Identifying endophenotypes associated with age-related macular degeneration in the Amish using OCT

Abstract

 

Purpose

 

Several genetic and environmental risk factors of age-related macular degeneration (AMD) have been identified, yet a substantial portion of variation in disease risk and heritability remains unexplained. The unexplained variation may partly reflect challenges defining the complex phenotype. We used Ocular Coherence Tomography (OCT) to quantify fine-scale features of AMD, and assess its potential to identify endophenotypes.

 

Methods

 

We sampled 359 related individuals from 116 Amish families across Pennsylvania, Ohio and Indiana. The Amish provide an excellent opportunity to analyze the heritability of complex traits given their large nuclear families. Initially we assessed the correlation between left and right eyes, defining the phenotype using both the traditional AREDS scale and OCT traits (subfoveal choroidal thickness and drusen area in a 5mm circle). We also measured the correlation between OCT traits and AREDS grade. Finally, we quantified the heritability (h², additive genetic variance) of OCT traits by fitting a polygenic model in SOLAR, accounting for age and gender.

 

Results

 

Approximately 41% of individuals were controls (grade 0 or 1), 33% had early AMD (grade 2), 20% had intermediate AMD (grade 3), 5% had geographic atrophy and 1% advanced neovascular AMD. AREDS grade (r_s=0.77, n=312), choroidal thickness (r_s=0.83, n=329) and drusen area (r_s=0.73, n=346) were highly correlated between eyes. Choroidal thickness (right eyes r_s=-0.33; left eyes r_s=-0.37) and drusen area (right r_s=0.48; left r_s=0.51) were moderately correlated with AREDS grade. Choroidal thickness was significantly heritable (right: h²=0.65±SE 0.15, n=318, p<0.01; left: h²=0.50±SE 0.15, n=316, p<0.01). For individuals with drusen, drusen area was not significantly heritable (left: h²=0.11 ±SE 0.30, n=87, p=0.35; right: h²=0.20 ±SE 0.34, p=0.27, n=84), although the sample size was small.

 

Conclusions

 

Our data confirm previous findings of strong correlation between eyes for most variables, with only drusen area not showing strong heritability. However, correlation coefficients also indicated some variation in the disease process, validating the inclusion of both eyes in future analyses. Moderate correlation between OCT traits and AREDS grade, and significant heritability of choroidal thickness highlighted the value of OCT for identifying endophenotypes, and furthermore, the value of Amish populations for such analyses.