Psychological autopsy is one of the most valuable tools of research on completed suicide. The method involves collecting all available information on the deceased via structured interviews of family members, relatives or friends as well as attending health care personnel. In addition, information is collected from available health care and psychiatric records, other documents, and forensic examination. Thus a psychological autopsy synthesizes the information from multiple informants and records. The early generation of psychological autopsies established that more than 90% of completed suicides have suffered from usually co-morbid mental disorders, most of them mood disorders and/or substance use disorders. Furthermore, they revealed the remarkable undertreatment of these mental disorders, often despite contact with psychiatric or other health care services. More recent psychological autopsy studies have mostly used case-control designs, thus having been better able to estimate the role of various risk factors for suicide. The future psychological autopsy studies may be more focused on interactions between risk factors or risk factor domains, focused on some specific suicide populations of major interest for suicide prevention, or combined psychological autopsy methodology with biological measurements.

The relationship between suicide attempts and different vulnerability factors was investigated in 61 male subjects during forensic psychiatric examination. Personality traits and psychopathy were determined by means of the Karolinska Scales of Personality (KSP) and the Psychopathy Checklist-Revised (PCL-R). Clinical diagnoses were determined by the use of Structured Clinical Interview for DSM-III-R (SCID). Platelet monoamine oxidase (MAO) activity was investigated in 58 subjects.

Suicide attempters had extremely low socialization and high impulsive aggression. This was independent of a history of repeatedly criminal violence or not. They had higher PCL-R scores and lower levels of platelet MAO activity than non-attempters. No relationship was found between suicide attempts and repeated violent criminality. However, among the subjects with repeated criminal violence there was a highly significant difference between suicide attempters and non-attempters, indicating different personality profiles in violent offenders with and without suicidal behavior. Suicidal behavior was significantly associated with borderline personality disorder, but not with any other single DSM-III-R disorder. The results show a far more severe personality disturbance in suicide attempters than in other violent patients in this type of population. Suicidal behavior should therefore be evaluated and treated per se.

The existence of mental disorders is almost constant in subjects who try to kill themselves. In addition, a majority of attempters have more than one diagnosis.

This is especially true if Axis II or Axis I subthreshold conditions are taken into account. The existence of a disorder largely explains the association between most socioeconomic variables (sex, marriage, education level) and suicidality. Depressive disorders are the major risk factor, a risk probably linked to a current episode just before the attempt. The association to depressive episodes of an anxiety disorder or the existence of impulsive traits (and/or cluster B personality disorder or drug abuse) increases the risk of acting out. Ideation and attempts show parallel onset curves peaking between the ages of 14 and 20 years, with the existence of a previous DSM III R diagnosis as a strong predictor. The number of associated disorders linearly increases the probability of attempting suicide and is the only significant predictor for lethality. A proper treatment of mental disorders could substantially lower suicidality.

Depression, the major cause of suicide, is prevalent but an under-detected, underdiagnosed and, under-treated illness and it is particularly true for depressed suicide victims. However, several studies consistently show that successful treatment of depression not only relieves depressive symptoms, but also decreases and makes suicidality vanish. If the rate of treated depressions in the population increases gradually, at a given point it will appear in the decline of the suicide rates. Although absolute evidence is lacking at present, recent reports from some European countries strongly suggest that increasing utilisation of antidepressants is one of the most important contributing factors in the decrease in suicide rates.

Alcohol-dependent populations have a high lifetime suicide rate (between 7 and 15%, relative risk = 7), and alcoholism is one of the two psychiatric disorders most frequently found in suicidal cases (between 15 and 25%). Biological factors that would detect patients at risk could thus be of value. Carbohydrate-deficient transferrin, monoamine oxydase B, soluble interleukin-2 receptor and cholesterol have been proposed as markers of suicidal risk in alcohol-dependent patients, although nonspecific and with low predictive value. On the other hand, there is large and convergent data stressing the importance of serotonin dysregulation as increasing the risk for aggressive behaviour toward the self, although it is not clear whether serotonin is involved through the altered behavior inhibition system, enhancement of anxiety and depression, or association with specific subtypes of alcohol-dependence, such as early-onset type II alcoholism. Considering the complex but significant impact of alcohol on serotonin metabolism and turnover, it is likely that serotonin mediates a large part of the proneness of ethanol to commit impulsive-aggressive behavior.

There is a tension between the need for scientifically valid trials of new psychotropic drugs and concern about conducting placebo-controlled trials, the trials psychopharmacologists consider the gold standard trial, when this requires that some patients be denied existing effective therapy. This paper will review the scientific principles supporting the need for placebo-controlled trials in depression and schizophrenia, and will provide preliminary data on failure rates of placebo-controlled trials for these disorders, as illustrations of the application of these principles. Next, the ethical issues pertinent to the conduct of placebo-controlled trials for these two serious psychiatric disorders will be reviewed. Preliminary data on suicides in placebo-controlled depression and schizophrenia trials will be presented to argue for the ethical acceptability of the conduct of placebo-controlled trials in these two conditions.

Several data are available about the implication of the dopaminergic system in the control of inward-directed aggression. Previously, we suggested an involvement of D2-dopaminergic function in the expression of suicidal behavior by demonstrating a smaller growth hormone (GH) response to apomorphine, a dopaminergic agonist, in depressed patients with a history of suicide attempts in comparison to non-attempters. In the present study, in order to test this hypothesis, GH responses to intravenous apomorphine were measured in non-depressed patients with a history of suicide attempts. The study was performed in 17 non-depressed male patients with a score less than 12 on the 17-item HAMD. The patients were subgrouped into suicide attempters (N = 7) and non-attempters (N = 10). Mean GH peak responses to apomorphine differed significantly between suicide attempters and non-attempters: (mean ± SD) for GH peak, 10.4 ± 8.2 ng/mL vs 27.3 ± 13.1 ng/ml, F = 9.0, P = 0.009. In conclusion, dopaminergic disturbances seem to play a role in the biology of inward-directed aggression in non-depressed patients.

The suicide attempts were assessed in 32 schizophrenic patients on whom the dexamethasone suppression test (DST) was done twice in the course of illness: in the years 1985–91 and 1996–97. In the 1985–91 period, both baseline and post-dexamethasone cortisol levels were significantly higher in the patients with previous suicide attempts and baseline cortisol was higher in the patients who were to make a future attempt. In 1996–97, DST non-suppression was shown in more than half of the patients with a history of suicide attempt and in none of those without such history: all cortisol levels were significantly higher in the patients with a history of suicide attempt. Although the mean intensity of depression was higher in the patients with a history of suicide attempt, no association between the intensity of depression and present or previous DST non-suppression status was found. It is suggested that the hyperactivity of the hypothalamic-pituitary-adrenal axis may constitute an element of diathesis for suicidal behavior in schizophrenic patients.