76. GENETIC STUDIES FOR SSRI ANTIDEPRESSANT RESPONSE: DEFINING PROXY PHENOTYPES IN ELECTRONIC HEALTH RECORDS IN UK BIOBANK

• Published in: European neuropsychopharmacology 2024-10, Vol.87, p.91-92
• Main Authors: Lo, Chris Wai Hang, Gillett, Alexandra C., Kamp, Michelle, Fabbri, Chiara, Wong, Win Lee Edwin, Handley, Dale, Iveson, Matthew H., Pain, Oliver, Vassos, Evangelos, Li, Danyang, Young, Allan H., Lewis, Cathryn M.
• Format: Article
• Language: English
• ISSN: 0924-977X
• DOI: 10.1016/j.euroneuro.2024.08.190

Selective serotonin reuptake inhibitors (SSRIs) are the first-line antidepressant treatment for major depressive disorder (MDD). A previous genome-wide association study (GWAS) meta-analysis of clinical trials estimated the heritability for remission after antidepressant treatment to be 13-40%, but no individual variants were identified. Electronic health records offer larger and more representative data in which to define outcomes for antidepressant treatment. Here we use SSRI drug switching as a proxy for antidepressant non-response in the UK Biobank (UKB), performing phenotypic analyses and GWAS to investigate the demographic and genetic profiles of SSRI treatment response. Using primary care prescription records from UKB, SSRI switchers were defined as participants prescribed an SSRI, who were then prescribed a different antidepressant within 90 days. Non-switchers were UKB participants with at least three consecutive prescriptions for an SSRI. MDD were defined by GP diagnoses records of depression from primary care. Associations between demographic characteristics of SSRI switchers and non-switchers were tested by logistic regression. Mean differences between switcher status and polygenic scores (PGSs) for antidepressant non-remission, MDD and schizophrenia (computed by PRS-CS) were also compared. A GWAS for SSRI switching was performed using REGENIE, with MDD defined using either one or two primary care diagnostic codes for depression. SNP-based heritability (h2) was estimated using genome-based restricted maximum likelihood-based methods, including GCTA and GCTB. In UKB, 5,133 SSRI switchers and 33,680 non-switchers were identified. Requiring one depression diagnostic code identified 2,868 SSRI switchers, and 18,360 non-switchers, reducing to 2,114 switchers and 12,667 non-switchers with two diagnostic codes. Higher education levels and annual income were associated with lower odds of switching (p