Association of trauma, post-traumatic stress disorder and non-affective psychosis across the life course: a nationwide prospective cohort study

Background We aimed to examine the temporal relationships between traumatic events (TE), post-traumatic stress disorder (PTSD) and non-affective psychotic disorders (NAPD). Methods A prospective cohort study of 1 965 214 individuals born in Sweden between 1971 and 1990 examining the independent effects of interpersonal and non-interpersonal TE on incidence of PTSD and NAPD using data from linked register data (Psychiatry-Sweden). Mediation analyses tested the hypothesis that PTSD lies on a causal pathway between interpersonal trauma and NAPD. Results Increasing doses of interpersonal and non-interpersonal TE were independently associated with increased risk of NAPD [linear-trend incidence rate ratios (IRR)(adjusted) = 2.17 [95% confidence interval (CI) 2.02-2.33] and IRRadjusted = 1.27 (95% CI 1.23-1.31), respectively]. These attenuated to a relatively small degree in 5-year time-lagged models. A similar pattern of results was observed for PTSD [linear-trend IRRadjusted = 3.43 (95% CI 3.21-3.66) and IRRadjusted = 1.45 (95% CI 1.39-1.50)]. PTSD was associated with increased risk of NAPD [IRRadjusted = 8.06 (95% CI 7.23-8.99)], which was substantially attenuated in 5-year time-lagged analyses [IRRadjusted = 4.62 (95% CI 3.65-5.87)]. There was little evidence that PTSD diagnosis mediated the relationship between interpersonal TE and NAPD [IRRadjusted = 0.92 (percentile CI 0.80-1.07)]. Conclusion Despite the limitations to causal inference inherent in observational designs, the large effect-sizes observed between trauma, PTSD and NAPD in this study, consistent across sensitivity analyses, suggest that trauma may be a component cause of psychotic disorders. However, PTSD diagnosis might not be a good proxy for the likely complex psychological mechanisms mediating this association.