To evaluate the effect of select baseline characteristics on geographic atrophy (GA) progression in eyes receiving intravitreal pegcetacoplan or sham.
Phase 2 multicenter, randomized, single-masked, sham-controlled trial.
Patients with GA received 15 mg pegcetacoplan monthly or every other month (EOM), or sham injection monthly or EOM for 12 months. Primary efficacy endpoint was change in GA lesion size (square root) from baseline. Post hoc analysis evaluated the effects of age; gender; lesion size, focality, and location (extrafoveal vs foveal); pseudodrusen status; best-corrected visual acuity (BCVA); and low-luminance deficit (LLD) on GA progression at Month 12.
Of 246 randomized patients, 192 with 12-month data were included in this analysis. Overall mean (standard deviation) change in lesion size (mm) was 0.26 (0.17) ( P < .01), 0.27 (0.27) ( P < .05), and 0.36 (0.21) in the monthly pegcetacoplan (n = 67), EOM pegcetacoplan (n = 58), and sham (n = 67) groups, respectively. In univariate analysis, patients with extrafoveal lesions ( P < .001), BCVA ≥20/60 ( P = .001), and larger LLD ( P = .002) had greater mean changes in lesion size. Multivariate analysis confirmed significant association of extrafoveal lesions ( P = .001) and larger LLD ( P = .023) with GA progression. Monthly and EOM pegcetacoplan significantly reduced progression ( P < .05) when controlling for these risk factors.
Extrafoveal lesions and larger LLD are potential risk factors for GA progression. Pegcetacoplan treatment significantly controlled GA progression even after accounting for these risk factors.
G eographic atrophy (GA) secondary to age -related macular degeneration (AMD) results in loss of retinal pigment epithelium, photoreceptors, and underlying choriocapillaris, which lead to progressive irreversible vision loss. Furthermore, approximately half of patients with unilateral GA may progress to bilateral GA within 7 years of diagnosis. Currently, no approved therapies exist to halt or slow the progression of GA lesions and ensuing vision loss. Although Age-Related Eye Disease Study (AREDS) dietary supplements are often recommended for patients with late AMD to reduce the risk of developing exudative AMD, both the AREDS and the Age-Related Eye Disease Study 2 (AREDS2) formulations have demonstrated no apparent effect on the progression of GA. ,
Multiple factors have been associated with an increased rate of GA progression, including GA lesion size, focality (uni-/multifocal nature), and location (foveal vs extrafoveal); junctional zone fundus autofluorescence patterns; reticular pseudodrusen; low-luminance deficit (LLD); and increasing percentages of surrounding choriocapillaris flow deficits. , , Although the majority of this evidence has been derived from natural history studies, it is clinically relevant to understand the impact of these characteristics on GA progression in studies involving therapeutic intervention.
Recently, intravitreal pegcetacoplan—a pegylated complement C3 inhibitor peptide—demonstrated statistically significant reductions in the growth of GA lesions compared with sham treatment in a Phase 2 trial (FILLY trial [ClinicalTrials.gov identifier NCT02503332]) for GA secondary to AMD. Using this data set, we conducted an exploratory post hoc analysis to study the role of select baseline characteristics on GA progression across the treatment groups.
The FILLY trial was a Phase 2, prospective, multicenter, randomized, single masked, sham injection–controlled study to assess the safety, tolerability, and efficacy of intravitreally administered pegcetacoplan in patients with GA secondary to AMD. The detailed study design has been published previously, and the trial is registered with ClinicalTrials.gov (identifier NCT02503332). A total of 246 patients were randomized in a 2:2:1:1 ratio to receive 15 mg pegcetacoplan monthly, 15 mg pegcetacoplan every other month (EOM), sham injection monthly, or sham injection EOM for 12 months. The study was performed in accordance with the Declaration of Helsinki, International Conference on Harmonization Good Clinical Practice guidelines, and all applicable regulations. Institutional review board or ethics committee approval was obtained at each site, and all patients provided written informed consent.
Eligible patients were aged >50 years with best-corrected visual acuity (BCVA) in the study eye of 24 letters or better using Early Treatment Diabetic Retinopathy Study (ETDRS) charts (approximately 20/320 Snellen equivalent), diagnosis of GA secondary to AMD confirmed using fundus autofluorescence imaging with GA area size of ≥2.5 mm 2 and ≤17.5 mm 2 , presence of any pattern of autofluorescence in the junctional zone of GA, and at least 1 focal lesion ≥1.25 mm 2 if GA was multifocal. An independent central reading center (Digital Angiography Reading Center, Great Neck, NY) confirmed lesion eligibility and assessed all images, in a masked fashion, throughout the duration of the study. Assessments were taken as the mean of GA lesion area measurements from 2 independent readers or the median of measurements from 3 independent readers.
