Estimation of disease onset for patients with geographic atrophy due to age-related macular degeneration

Abstract

Introduction: Patients of the longitudinal FAM Study receive a follow-up after the diagnosis of AMD. The patient's natural course of geographic atrophy (GA) is observed and related to lifestyle factors. Since patients are recruited after disease onset it is of interest to estimate the patient's age at disease onset and its link to risk factors. Material and Methods: Individual GA time courses are modelled by linear mixed effect models which allow to fit linear and non-linear curves and to extrapolate the individual GA time course to the time of onset for each patient eye. Linear or quadratic time course models result in specific onset scenarios and age of onset with regard to putative risk factors (e.g. smoking history, BMI, gender, hypertension). These onset distributions can be compared with data from other studies and used for the appropriate choice of a time course model for the FAM data. GAMLSS models assess differences between the onset distributions with respect to underlying risk factors. Results: 179 patients and 279 eyes with GA were observed over a follow-up time with IQR 1.01–3.57 years. Given the GA size at baseline and the individual linear (quadratic) growth curve, the onset of the disease process was estimated to start on average 4.75 y (5.31 y) before time of first GA measurement. The median age at onset is 69.80 (69.85) (IQR: 63.95–75.15; 64.33–75.55). In the linear model the density functions of age at onset in the exposed vs. non exposed showed differences with respect to gender (p=0.008) and BMI>/≤30 (p=0.088). Discussion: We try to validate onset scenarios by comparing the estimated distribution with data of the Munster Aging and Retina Study and the Rotterdam Study. Risk factors measured at baseline are considered as surrogate parameters for lifestyle factors in the past.