Patterns of genetic liability in individuals treated for depression in different care settings
K. L. Musliner, J. Ø. Sørensen, E. Agerbo, A. J. SchorkIntroduction
Genetic studies of psychiatric disorders often select cases using information from treatment settings such as primary and secondary care. Although generally acknowledged that these patient populations vary in terms of clinical characteristics such as severity, the potential for differences in their underlying molecular genetic liabilities has not been explored.
Objectives
To examine polygenic profiles of a range of psychiatric phenotypes in individuals treated for depression in primary care only, secondary care only, or both primary and secondary care.
Methods
Data were obtained from the iPSYCH case-cohort study, which includes 141,265 individuals born in Denmark between 1981–2008. Psychiatric diagnoses from secondary care and antidepressant prescriptions from primary care were linked to genetic data via national registers. Individuals who died, emigrated or received a bipolar disorder (BD) or schizophrenia (SCZ) diagnosis before their 10 th birthday were excluded. Depression was classified by treatment setting (primary care only, secondary care only, both primary and secondary care). Polygenic scores (PGS) were calculated using LDpred2. Individuals were followed from their 10 th birthday until their first depression treatment, emigration, death, or December 31, 2018, whichever came first. Hazard ratios were calculated using weighted Cox regressions with robust standard errors adjusted for birth year, sex and the first 5 principal components.
Results
The sample included 117,305 individuals (49% female), aged 10–35 at follow-up. 12.3% were treated for depression: 8.0% in primary care only, 0.5% in secondary care only and 3.7% in both. Figure 1 shows the hazard ratios for PGSs and depression stratified by treatment setting. PGSs for depression and neuroticism showed the strongest associations with depression with only minor variation across treatment settings. The effect of PGS-depression was strongest for secondary care only (HR=1.49, 95% CI [1.36-1.64]) while the effect of PRS-neuroticism was strongest for treatment in both settings (1.27 [1.22-1.32]). Associations for PGSs for the p-factor, BD and SCZ were notably elevated for secondary care only (p factor: 1.84 [1.60-2.10]; BD: 1.35, [1.22-1.51]; SCZ: 1.31 [1.18-1.46]) relative to the primary care only (p factor: 1.29 [1.24-1.35]; BD: 1.11, [1.07-1.15]; SCZ: 1.12 [1.08-1.16]) and those treated in both settings (p factor: 1.62 [1.54-1.71]; BD: 1.15, [1.10-1.21]; SCZ: 1.18 [1.13-1.24]).
Image 1: Long description.