Antipsychotics, mood stabilisers, and reductions in violence

www.thelancet.com Vol 384 September 27, 2014 1167 In the past 25 years, evidence has accumulated that people with schizophrenia are at increased risk to commit violent crimes and, to a lesser degree, non-violent crimes compared with the general population. A smaller amount of evidence suggests that patients with bipolar disorder are also at increased risk of committing violent off ences. Although both disorders are mainly treated with drugs that reduce risk of relapse, the eff ect of antipsychotic medications and mood stabilisers on violence has been unclear. In view of the human suff ering, stigma, and costs resulting from violence by people with these disorders, the identifi cation of humane and eff ective strategies to reduce such behaviours is an urgent unmet need. In their study in The Lancet, Seena Fazel and colleagues used Swedish national registers to investigate whether prescriptions for antipsychotic drugs and mood stabilisers were associated with reductions in violent crime. During 2006–09, antipsychotics or mood stabilisers were prescribed to 40 937 men and 41 710 women in Sweden, of whom 2657 (6·5%) men and 604 (1·4%) women were convicted of a violent crime during the study period. Within-individual analyses were used to compare convictions for violent crime during periods when drugs were prescribed compared with periods when no prescriptions were dispensed. Violent crime was reduced by 45% in patients receiving antipsychotics (hazard ratio [HR] 0·55, 95% CI 0·47–0·64) and by 24% in patients prescribed mood stabilisers (0·76, 0·62–0·93). Depot medications (HR adjusted for concomitant oral drugs 0·60, 95% CI 0·39–0·92) and higher doses of antipsychotics were also associated with reductions in violent crime (p=0·019). Similar reductions in the risk of violent crime were recorded in people with diagnoses of schizophrenia who were prescribed antipsychotic drugs, and in men, but not women, with diagnoses of bipolar disorder who were prescribed mood stabilisers. However, about 60% of the patients prescribed antipsychotics or mood stabilisers did not have diagnoses of schizophrenia, bipolar disorder, or other psychoses. Despite this fi nding, the reductions in violent crime during periods when medications were dispensed were detected in analyses that included all people prescribed these drugs. These results might represent the increased risk of violent behaviour in people with psychotic-like experiences who do not have psychotic disorders. This issue warrants further investigation. The study was undertaken carefully, and sensitivity analyses suggest that the results are robust. No age eff ects were detected in the association between antipsychotics and reductions in violent crime, while the association of mood stabilisers in older patients with bipolar disorder needs further investigation, as do diff erences in associations in male and female patients. Similar reductions (22–29%) during periods when antipsychotics or mood stabilisers were prescribed were noted for any crime, drug-related crimes, less severe crimes, and suspected violent crimes. Thus, this study provides a basis for future investigations aimed at establishing the type of patient, phase of illness, and type and dose of medication that reduces physically aggressive behaviour. However, many uncertainties remain regarding the treatment of aggression in schizophrenia and bipolar disorder. In their study, Fazel and colleagues used data from offi cial criminal records of violent crime. However, such criminal records capture only some physically aggressive behaviours, since family members and social and health-care workers—the usual victims— often do not report such behaviours. Methods that ask patients and family and staff to report such behaviours have been shown to be valid. Importantly, in people with schizophrenia the correlates of aggressive behaviour and violent crime are similar. Fazel and colleagues note that randomised controlled trials studying the eff ect of drug treatments on violence are rare because of low rates of violence. However, when physical aggression towards another is used as the dependent variable, this is not the case. For example, and consistent with previous evidence, two recent studies that followed patients for 1 year after a fi rst episode of psychosis reported that 38% and 14% of the patients engaged in aggressive behaviour. These rates are suffi ciently high to warrant examination of the characteristics of patients who hurt others, and of the eff ects of drug treatment on such behaviours. These studies identifi ed anger due to delusions and previous antisocial and aggressive behaviour as risk Antipsychotics, mood stabilisers, and reductions in violence