Abstract
CYP2C19 metabolizes various selective serotonin reuptake inhibitors (SSRIs) and genetic CYP2C19 variants are associated with SSRI tolerability and response. Yet, whether CYP2C19 variability also impacts citalopram response remained unclear. We here evaluated associations between CYP2C19 genotypes and citalopram prescription data of 11,079 patients from the UK Biobank. Importantly, CYP2C19 ultrarapid metabolizers (UMs) were more likely to experience therapeutic failure (OR 2.03, P = 0.002) and both CYP2C19 poor metabolizers (PMs) and UMs exhibited an increased likelihood of treatment resistance (PMs, OR 1.85, P = 0.06; UMs, OR 1.74, P = 0.05). Furthermore, inferred CYP2C19 metabolizer status showed significant monotonic trends with citalopram maintenance dose (from 19.7 mg/d in PMs to 24.4 mg/d in UMs; Jonckheere-Terpstra P = 0.017) and time to dose escalation (2.5 years in PMs to 1.7 years in UMs; P = 0.006). Besides CYP2C19, rare variant analyses identified BMP2K as a novel candidate locus for SSRI response. BMP2K variants were significantly associated with therapeutic failure in citalopram patients (P = 5.49 × 10-6) and this association was replicated in 220 escitalopram users (P = 0.042). Furthermore, higher BMP2K variant burden significantly correlated with lower maintenance dose. These findings provide the first report of a U-shaped association between CYP2C19 metabolizer status and citalopram therapeutic outcomes and pinpoint BMP2K as a novel candidate gene with potential contribution to interindividual differences in citalopram response. Collectively, these results demonstrate that CYP2C19 genotype is significantly associated with citalopram outcomes in a real-world setting in which patient treatment was not guided by pharmacogenetics, thus avoiding open-label awareness, eliminating expectation bias of genetic testing and highlighting the value of large-scale biobank data for advancing precision psychiatry.</p>