To assess the incidence and progression of macular atrophy and other key anatomic outcomes over 7 to 8 years in an early cohort of ranibizumab-treated exudative age-related macular degeneration patients.
Follow-up analysis of long-term outcomes in a multicenter treatment cohort.
Fourteen study sites enrolled 65 previous subjects from the ranibizumab treatment arms of the ANCHOR, MARINA, and HORIZON trials. In a single update visit, clinical assessment and retinal imaging studies were performed, with comparison with each subject’s prior results from the previous trials. Early Treatment Diabetic Retinopathy Study visual acuity was the primary outcome. Secondary outcomes, including area of macular atrophy and selected anatomic factors, were analyzed for associations with long-term vision outcomes.
At a mean 7.3 years after ANCHOR or MARINA enrollment, mean visual acuity was 54 letters, study eyes having received a mean 1.6 injections per year since the HORIZON study. Macular atrophy was present in 98% of study eyes, the mean area increasing from 0.83 ± 0.96 mm 2 at the ANCHOR or MARINA year 2 exit to 2.22 ± 1.6 mm 2 at the SEVEN-UP visit, a growth rate of 0.28 mm 2 /year. Progression of macular atrophy was associated significantly with visual decline over this 5-year period ( P < .001), and final macular atrophy lesion size was related significantly to final vision ( P < .001). Other key anatomic outcomes (macular thickening, thinning, or fluid and submacular fibrosis) did not have significant effects on vision outcomes.
Seven years after initiation of intensive ranibizumab therapy for exudative age-related macular degeneration, macular atrophy progression and severity were the primary anatomic determinants of visual outcomes.
Exudative age-related macular degeneration (AMD) is a chronic disease that can require protracted treatment with ocular anti–vascular endothelial growth factor (VEGF) therapies. Since intravitreal ranibizumab (Lucentis; Genentech, Inc, South San Francisco, California, USA) was clinically approved in 2006, many patients have now been receiving this treatment for 7 years or more. Although the significant visual benefits of ranibizumab, off-label intravitreal bevacizumab (Avastin; Genetentech, Inc) and intravitreal aflibercept (Eylea; Regeneron Pharmaceuticals, Tarrytown, New York, USA) have been demonstrated for AMD patients in major phase 3 randomized controlled multicenter trials, available results are limited largely to 1- or 2-year studies. Longer-term evidence to describe vision outcomes and to guide the ongoing management of exudative AMD patients remains incomplete. In addition, macular anatomic outcomes over an extended therapeutic course and their relationship to long-term vision outcomes have not been elucidated.
The Seven-Year Observational Update of Macular Degeneration Patients Post-MARINA/ANCHOR and HORIZON Trials (SEVEN-UP) study ( ClinicalTrials.gov identifier, NCT01256827 ) previously reported visual outcomes in a cohort of AMD patients who were among the first to receive ranibizumab: those who had been study participants 7 to 8 years previously in the phase 3 Anti-VEGF Antibody for the Treatment of Predominantly Classic Choroidal Neovascularization in Age-Related Macular Degeneration (ANCHOR) trial ( ClinicalTrials.gov identifier, NCT00061594 ) and Minimally Classic/Occult Trial of the Anti-VEGF Antibody Ranibizumab in the Treatment of Neovascular Age-Related Macular Degeneration (MARINA) trial ( ClinicalTrials.gov identifier, NCT00056836 ) and who had continued in the treatment arms of the Open-Label Extension Trial of Ranibizumab for Choroidal Neovascularization Secondary to Age-Related Macular Degeneration (HORIZON) extension study ( ClinicalTrials.gov identifier, NCT00379795 ). In these subjects, who offer the longest available treatment history, the SEVEN-UP study found divergent vision outcomes. After 2 years of intensive monthly ranibizumab treatments followed by what is now considered low treatment frequency in the HORIZON study and off-study thereafter with ranibizumab and intravitreal bevacizumab (Avastin), 37% of eyes had 20/70 or better Early Treatment Diabetic Retinopathy Study (ETDRS) best-corrected visual acuity (BCVA), whereas 37% of eyes had 20/200 or worse ETDRS BCVA.
