Childhood trauma and psychosis: new perspectives on aetiology and treatment

Over the last 10 years, it has become increasingly apparent from well-conducted research studies that childhood trauma is a risk factor for psychosis. A recent meta-analysis of the role of childhood adversity on the risk of developing psychosis found that the odds of developing psychosis were significantly increased as a result of sexual abuse (odds ratio = 2.38), physical abuse (odds ratio = 2.95), emotional abuse (odds ratio = 3.40), bullying (odds ratio = 2.39), neglect (odds ratio = 2.90), but not parental death. Furthermore, in the studies that had controlled for potential confounders such as cannabis use, genetic liability (i.e. family history of psychosis), co-morbid psychopathology, ethnicity and education level, the effect of childhood trauma remained significant. This suggests that the effect of childhood trauma on psychosis is at least partially independent of these factors. A review of 46 studies found that between 28% and 73% of people with psychosis have experienced trauma such as sexual, physical or emotional abuse in childhood. A study by Wang et al., published in this issue of Early Intervention in Psychiatry, shows that in a large Chinese sample both people with first-episode psychosis (FEP) (n = 128) and chronic psychosis (n = 571) had an equally high prevalence of childhood trauma, 71% and 69%, respectively. However, there has been conjecture as to the strength of the relationship between childhood trauma and psychosis. Most of the evidence for the relationship is based on retrospective recall of childhood trauma by adults or older children. It has been recently proposed that this method may inflate the association between trauma and psychosis as those with psychosis may be biased towards overreporting childhood trauma in search of an explanation for their symptoms. However, there is evidence to suggest that this is unlikely to be the case. Firstly, in the aforementioned meta-analysis the odds of developing psychosis in groups with trauma versus no trauma are similar whether childhood trauma was reported before or after the onset of psychosis (2.75; 2.72 respectively). Further, one large prospective study in the meta-analysis, that utilized documented evidence of childhood sexual abuse, found that the odds of developing psychosis or schizophrenia in a group with documented childhood sexual abuse versus a matched (for sex and age) community control group without documented childhood sexual abuse were 2.1 and 2.6 respectively. By comparison, the meta-analysis of all studies investigating the association between childhood sexual abuse and development of psychosis (based largely on studies of retrospective recall of childhood sexual abuse) yielded an odds ratio of 2.38. The consistency of the odds ratios across different types of study points to no greater ‘over-reporting’ in groups with established psychotic symptoms or disorder than with groups where symptoms have not yet developed or with objective evidence of trauma. In fact, Bentall et al. have recently argued that there is now consistent evidence of a relationship between childhood trauma and psychosis, a dose–response effect, a temporal sequence of events, and emerging evidence of specificity of effects and the existence of plausible mechanisms. The presence of these factors has been suggested as a determinant of whether effects are causal from the epidemiological data. We suggest, based on the above and other data, that childhood trauma is likely to be a contributory cause in a set of disorders that include schizophrenia and other psychoses.