Polygenic score analyses on antidepressant response in late-life depression, results from the IRL-GRey study

Abstract

Late-life depression (LLD) is often accompanied by medical comorbidities such as psychiatric disorders and cardiovascular diseases, posing challenges to antidepressant treatment. Recent studies highlighted significant associations between treatment-resistant depression (TRD) and polygenic risk score (PRS) for attention deficit hyperactivity disorder (ADHD) in adults as well as a negative association between antidepressant symptom improvement with both schizophrenia and bipolar. Here, we sought to validate these findings with symptom remission in LLD. We analyzed the Incomplete Response in Late Life Depression: Getting to Remission (IRL-GRey) sample consisting of adults aged 60+ with major depression (N = 342) treated with venlafaxine for 12 weeks. We constructed PRSs for ADHD, depression, schizophrenia, bipolar disorder, neuroticism, general intelligence, antidepressant symptom remission and antidepressant percentage symptom improvement using summary statistics from the Psychiatric Genomics Consortium and the GWAS Catalog. Logistic regression was used to test the association of PRSs with venlafaxine symptom remission and percentage symptom improvement, co-varying for the genomic principal components, age, sex and depressive symptoms severity at baseline. We found a nominal (i.e., p value ≤ 0.05) association between symptom remission and both PRS for ADHD and (OR = 1.36 [1.07, 1.73], p = 0.011) and PRS for bipolar disorder (OR = 0.75 [0.58, 0.97], p = 0.031), as well as between percentage symptom improvement and PRS for general intelligence (beta = 6.81 (SE = 3.122), p = 0.03). However, the ADHD association was in the opposite direction as expected, and both associations did not survive multiple testing corrections. Altogether, these findings suggest that previous findings regarding ADHD PRS and antidepressant response (measured with various outcomes) do not replicate in older adults.