Polygenic Score for Cardiometabolic and Psychiatric Phenotypes Predict Response to Treatment with Selective Serotonin Reuptake Inhibitors in Major Depressive Disorder

Abstract

Selective serotonin reuptake inhibitors (SSRIs) are the first line antidepressants to treat major depressive disorder (MDD). Among other factors, the co-morbidity of depression with other psychiatric and cardiometabolic disorders may negatively influence antidepressants treatment response. Moreover, response to antidepressants is considered as a complex trait with substantial contribution from a large number of common genetic variants of small effect. Hence, polygenic scores (PGSs) could be a useful tool to estimate the overall effect of genetically driven factors on the treatment outcome.Here, we evaluated whether PGSs derived from 25 recognized comorbid cardiometabolic and psychiatric disorders, personality traits and educational parameters predict treatment response to SSRIs in MDD. In order to create the PGSs, we used genome-wide genotype data obtained from 865 MDD patients who received SSRIs treatment and GWAS summary statistics for the 25 disorders and phenotypes. The PGSs were calculated within a P-value threshold (PT) as the sum of the single nucleotide polymorphisms effect allele multiplied by its corresponding GWAS effect size (β- coefficient or log (OR)) weighted by the sum of the GWAS effect sizes.Clinical, demographic and SSRIs treatment response data were collected from 865 MDD patients by the ISPC. Baseline and follow-up SSRIs treatment response were measured using the 17-item Hamilton Rating Scale for Depression (HRSD-17). After 4 weeks of treatment, patients with at least 50% reduction in HRSD-17 from the baseline score were classified as responders and patients with a score of less than or equal to seven were considered as remitters. Once the PGSs were calculated, binary logistic regression analyses were performed between the dependent variables (response or remission to SSRIs) and the PGSs in each of the 25 disorders and phenotypes adjusted for the covariates (age, sex and study sites). Finally, we reported the best-predictive PGS for each disorder and phenotype that predicts the treatment outcome with the smallest p-value. A statistically significant association was determined at P < 0.05.We found that the PGSs for MDD, depressive symptoms, type 2 diabetes, body mass index, plasma levels of fasting proinsulin or fasting glucose, openness personality trait, subjective wellbeing and educational attainment significantly (P < 0.05) predicted both response and remission to SSRIs treatment in MDD patients. Additionally, the PGS for coronary artery disease, plasma level of total cholesterol, agreeableness and conscientiousness personality traits showed a significant association with remission, but not with response. Our findings suggest that the influence of cardiometabolic and psychiatric co-morbidities on antidepressants treatment response could be at least partly determined at a genetic level.