Associations between mood, anxiety or substance use disorders and inflammatory markers after adjustment for multiple covariates in a population-based study.