presented in the Table. The main findings are that fellow eyes presented with better VA (mean VA, 20/50 [0.40 logMAR]) than index eyes (mean VA, 20/90 [0.67 logMAR]) at the time of conversion to nvAMD (difference, 14 letters [0.27 logMAR]; 95% CI, 10-17 [0.20-0.34]; P < .001), and index eyes did not achieve the same level of VA as fellow eyes after an equivalent postconversion follow-up period of approximately 20 months (mean VA: index eye; 20/70 [0.56 logMAR]; fellow eye, 20/50 [0.40 logMAR]; difference, 8 letters [0.15 logMAR]; 95% CI, 4-11 [0.08-0.22]; P = .001). The proportion of pseudophakic eyes was similar among index and fellow eyes, both at the time of conversion to nvAMD (38% vs 35%, respectively) and at equivalent postconversion follow-up visits (71% vs 76%). There was no difference in the mean number of anti-VEGF injections received by index and fellow eyes during equal periods of follow-up (10.0 vs 9.7 injections, respectively; difference, 0.23; 95% CI, −0.95 to 0.49; P = .53).

Table. Visual Outcomes of Index and Fellow Eyes.

Measurement	Eyes		Difference (95% CI)	P Value
	Index	Fellow
Mean VA at time of conversion to nvAMD	20/90	20/50	14 (10 to 17) letters; 0.27 (0.2 to 0.34) logMAR	<.001a
Follow-up after conversion to nvAMD, mean (SD), mo	19.8 (0.7)	20.3 (0.7)	0.6 (0.36 to 0.77) mo	<.001a
Mean VA after equivalent postconversion follow-up	20/70	20/50	8 (4 to 11) letters; 0.15 (0.08 to 0.22) logMAR	<.001a
Mean VA at most recent examination date	20/85	20/50	12 (8 to 16) letters; 0.24 (0.15 to 0.32) logMAR	<.001a
No. of anti-VEGF injections during follow-up, mean (SD)	10.0 (6.5)	9.7 (6.1)	0.23 (−0.95 to 0.49) injections	.53a
VA ≥20/40 at time of conversion to nvAMD, No. (%)	69 (26.1)	145 (54.9)	29% (21 to 37)	<.001b
VA ≥20/40 after postconversion follow-up, No. (%)	130 (49.2)	164 (62.1)	13% (5 to 21)	.003b
≥10 Letter gain, No. (%)	117 (44.3)	59 (22.3)	22% (12 to 31)	<.001b
≥10 Letter loss, No. (%)	44 (16.7)	48 (18.2)	1% (−10 to 12)	.86b

Abbreviations: nvAMD, neovascular age-related macular degeneration; VA, visual acuity; VEGF, vascular endothelial growth factor.

^aPaired t test.

^bχ² Test.

Discussion

Fellow eyes of previously treated patients with nvAMD achieved better final VA than their index eye counterparts after an equivalent amount of follow-up. There was no difference in the mean number of anti-VEGF injections between index and fellow eyes, and the duration of nvAMD was controlled for. Thus, our results suggest that early detection and treatment of newly diagnosed CNV in fellow eyes of previously or concurrently treated patients with nvAMD may result in superior visual outcomes in fellow eyes compared with index eyes.

The idea that prompt detection and treatment of CNV may preserve VA is not novel.⁵ However, despite the results of prior studies that show a benefit of early diagnosis and treatment of nvAMD,⁴ some practitioners still do not consistently screen fellow eyes with nonneovascular AMD for conversion to nvAMD at each index eye treatment session. Our study suggests that it may be prudent to screen the fellow eye with nonneovascular AMD with an OCT and fundus examination at every office visit while the index eye with nvAMD is receiving anti-VEGF injections. If this is done, CNV can be detected early in the fellow eye and treated before vision loss occurs.

Strengths and Limitations

The main strengths of this study are the large number of patients included in the analysis and the paired comparison of fellow and index eye data that allows patients to serve as their own internal control, thus reducing potential variables that could influence final visual outcome, such as differences in physician treatment strategies and medical and ocular comorbidities. However, the retrospective nature of the study precludes us from eliminating all potential confounders and biases from the analyses. Also, the data presented are from a single vitreoretinal practice and thus may not be generalizable to all practicing vitreoretinal specialists who may care for patients with AMD with different demographic makeups and/or lesion characteristics.

Conclusions

We conclude that routine screening for new CNV in fellow eyes with nonneovascular AMD of patients currently being treated for nvAMD at each office visit may be worthwhile because it could lead to the early detection and treatment of CNV when VA is still good, thus preserving VA among fellow eyes. Future studies may aim to determine whether the frequent office visits, VA measurements, or OCT testing might be responsible for the detection of new CNV at a better level of VA.