Demencia y suicidiouna revisión sistemática y meta-análisis

Abstract

Suicide is a problem of first order. At epidemiological level, moments of life with highest risk occur in adolescence and old age. On the other hand, rates of dementia increase with age of person. Two systematic reviews suggested a relationship between dementia and suicide. To date, meta-analysis analyzing this association has not been published yet. Aims of the study are: i) to quantify the meaning and magnitude of association between dementia and suicide; ii) to analyze association between dementia and different types of suicidal behavior: ideation, attempts, and completed suicides; iii) to examine the influence of moderating variables with the greatest impact on effect sizes; and iv) to develop a predictive model to allow the estimation of suicide risk in people with dementia. Methods: Present study consists of a Systematic Review and Meta-Analysis on the relationship between dementia and suicide (ideation, attempt or completed suicide). A systematic search was carried out in 5 databases (PubMed, EMBASE, CINAHL, IME and Lilacs ) looking for scientific articles on this relationship, published until December 31, 2018. A manual search and analysis of bibliography of selected articles was performed and the most relevant authors were contacted. Inclusion criteria were: studies on dementia and suicide; Analytical, Observational or Descriptive type of studies; existence of a comparison group; published in English, French, Italian, Portuguese or Spanish. The processes of article selection and data extraction using a previously developed protocol were carried out independently by two reviewers. The analyses were: i) effect size index using Odds Ratio (OR) and 95% confidence interval (using random effects model); ii) heterogeneity evaluation using forest plots, Cochran's Q and I2 index; iii) publication bias assessment using funnel plots ("trim-and-fill" method) and Egger's test; iv) analysis moderating variables using multiple mixed effects meta-regression model. Results: Articles (N = 2 199) were located from 5 databases, plus 93 articles from manual review of bibliographic references. Finally, after meeting inclusion-exclusion criteria, 37 studies were finally included in Meta-analysis with 47 basic study units were obtained. The effect size of dementia association was: with Suicidal Ideation OR= 1.37 (95% CI: 0.78-2.39); Suicide Attempt: OR= 2.24 (95% CI: 1.01-4.97); and Suicide Completed: OR= 1.28 (95% CI: 0.77-2.14). Greater probability of suicidal ideation at a younger age in subjects with dementia (bj = -0.046; F (1, 2) = 22.453; p = 0.040; k = 4; R2 = 0.99). Effect size was significantly associated with dementia (F (2, 16) = 4.842, p = 0.023, R2 = 0.35), community sample (OR+ = 5.379); suicide measurement (F (1, 17) = 4.480, p = 0.049, R2 = 0.20), measured by records (OR+ = 1.656). The meta-regression model explained 59% of variance and included: type of sample, measure of completed suicide and measure of effect size, statistical significance was reached (F (4, 14) = 6.348, p = 0.004),. Remarkable heterogeneity was observed between included studies and a potential publication bias could be ruled out. Conclusions: Dementia, especially at its early stages, can increase the risk of suicide events, particularly suicidal ideation (especially in those of younger age), as well as suicide attempts and completed suicide, compared to people without dementia. An adequate assessment of suicide risk in people with dementia would be advisable, especially at its initial stages.