Clinical studies of biomarkers in suicide prediction

Abstract

Suicide is a major clinical problem in psychiatry and suicidal behaviours can be seen as a nosological entity per se. Predicting suicide is difficult due to its low base-rate and the limited specificity of clinical predictors. Prospective biological studies suggest that dysfunctions in the hypothalamo pituitary adrenal (HPA) axis and the serotonergic system have predictive power for suicide in mood disorders. Suicide attempt is the most robust clinical predictor making suicide attempters a clinical high-risk group. A prediction model that incorporates biological testing to increase specificity and sensitivity of prediction of suicide risk is of potential clinical value.

The aim of these studies was to investigate the predictive potential of two biomarkers: the nonsuppression in the dexamethasone suppression test (DST) and low 5-hydroxyindoleacetic acid (5-HIAA) in the cerebrospinal fluid(CSF) in suicide prediction in clinical high-risk groups and to investigate relationships of the two neurobiologic correlates and some psychological components of the vulnerability to suicide.

DST and CSF monoamine metabolite data from two cohorts of suicide attempters and mood disorder inpatients (nearly 400 patients) were analysed in relation to subsequent death by suicide and to suicide intent and hopelessness. Optimal threshold of the DST in suicide prediction was analysed by Receiver Operating Characteristic (ROC). Interrelationship of two potential biomarkers was analysed in a subgroup of patients stratified by suicide attempt as a clinical predictor.

Suicide mortality rates were 9.4% for unselected mood disorder inpatients and 22% for those hospitalised after a suicide attempt. DST nonsuppression was found to be a biologic predictor of suicide in mood disorder inpatients with index suicide attempt yielding a risk ratio of 2.8; the optimal threshold for DST nonsuppressor status in suicide prediction was different for males and females. CSF 5-HIAA predicted suicide in short term while DST non-suppression seemed to be a long-term predictor of suicide risk for male suicide attempters. The interrelationship of two biomarkers was different in suicide attempters compared to mood disorder inpatients without suicide attempt and in male suicide victims compared to survivors indicating that the two biomarkers can be seen as independent biologic risk factors for suicide. Suicide intent measured by Beck Suicide Intent Scale is correlated to CSF HVA/5-HIAA ratio but failed to predict suicide in the clinical high-risk group.

DST non-suppression and low CSF 5-HIAA were independent biomarkers of suicide risk in suicide attempters. Support was lent to reintroduction of the DST as a complementary measurement of biological vulnerability in the clinical high-risk group of hospitalised male suicide attempters with mood disorder.