Pharmacological profiling of major depressive disorder-related multimorbidity clusters

We previously identified seven distinct multimorbidity clusters associated with major depressive disorder through a comprehensive analysis of 1.2 million individuals of multiple cohorts. These clusters, characterized by unique clinical, genetic, and psychiatric and somatic illness risk profiles, implicate divergent treatment pathways and disease management strategies. This study aims to deepen the understanding of these clusters by analyzing drug prescriptions, evaluating the effectiveness of antidepressant treatment strategies, and identifying potential markers for personalized medicine. Utilizing drug prescription data in the format of ATC codes, we performed epidemiological assessments, including multimorbidity (number of diseases), polypharmacy (number of chemical substances), and drug burden (number of prescriptions) analyses across the clusters. We applied linear regression models to assess strength and predictive capability of cluster membership on various metrics, and logistic regression to explore associations with treatment-resistant depression. We also quantified and visualized common antidepressant treatment sequences within each cluster. Our findings indicate significant variations in polypharmacy and drug burden across clusters, with distinct patterns emerging that correlate with the clusters’ profiles. Clusters liable to multimorbidity have higher drug burden, even after correction for number of diseases. Furthermore, the three clusters with higher risk for MDD showed different antidepressant treatment profiles; two required significantly more antidepressant prescriptions and had a higher risk for TRD. The detailed pharmacological profiling presented in this study not only corroborates the initial cluster definitions but also enhances our predictive capabilities for treatment outcomes in MDD. By linking pharmacological data with comorbidity profiles, we pave the way for targeted therapeutic interventions.