Abstract
Purpose :
To evaluate the efficacy of future gene augmentation therapy trials for Stargardt disease, the endpoint needs to distinguish a treatment effect from the effect of time. Visual acuity demonstrates an extremely variable decline rate; yet, it may well be an appropriate endpoint in children progressing rapidly. We performed a retrospective, observational clinical study to determine if best-corrected visual acuity (BCVA) decline is an appropriate endpoint at different ages at presentation.
Methods :
We selected 218 patients (422 eyes) with clinically confirmed Stargardt disease supported by at least one variant in the ABCA4 gene. We collected BCVA over a maximum period of five years. Patients were grouped based on their age at presentation <18 years, 18–26 years, or >26 years. We estimated the BCVA decline using a linear mixed-effects model with time, subgroups, and an interaction term between time and subgroups as covariates. We performed sample size calculations for a theoretical intervention with 50% effect in slowing BCVA decline, using an unpaired t-test with a two-sided significance level of 0.05 and a power of 0.80.
Results :
A total of 1446 eye-visits were included with a median follow-up of 3.3 years (interquartile range [IQR], 3.8), 1.1 years (IQR 3.9) and 0.71 years (IQR 3.2) in the <18, 18-26 and >26 group, respectively (p<0.001). The average baseline BCVA was 0.29 logMAR (95% CI, 0.26 – 0.33) for the age subgroup >26 years, and was 0.20 logMAR (0.15 – 0.25) and 0.20 logMAR (0.14 – 0.25) higher for age groups < 18 years, and 18 – 26 years, respectively. The average BCVA decline was 0.098 logMAR/year (0.07 – 0.13). The model failed to detect differences between age groups.
To detect a 50% treatment effect within 2 years of follow-up, 860 patients are needed.
Conclusions :
Best-corrected visual acuity failed to identify differences in progression speed between children and adults, and thus does not support a preferred selection of children in therapeutic trials. However, failure in detecting differences in progression speed is likely because of the insensitivity of BCVA itself rather than the absence of such differences. As a large number is needed to identify a treatment effect, BCVA is unlikely to be a sensitive endpoint for short-term treatment efficacy.
This is an abstract that was submitted for the 2017 ARVO Annual Meeting, held in Baltimore, MD, May 7-11, 2017.