DOP09 Phenotypic and genetic determinants of medication response in patients in the UK IBD BioResource

Abstract

Abstract Background Inflammatory bowel disease (IBD) exhibits heterogeneity at genetic, phenotypic, and therapeutic levels [1]. Although several studies have investigated genetic effects on IBD subtypes and drug adverse events [2,3], few have comprehensively explored the phenotypic and genetic determinants of IBD drug efficacy in a sufficiently powered cohort. Methods We used data from 32,199 patients in the UK IBD BioResource to investigate the effects of clinical phenotypes and demographics on drug efficacy and combined this with genome-wide genetic data for a subset of 11,536 individuals (Table 1). Drug efficacy was defined using a combination of clinician reported efficacy and persistence on drug without treatment escalation. Anti-TNF, thiopurine, steroids, and mesalazine were explored. We estimated phenotypic effects on drug efficacy using multivariable logistic regression and the genetic effects by genome-wide logistic regression. To explore whether drug efficacy and IBD disease susceptibility share a genetic basis, we estimated the proportion of variance in drug efficacy explained by known IBD risk variants [4]. Associations with Bonferroni corrected P-values < 0.05 were defined as statistically significant in phenotypic analyses and a genome-wide significance threshold of P=5x10-8 was adopted for genetic analyses. Results Drug efficacy was generally lower in patients with Crohn’s disease (CD) compared to those with other subtypes (OR ranges from 0.40 to 0.79), but anti-TNF showed a higher efficacy rate (OR = 1.21) in CD patients. Increasing age at diagnosis was associated with evidence of increased efficacy of thiopurine and mesalazine (Table 2). We found evidence of a genetic contribution to variation in drug efficacy for most drugs studied. However, known IBD risk variants showed little contribution (Figure 1). Using genome-wide association testing, we identified three loci showing a significant association with drug efficacy; two were related to steroid response and one to thiopurines (Table 3). None of these was an IBD disease susceptibility locus (P > 0.05). Conclusion Using a large, well-characterised cohort we found both genetic and phenotypic determinants of drug efficacy. Three loci were reported to be associated with drug efficacy in the first phase of the genetic analysis; at least 4,200 extra genotyped samples will be included before the ECCO meeting, thus increasing the power to detect additional loci. Our results suggest the genetic causes of drug efficacy and disease susceptibility are largely independent. These findings may provide opportunities for exploring the biology of drug efficacy and improving medication prioritization in IBD patients.