LAUSR

BACKGROUND We examined mortality in schizophrenia spectrum disorder (SSD) and non-schizophrenic psychosis (NSSD) compared to individuals without psychosis, and whether perinatal factors predict mortality. METHODS Within Northern Finland Birth Cohort 1966 (n=10 933; 203 with SSD, 178 with NSSD), mortality was followed until end of 2011 by national register. Wantedness of pregnancy, mother's antenatal depression, smoking and age, parity, paternal socio-economic status (SES) and family type at birth were examined as predictors of mortality. RESULTS Mortality was higher in SSD (hazard ratio (HR) 3.60; 95% confidence interval (CI) 2.38-5.45) and NSSD (4.05; 2.65-6.17) compared to persons without psychoses after adjustment for gender. HR for natural death was 2.01 (0.82-4.91) in SSD and 4.63 (2.43-8.80) in NSSD after adjustment for gender. Corresponding figures for unnatural deaths were 4.71 (2.94-7.54) and 2.94 (1.56-5.55), respectively. Among non-psychotic persons, mother's depression, smoking and low SES predicted mortality after adjustment for gender and parental psychoses (and SES), whereas among psychosis those whose father was a farmer had lower risk of mortality compared to those with high SES. CONCLUSIONS Individuals with SSD had a higher risk of unnatural death and individuals with NSSD of natural and unnatural deaths. Perinatal factors seem to be more important predictors of mortality in individuals without psychoses than with psychoses. According to population-based long follow-up data, it is important to pay attention to somatic morbidity behind natural causes of death in psychoses and to prevent suicides in order to prevent excess mortality.