The primary efficacy endpoint was the change from baseline to Month 12 in the square root of the GA lesion area as assessed using fundus autofluorescence imaging. Results on the primary and secondary endpoints at Month 12 have been previously published. This study presents a post hoc analysis of the effect of key baseline characteristics on GA progression at Month 12 for patients with both baseline and 12-month data, including age; gender; GA lesion size, focality, and location (foveal vs extrafoveal, determined by fundus autofluorescence imaging); pseudodrusen status (present vs absent, deemed present for ≥1 pseudodrusen detected on near-infrared reflectance imaging); BCVA; and LLD.
One hundred ninety-two patients with GA lesion area assessed at baseline and Month 12 were included in this post hoc analysis (pegcetacoplan monthly [n = 67], pegcetacoplan EOM [n = 58], and sham [n = 67]). In the binary and tertile analyses, subgroups for each baseline characteristic were defined as follows: age (by tertile, grouped as <77 years, ≥77 to <83 years, and ≥83 years); gender (male or female); untransformed GA lesion size area (by tertile, grouped as <5.55 mm 2 , ≥5.55 to <10.55 mm 2 , and ≥10.55 mm 2 ); focality (unifocal vs multifocal); GA location (foveal vs extrafoveal); pseudodrusen status (present vs absent); BCVA (by tertile, grouped as <53 ETDRS letters, ≥53 to <68 ETDRS letters, and ≥68 ETDRS letters); and LLD (by tertile, grouped as <17 ETDRS letters, ≥17 to <30 ETDRS letters, and ≥30 ETDRS letters). Forest plots were used to illustrate treatment differences for each baseline characteristic. Analyses with baseline characteristics divided into categories and subgroups by treatment arms are descriptive in nature.
Overall, 176 patients without any missing baseline characteristics were included in univariate and multivariate analyses. Univariate analysis was performed using 1-way analysis of variance or linear regression for categorical and continuous predictors, respectively. Multivariable analysis of covariance was performed including all potential risk factors (identified in univariate analysis) and excluding treatment designation in the model. A 2-sided alpha level of 0.05 was used, with no adjustments for multiple comparisons. The risk factors included in the multivariate analysis were those that were statistically significant in the univariate analyses as well as gender and age. The presence of interactions among the risk factors identified in univariate analysis was examined. Final multivariable analysis of covariance was performed to assess treatment differences controlling for risk factors identified in the previous step using a significance level of 0.05. SAS 9.4 software (SAS Institute, Cary, NC) was used for analysis.
Figure 1 depicts a flowchart of the number of patients used for analysis by treatment group. Of 246 randomized patients, 192 had FAF measurements of GA area at both baseline and Month 12; after excluding patients with missing baseline lesion data (location, focality, and presence of pseudodrusen), the final sample size for this analysis was 176. Overall mean change in GA lesion size (mean [standard deviation] in millimeters) from baseline to Month 12 in patients included in this analysis was 0.26 (0.17) ( P < .05), 0.27 (0.26) ( P < .05), and 0.36 (0.21) in the monthly pegcetacoplan, EOM pegcetacoplan, and sham, respectively ( Figure 2 ).
Evaluation of Effect of Key Baseline Characteristics
Gender and age
The mean change in the square root of GA lesion size area by gender is also illustrated in Figure 2 . The majority (60%-65%) of patients were female, and a higher rate of progression was observed in the female subgroup that received sham injections. No specific trend was observed in the male subgroup. The mean change in the square root of GA lesion size area by age tertile (<77 years, ≥77 to <83 years, and ≥83 years) is shown in Figure 2 . The rates of progression trended to be higher in the sham group for age <77 years and age ≥83 years, but not for the age tertile ≥77 to <83 years compared with the pegcetacoplan-treated groups.
Lesion characteristics—size, focality, and location
The mean change in the square root of GA lesion size area for each tertile of untransformed baseline GA area (<5.55 mm 2 , ≥5.55 to <10.55 mm 2 , and ≥10.55 mm 2 ) is shown in Figure 3 A, demonstrating that the rate of progression was higher in the sham group for all lesion size categories. Figure 3 B shows that the majority (62%-72%) of patients presented with multifocal lesions, and growth rates were lower in the pegcetacoplan treatment groups regardless of lesion focality (unifocal or multifocal). Similar trends were observed for the foveal and extrafoveal lesion location subgroups ( Figure 3 , B). The statistical significance of monthly vs EOM dosing of pegcetacoplan with respect to unifocal lesions and extrafoveal lesions could not be assessed because of the small sample sizes.
The majority (60%-72%) of participants had pseudodrusen present at baseline per the reading center definition. Irrespective of baseline pseudodrusen status, the GA growth rate was lower in pegcetacoplan-treated groups compared to sham ( Figure 4 ).