Polygenic, sociodemographic, and clinical associations with five-year clinical outcomes in patients with major depressive disorder in a hospital setting: a danish nationwide study

Major depressive disorder (MDD) has a highly heterogeneous clinical course. The disorder is heritable, and research shows that treatment response and likelihood of recurrence also have a genetic basis. There is growing interest in prediction of treatment response from baseline characteristics. However, studies of symptom change following treatment often use a timeframe that is too short to assess long-term outcomes that may be equally important as immediate improvement. We examined associations with 5-year clinical outcomes in patients with MDD and included genetic information in addition to traditionally studied sociodemographic and clinical variables. Using nationally representative data from iPSYCH, we studied 13,244 patients (71% female, mean age at first hospital contact of 19 years) admitted to a Danish psychiatric hospital for MDD (ICD10 F32–F33). Follow-up began six months after discharge from their first episode of MDD in a hospital and lasted for 5 years. We categorized patients based on clinical outcomes during this period: readmission for another MDD episode, readmission for another psychiatric disorder, use of antidepressants, or no further contact with the healthcare system for a psychiatric disorder. This categorization was hierarchical and groups were mutually exclusive. Sociodemographic, clinical, and genetic risk factors were tested for associations with 5-year outcomes, with “no further contact” as the reference. Multinomial logistic regression models incorporated all risk factors. The first five ancestry principal components, genotyping batch, and calendar year of diagnosis were included as covariates. We used a Bonferroni-adjusted alpha to account for multiple testing (0.05/16=0.003). The strongest predictor was prior use of antidepressant medication: receiving antidepressants afterwards (odds ratio [OR] = 2.39, 95% confidence intervals [CIs] = 2.38-2.49), readmission for another disorder (OR = 2.05, 95% CIs = 2.04-2.07), readmission for MDD (OR = 2.34, 95% CIs = 2.33-2.35). Compared to those with mild MDD at their first hospital episode, those with moderate MDD had an increased risk of receiving antidepressants (OR = 1.44, 95% CIs=1.43-1.45) and of readmission for MDD (OR = 1.47, 95% CIs = 1.45-1.48). Similarly, those with severe MDD first contact had an increased risk of receiving antidepressants (OR = 1.48, 95% CIs=1.46-1.50) and of readmission for MDD (OR = 1.84, 95% CIs = 1.82-1.86). In comparison to male sex, female sex was associated with increased risk for receiving antidepressants (OR = 1.39, 95% CIs=1.38-1.40), readmission for another psychiatric disorder (OR = 1.95, 95% CIs = 1.94-1.97), and readmission for MDD (OR = 1.80, CIs = 1.79-1.82). In terms of genetic risk factors, having a higher polygenic score for MDD was associated with readmission for MDD (OR = 1.12 per unit increase, 95% CIs = 1.11-1.3) and readmission for another psychiatric disorder (OR = 1.11 per unit increase, 95% CIs = 1.10-1.12), while having a parental history of MDD was associated only with readmission for MDD (OR = 1.29, 95% CIs = 1.28-1.31). Currently, sociodemographic and clinical risk factors are more robust predictors of clinical outcomes in hospital-treated MDD patients. Nonetheless, having a higher MDD polygenic score was associated with being readmitted to hospital for either another MDD episode or another psychiatric disorder. This supports previous research showing that individual polygenic liability influences the clinical course of MDD.