Latent Correlations Between Longitudinal Systemic Complement Factors and Geographic Atrophy Growth Rate

Purpose: Investigate associations between square root-transformed (SQRT) geographic atrophy (GA) growth rate and longitudinal systemic complement levels in non-neovascular AMD. Methods: Prospective longitudinal cohort study of patients with GA in the University of Colorado AMD Registry enrolled from August 2014 to September 2022, with ≥1month of follow-up imaging through April 2023. Systemic complement factor level ratios (Bb/Factor B, Ba/Factor B, Factor H, Factor I, C4b/C4, C3a/C3, C3b/C3, sC5b-9/C5) were collected approximately annually. SQRT GA growth rate was quantified on fundus autofluorescence. Correlations between GA growth rate and longitudinal biomarker levels were evaluated using a Bayesian hierarchical joint model. Results: One hundred fifty-five eyes from 88 patients were included. Average SQRT GA growth rate was 0.27mm/year over a mean of 3.3 years and was positively associated with female sex. Ba/Factor B (correlation [credible interval {CrI}]:0.35, 95% CrI: 0.02 to 0.64), and Bb/Factor B (0.20, 95% CrI: -0.086 to 0.48) demonstrated posterior distributions largely above zero, consistent with a positive association. Posterior distributions for Factor I (-0.08, 95% CrI: -0.35 to 0.20) and C3a/C3 (-0.14, 95% CrI: -0.44 to 0.18) were centered below zero, whereas those for Factor H (0.09, 95% CrI: -0.20 to 0.37) and sC5b-9/C5 (0.09, 95% Crl: -0.19 to 0.37) were centered above zero. Conclusions: Systemic alternative pathway biomarkers demonstrated the strongest positive associations with GA growth rate. The pattern of associations in this exploratory analysis supports a hypothesis-generating framework in which differences in alternative pathway regulation may be associated with heterogeneity in GA progression. Future larger studies are needed for external validation of this hypothesis-generating work.