Predicting who will continue, switch, discontinue or require augmentation of antidepressant therapy remains challenging. This study investigated whether reported polygenic associations reflect drug-specific non-response or a broader tendency toward treatment modification. The authors analysed genetic and familial contributions to four real-world treatment outcomes—continuation, switching, discontinuation and augmentation—across two large cohorts to clarify the relationships between polygenic liability, clinical features, and family aggregation.
Two cohorts were analysed. The All of Us (AoU) cohort included 98,357 participants and the Pharmlines resource (Lifelines linked to IADB.nl) included 12,884 participants. Analyses were restricted to individuals with at least one antidepressant monotherapy episode lasting 28 days or longer. Treatment outcomes compared were: continued monotherapy (continuation), a change to a different antidepressant (switching), stopping antidepressant therapy (discontinuation), and augmentation defined as addition of an atypical antipsychotic or lithium.
The study evaluated associations between treatment outcomes and polygenic scores (PGS) for psychiatric disorders, including a depression PGS and scores for other disorders such as schizophrenia and bipolar disorder. Effect estimates reported include odds ratios (OR) per standard deviation (SD) change in PGS where available. The authors also tested whether PGS associations were drug-class-specific for switching, and whether genetic liability tracked progressively across ordered treatment outcomes.
Among participants with recorded major depressive disorder, switching—but not discontinuation—was consistently associated with comorbid anxiety, higher counts of depressive symptoms, and stress-related measures in both cohorts. These clinical correlations suggest that switching often reflects greater clinical severity or symptom burden rather than simple intolerance or immediate non-adherence. Augmentation was considered separately as an indicator of more complex or refractory illness leading to addition of antipsychotics or lithium.
A depression PGS was associated with antidepressant switching in All of Us (OR 1.16 per SD, 95% CI 1.12–1.20) and showed a concordant nominally significant estimate in Pharmlines (OR 1.11, 95% CI 1.01–1.22). In the AoU cohort, the depression PGS increased progressively across ordered treatment outcomes from continuation to switching to augmentation; the reported per-step association was OR 1.18 (95% CI 1.15–1.21), indicating a graded relationship between greater genetic liability for depression and movement toward more intensive or complex treatment strategies.
By contrast, schizophrenia and bipolar disorder PGS were selectively associated with augmentation, consistent with augmentation capturing cross-disorder psychiatric complexity rather than isolated depression liability. The authors report that no PGS showed drug-class-specific associations with switching, indicating that the polygenic signals identified did not map to non-response specific to particular antidepressant classes in this dataset.
Familial aggregation analyses in the Pharmlines cohort detected family-level clustering for continuation, including continuation specific to SSRIs and SNRIs. In contrast, familial aggregation was not detectable for switching or discontinuation. This pattern implies that sustained, switch-free continuation may reflect familial influences—genetic or shared environmental factors—whereas treatment modification behaviors such as switching do not show the same family-level clustering in Pharmlines.
The combined clinical and genetic findings led the authors to conclude that antidepressant switching partially indexes depression severity and symptom burden rather than representing only drug-specific pharmacologic non-response. Augmentation appears to capture greater cross-disorder psychiatric complexity, consistent with associations with schizophrenia and bipolar PGS. Familial aggregation was confined to sustained continuation, suggesting that families are more likely to share patterns associated with continued, switch-free treatment than patterns of switching or early discontinuation.
This report is a preprint and has not been certified by peer review; the authors note it should not be used to guide clinical practice. The article states that data are not available upon request. Ethical approvals were obtained from the All of Us Research Program Institutional Review Board and the Medical Ethical Committee of the University Medical Center Groningen (METc 2007/152). The authors declared no competing interests. Further methodological details, supplementary material and code/data links are referenced in the source but primary data access is restricted as reported.
Funding sources reported in the manuscript include the Dutch Ministry of Health, Welfare and Sport; the Dutch Ministry of Economic Affairs; University Medical Center Groningen and regional provinces; Michael Davys Trust; University of Oxford; the National Health and Medical Research Council; and the Wellcome Trust. The authors declared no competing interests in the source document.