As previously reported, a high proportion of eyes at this late stage displayed continued choroidal neovascular leakage and macular edema, submacular fibrosis, and macular atrophy (loss of retinal pigment epithelium [RPE], outer retina, and choriocapillaris). In the current analyses, additional access to the original ANCHOR and MARINA imaging databases made it possible to track the progression of macular atrophy in individual subjects over an approximate 5-year period to assess the effect on vision change in this group of treated AMD patients. Also, in multivariate regression analyses, key anatomic outcomes could be evaluated to distinguish their relative impact on vision outcomes. The hypothesis is that long-term visual outcomes in treated AMD patients are affected by the development of specific types and severity of macular pathologic features. Identifying such anatomic markers could assist clinicians in understanding the highly variable long-term vision outcomes observed among AMD patients and could guide treatment strategies over extended durations.
The SEVEN-UP study was a follow-up analysis of long-term outcomes in a multicenter treatment cohort of AMD patients who were originally in the ANCHOR, MARINA, and HORIZON trials. Health Insurance Portability and Accountability Act and Declaration of Helsinki guidelines were followed, and approval granted by Institutional Review Boards for each participating center. All subjects gave informed consent to participate. Study visits were conducted from November 2010 through November 2011.
The predetermined primary outcome measure was the percentage of study eyes with ETDRS BCVA of 20/70 or better. Predetermined secondary outcome measures included the presence of cystoid macular edema, subretinal fluid, or both, as well central subfield thickness (CST) and the presence and location of submacular fibrosis. Macular atrophy was another predetermined secondary outcome measure in terms of percentage of eyes affected, area, and progression from baseline in ANCHOR or MARINA.
The SEVEN-UP study represented 14 clinical trial sites in the United States from the original ANCHOR, MARINA, and HORIZON studies. All subjects who had participated previously at 1 of these 14 sites in the ranibizumab-treatment arms of ANCHOR or MARINA (including both the 0.3-mg and 0.5-mg dosing assignments) who also completed 24 months in the treatment arms (0.5 mg ranibizumab as needed) of the HORIZON extension study were eligible and were recruited by invitation. Through ANCHOR or MARINA and HORIZON, 357 patients completed their assignment in the ranibizumab treatment arms; 155 of these were eligible from the 14 participating SEVEN-UP study clinical sites. By outreach from the original sites of their participation, 65 subjects were enrolled for a single update visit. For this group, the mean duration since entry into ANCHOR or MARINA was 7.3 years, and the mean duration since completing month 24 of HORIZON was 3.4 years.
Seven-up Study Protocol
The single visit study evaluation included BCVA measurements with ETDRS vision test by a certified examiner using standard protocols. A complete ophthalmologic examination was performed, and clinical report forms recorded assessments by investigators. Patient records also were reviewed retrospectively when available for key events in the previous 6 months, as well as for treatment history over the interim since exit from the HORIZON study. A panel of retinal imaging studies included spectral-domain optical coherence tomography (SD OCT), fundus photography, fluorescein angiography (FA), and fundus autofluorescence (FAF) imaging. No interventions or treatments were included as part of the SEVEN-UP protocol.
All image data collected at SEVEN-UP visit were analyzed at the Doheny Imaging Reading Center by independent, masked, certified graders according to a predetermined standardized and validated grading protocol. Imaging studies from prior study participation were obtained for SEVEN-UP subjects from the ANCHOR and MARINA databases, provided by the Wisconsin Reading Center.
