Suicide : a pharmacoepidemiological database study in the region Friuli Venezia Giulia, Italy

Abstract: Aims: The objective of this thesis is to examine the relationship between suicide and health care in Italy’s Friuli Venezia Giulia (FVG) region. First, it explores the correlation between suicide rates and antidepressant sales. Second, it analyses, at the individual level, the risk of suicide associated with the main suicidal risk factors, such as non-fatal self-harm, psychiatric disorders and somatic disorders. Third, it investigates the differences in suicide risk related to qualitative parameters in the use of antidepressants, such as adherence and treatment modifications. The overall aim is to help improve interventions to prevent suicide. Methods: All data were retrieved from the FVG Regional Social and Health Information System (SISSR), which links data using a unique anonymous key from different regional databases. Paper I analyses changes in individual-based data on antidepressant use and the rates of suicide during years 1997-2006. The other three papers are designed as case-control studies. All suicides that occurred in the region during years 2002-2008 (Paper II) and 2003-2013 (Paper III) were classified as cases, which were then age- and gender- matched to controls from the general population. In Paper IV, cases and controls from 2005 to 2014 must have had at least one prescription of antidepressant in the 730 days prior to the index date. Regression analysis was used to assess the association between suicide risk and its predictors. Results: In Paper I, suicide rates decreased by one-third in all genders and age groups. In parallel, both the number of individual users of antidepressant and the number of Defined Daily Doses per patient increased by 5-fold and 7-fold, respectively. In Paper II, the risk of suicide was highly increased by previous self-harm (OR = 53.1 for a single episode and OR = 98. for repeated episodes), as by psychiatric disorders (OR = 19.5). In Paper III, somatic disorders were strong predictors of suicide (OR = 2.9), particularly in case of comorbid disorders (OR from 2.6 to 9.8 when the number of disorders raised from 1 to ≥4) and in the elderly (OR = 4.3). No significant risk of suicide was found when medically-ill patients adhere to antidepressants. In Paper IV, none of the antidepressants compounds and classes was associated to suicide except SSRI (OR = 1.6). The association to suicide tended to decrease with adherence or current use of antidepressants. In all studies, on average only 10-20% of suicide cases adhere to antidepressants and 20-40% were currently using them at the time of death. Conclusions: Our findings support the hypothesis that treatment with antidepressant medication lowers the risk of suicide. The treatment at a proper dosage and for a proper length of time further decreased the risk, particularly in somatic-ill persons. Well-known risk factors, such as non-fatal self-harm and psychiatric and somatic disorders, were confirmed to highly increase the risk of suicide. Only a minority of suicides, however, had adhered to antidepressants or were under antidepressant treatment at the time of death.