The toxicological pharmacoepidemiology of suicide : population-based studies on psychotropic-medication use in Sweden

Abstract

The act of completed suicide is a complex phenomenon associated with numerous identified individual and combined risk factors. Mental illness and related disorders confer among the highest risks of completed suicide and have consequently been targeted through pharmacological interventions. However, psychotropic medications and their possible effects on completed suicide have proved notoriously difficult to study. Nonetheless, it is of great importance to further investigate these classes of medications and their role in the suicidal process. The thesis focuses on toxicological verification of psychotropic medication among individuals who have committed suicide or died of other causes in Sweden. We have investigated the potential role of selective serotonin re-uptake inhibitors in the choice of suicide method and demonstrated the level of adherence to and recreational use of the most commonly prescribed non-addictive psychotropic medications in the Swedish population. In yet-to-be-published Study III, we have examined purchasing patterns for psychotropic medications during the year prior to death, as well as the impact of the level of medication adherence upon the risk of completed suicide.

All studies are based on Swedish national registry data from a period of nearly ten years (2005-2014). Full coverage on dispensed prescriptions, results of medico-legal autopsies, causes of death and diagnoses from inpatient care have been used to operationalize investigated exposures and outcomes and facilitate adjustments for confounders in two case-control studies and one methodological study. By the use of the prescription-based algorithm PRE2DUP, purchases of psychotropic medications and calculated lengths of use were compared with forensic-toxicological findings as a measure of adherence. Logistic regression and Cohen’s kappa have been the principal statistical methods used. A new measure of continued need of treatment – the dispensation ratio (the ratio of initiated and discontinued prescriptions) – was also developed.

Two of the main findings are that selective serotonin re-uptake inhibitors are associated with violent completed suicide during early treatment among elderly subjects and that presence of other substances (including other medications and illegal drugs) is an important confounder of associations between treatment with selective serotonin re-uptake inhibitors and choice of suicide method. Overall, rates of adherence in the Swedish general population to psychotropic medication, as reflected by therapeutic blood concentrations at forensic autopsy, are good. Biochemically verified incomplete adherence to antipsychotics, as shown in Study III, is associated with a markedly increased risk of completed suicide. The corresponding effect size for antidepressants was shown to be smaller and no longer statistically significant after adjustments for the dispensation ratio. Forensic-toxicological data are largely conservative biochemically verified measures of the possible impact of psychotropic medication upon the risk of death. By combining such data with national registry data, we have revealed findings of great importance to further suicide prevention and, at the same time, demonstrated a public-health-oriented application of results from medico-legal autopsies, beyond their immediate use in forensic investigations.