We studied prevalence of depressive symptoms in primary care (PrC) and in psychiatric outpatient care (PsC), and how psychotic and manic symptoms are associated with current depressive symptoms. Altogether 563 patients attending PrC and 163 patients attending PsC filled in a questionnaire including the Depression Scale (DEPS), the Mood Disorder Questionnaire (MDQ) and questions on psychotic symptoms from the Composite International Diagnostic Interview (CIDI). Patients with depressive symptoms (DEPS score > 8) were interviewed by phone using the same checklist 6 months after baseline examination. From the PrC sample, 19.5% and from the PsC sample 73.0% were DEPS positive. In the PrC but not in the PsC sample, patients' background associated strongly with occurrence of depressive symptoms. Both at baseline and at follow-up, depressive symptoms correlated significantly with psychotic and manic symptoms. In multivariate analyses, when the effects of background, health and functioning were taken into account, baseline depressive symptoms associated significantly with lifetime psychotic symptoms. Depressive symptoms at follow-up associated significantly with psychotic symptoms during the follow-up period. In the PrC sample, this association was significant even when the effect of baseline depressive symptoms was controlled. About one-fifth of patients attending primary care and about three-quarters of patients attending psychiatric outpatient patient care suffer from depressive symptoms. Vulnerability to psychosis, indicated by occurrence of psychotic symptoms, increases the risk of and slower recovery from depressive symptoms in the patients attending primary care. Therefore, vulnerability to psychosis should be evaluated when treatment intervention for patients with depressive symptoms is planned.