Pharmacogenetic associations in neovascular age-related macular degeneration. Results from the IVAN study

Abstract

Purpose: To determine if pre-specified genetic polymorphisms influence treatment responsiveness to VEGF inhibition in patients with neovascular age-related macular degeneration (nAMD).

Methods: Participants enrolled in the IVAN clinical trial and undergoing treatment for nAMD with either ranibizumab or bevacizumab were classified as treatment responders or non- responders based on the optical coherence tomography (OCT) metric of total retinal thickness (TRT). We computed the change in TRT from baseline to the latest time point for which OCT data was available (3, 6, 9 or 12 months). Study eyes at or above the 75th percentile for change were classified as responders and those at or below the ≤25th percentile for change as non-responders. Two analyses were performed using data for responders and non-responders. Firstly, we tested for replication of three previously reported pharmacogenetic associations of response to VEGF inhibition in nAMD at the CFH, FZD4 and HTRA1/ARMS2 loci. Secondly, we tested an additional 500 SNPs for pharmacogenetic associations using a candidate gene approach. These analyses compared single nucleotide polymorphism (SNP) frequencies in responder and non-responder groups (2 x 2 tables), in accordance with a prespecified analysis plan.

Results: 512 patients were available with both phenotype and genotype information. Three were removed from further analysis due to differing ethnicity. 423 had OCT data available at 12 months, 53 at 9 months, 20 at 6 months and 13 at 3 months. 255 were in the middle 50% and not analysed for the main “case/control” association tests, 126 were classified as non-responders and 128 as responders. SNP rs10490924 in HTRA1/ARMS2 showed a borderline association after Bonferroni correction (pc = 0.06; p = 0.02 without correction). For the 500 additional SNPs tested for association, none were significant after Bonferroni correction for multiple testing. The smallest p-value was 0.0005 (pc =0.24) for rs9679290 in the EPAS1 (HIF2A) gene on chromosome 2.

Conclusions: Our analysis tested pharmacogenomic association using high quality phenotype data from a randomised controlled trial of nAMD. SNPs investigated covered several biological pathways which could be plausibly linked to nAMD. No replication of previous associations or other significant associations was observed.

Keywords: 412 age-related macular degeneration • 464 clinical (human) or epidemiologic studies: risk factor assessment • 539 genetics