Macular Atrophy Assessment
For an analysis of macular atrophy progression over time, study eyes were assessed by red-free photographs and early and late FA frames at the month 24 ANCHOR or MARINA exit visit, and by red-free images, early and late FA frames, and FAF images at the SEVEN-UP visit. FAF imaging was available only for year 7 and was used for analyses restricted to year 7 outcomes. Where analyses required comparison with earlier time points, measurements were based on red-free images and FA. Interpretation and measurements were performed in a masked fashion by an experienced retinal specialist, with each study eye measured independently at least twice to reconcile disparities and determine a mean value. Macular atrophy was defined as a demarcated area of RPE defect or abnormal visibility of large choroidal vasculature, as an area of window defect without leakage on FA, and as an area of decreased autofluorescence on FAF imaging. Automated measurements were made by manually delineating the lesion borders using OIS integrated software (Ophthalmic Imaging Systems, Inc, Sacramento, California, USA). A minimum greatest linear dimension of 175 μm defined the presence of macular atrophy.
All data were collected by clinical investigators in case report forms and forwarded for entry into an encrypted electronic database (REDCap) at the University of California, San Francisco. For statistical analysis, data were de-identified and analyzed using Stata software version 10.1 (Stata Corporation, College Station, Texas, USA). Linear regression models were used to identify clinical and anatomic factors that were associated with ETDRS letter score at the SEVEN-UP visit. All patients had only 1 eye in the study sample. Based on previous literature, initial vision score, patient age, gender, race, SD OCT indicators of leakage and macular thickness, submacular fibrotic scar formation, and macular atrophy were evaluated as potential predictors. These were chosen a priori, and the associations were evaluated for each in a preliminary bivariate regression analysis against vision outcomes. Subsequently, multivariate regression models were performed restricted to the 4 key anatomic variables, with the inclusion of prior baseline vision score where appropriate. Patient age, although found in the bivariate analysis to be associated significantly with vision outcomes, was not included so as to limit the number of variables in the statistical model given the sample size.
Figure 1 shows comparisons of vision outcomes for patient demographic factors and baseline vision, analyzed for independent associations to BCVA outcomes. For ethnicity, a minority of SEVEN-UP subjects were nonwhite (7, 11%); the white and nonwhite patient groups had comparable ETDRS letter scores (mean ± standard deviation ETDRS letter score, 46.6 ± 26.0 letters vs 39.0 ± 27.3 letters; P = .51). Male and female patient groups also had similar ETDRS vision score profiles (mean letter scores, 48.2 ± 26.7 vs 44.1 ± 25.2; P = .55). The mean patient age at the time of the SEVEN-UP visit was 82.1 years. The scatterplot of patient age against final ETDRS letter score indicates that older patients had significantly poorer visual outcomes (coefficient, −0.93; 95% confidence interval [CI], −1.75 to −0.11; P = .027).
Prior vision result at the completion of the ANCHOR or MARINA trials also was a significant predictor of final vision results at the SEVEN-UP visit approximately 5 years later. In Figure 1 , ANCHOR or MARINA month 24 exit vision for each subject is plotted against the final SEVEN-UP vision; better prior vision score was associated significantly with better final vision score in a bivariate analysis (coefficient, 0.97; 95% CI, 0.62 to 1.32; P < .001; correlation coefficient, 0.58) and was confirmed in multivariate analysis that accounted for the effects of anatomic factors (coefficient, 0.52; 95% CI, 0.07 to 0.96; P = .024; Table 1 ).
SD OCT studies were available from 56 subjects at the SEVEN-UP visit; OCT imaging had not been obtained uniformly as part of ANCHOR, MARINA, and HORIZON, and for the purpose of these analyses, no OCT data were used from those previous clinical trials. In a preliminary bivariate analysis of 53 eyes with interpretable OCT images ( Figure 2 ), the visual outcomes for eyes with the presence of intraretinal or subretinal fluid or both (72% of study eyes) corresponded closely to eyes without fluid (mean ETDRS scores, 44.1 ± 25.0 letters and 47.7 ± 22.7 letters, respectively; P = .61). There was a lack of a significant effect of intraretinal or subretinal fluid on final visual outcome while adjusting for other key anatomic factors in the multivariate regression analyses ( P = .97; Table 1 ).
