OBJECTIVE: To evaluate the influence of extensive genetic and psychosocial confounding on the association between early childhood infection and five major psychiatric disorders.

METHODS: A case-cohort study including participants from the Danish iPSYCH2012 sample, a case-cohort sample where all cases born between May 1, 1981, and December 31, 2005, diagnosed with ADHD, autism spectrum disorder (ASD), bipolar affective disorder (BIP), affective disorder (MDD) or schizophrenia (SCZ), were identified and pooled with a representative sample (subcohort) of the Danish population. We used Cox proportional hazards regression customized to the case-cohort setup to calculate hazard ratios (HRs) of outcome with 95% confidence intervals (CIs), following exposure to early childhood infection before the age of 5 for ADHD and ASD, and before the age of 10 for BIP, MDD and SCZ. To evaluate psychosocial confounding we included sex, calendar period, sibling infections, urbanicity, parental socio-economic status, parental mental health information and polygenic risk scores for all five disorders, as covariates. To estimate how liability for psychiatric disorders measured through the PRS influenced the risk of early childhood infection, we calculated odds ratios (ORs) with 95% CIs, using logistic regression.

RESULTS: Early childhood infection was associated with ADHD, ASD, MDD and SCZ with number of childhood infections increasing the hazard. The HR was still significant in the model with full adjustments after 1 infection for ADHD (HR 1.29, 95% CI: 1.19 - 1.41), ASD (HR 1.28, 95% CI: 1.18 - 1.40), MDD (HR 1.23, 95% CI: 1.14 - 1.33) and SCZ (HR 1.21, 95% CI: 1.07 - 1.36), but not for BIP (HR1.17, 95% CI: 0.96 - 1.42). Probands exposed to 4 or more sibling infections, but not own infection had an absolute risk of ADHD, BIP, MDD and SCZ that closely approached the absolute risk for individuals exposed to 4 or more own infections. We found evidence of gene-environment correlation with higher PRS of MDD and to some extent SCZ increasing the risk of infections and higher PRS of BIP associated with significantly decreased risk.

CONCLUSION: Early childhood infection is significantly associated with ADHD, ASD, MDD and SCZ and not explained by genetic or psychosocial confounding. Although we found evidence of gene-environment correlation, it had minor impact on the results.