Submacular fibrosis, an end-stage characteristic of choroidal neovascularization and exudation, was present in 35 (61%) of 65 study eyes and was localized under the foveal center in 39%. In a bivariate analysis ( Figure 2 ), comparable vision outcomes were found between the group with subfoveal fibrosis and the group without (mean ETDRS letter scores, 42.3 ± 29.3 vs 50.9 ± 20.7; coefficient, −8.58; 95% CI, −21.97 to 4.81; P = .20). Furthermore, in the multivariate regression analyses of the 4 key anatomic end points, no significant effect for subfoveal fibrosis was found in terms of final vision at SEVEN-UP ( P = .75; Table 1 ).
On SD OCT measurements of macular thickness, the mean CST (average thickness of central 1 mm, measuring neurosensory retina only) was 170.0 μm. Seven percent of study eyes had macular thickening with CST of more than 250 μm. Macular thinning, defined as CST less than 200 μm, was seen in 75%. Although the isolated comparison of macular thickness categories against ETDRS letter score suggested reduced vision outcomes in both the group with macular thinning and in the group with macular thickening (mean ETDRS letter scores, 44.3 ± 21.7 and 28.0 ± 18.4, respectively, compared with normal group score of 67.9 ± 25.3; Figure 2 ), no significant effect was borne out when adjusting for the effects of other key anatomic factors in the multivariate regression analysis for final vision at SEVEN-UP visit (macular thinning, P = .25; macular thickening, P = .06; Table 1 ).
Multivariate regression analysis was performed to assess the independent effects of selected clinical and anatomic factors. For the final visual outcome at the SEVEN-UP visit, Table 1 shows the results of a regression model on 46 study eyes with complete data available ( R 2 = 0.46). The only macular anatomic variable demonstrating a significant association with final vision was macular atrophy lesion size, measured as area of definite decreased fundus autofluorescence (coefficient, −1.63; 95% CI, −2.45 to −0.81; P < .001), indicating that for each 1 mm 2 in area of macular atrophy, there was an associated vision decrement of 1.6 ETDRS letters.
A predetermined subanalysis assessed the incidence, severity, and progression of macular atrophy, defined clinically as a demarcated area of absent RPE pigmentation, window defect on FA, and decreased autofluorescence on FAF imaging. Macular atrophy was nearly ubiquitous by the time of the SEVEN-UP visit, detected on FAF imaging in 98% of study eyes. Figure 3 shows examples of 4 patients. From the month 24 exit visit in the ANCHOR or MARINA trials (year 2) to the SEVEN-UP visit (year 7), the macular atrophy in each case increased in extent. To quantify macular atrophy progression over time, an analysis compared individual patients at their year 2 visit and their year 7 visit based on red-free images and FA ( Table 2 ). Prior FA studies from ANCHOR or MARINA month 24 were available for 48 patients from the Wisconsin Reading Center dataset, 40 of which allowed accurate measurements. The progression analysis showed a visual decline of 21 letters over this mean 5.4-year interval. Every eye showed an increase in macular atrophy, with 57% increasing by 1 disc area or more. At year 2, in 22.5% of eyes the fovea was spared from macular atrophy, but by year 7, only 5% were fovea sparing.
|Variable||Coefficient||Confidence Interval||P Value|
|Visual acuity at year 2 (letter score)||+0.52||0.07 to 0.96||.02|
|Intraretinal and/or subretinal fluid||−0.19||−11.65 to 11.26||.97|
|Macular thinning a||−9.12||−24.92 to 6.67||.25|
|Macular thickening b||−27.04||−55.51 to 1.43||.06|
|Subfoveal fibrosis||−1.84||−13.39 to 9.71||.75|
|Macular atrophy||−1.63||−2.45 to −0.81||<.001|