pression of cortisol after dexamethasone is interpreted as a consequence of hyperactive HPA axis. This “dexamethasone suppression test” (DST) has been evaluated with suicidal patients. The results have been mixed and subject of some controversy. A number of studies have indicated that abnormal DST results and changes in daily rhythms of stress hormone release correlate with recent suicide attempts independently of psychiatric diagnosis (Banki et al., 1984; Lopez-Ibor et al., 1985; Pfeffer et al., 1991; Targum et al., 1983). Other studies, however, have failed to demonstrate this relationship between DST non-suppression and suicide attempt (Brown et al., 1986). In contrast, strikingly more consistent results have been obtained for the association of an abnormal DST response with completed suicide. Several studies have suggested that non-suppression of cortisol in the DST is a good predictor of future suicide. Carroll et al. (1980) evaluated 250 patients with melancholy. Only about half of them were nonsuppressors but all of the 5 subsequent patients who completed suicide were in this group. Similarly, Coryell and Schlesser (1981) tested 205 patients with unipolar depression and found that 45.8 percent had abnormal DST results but all four suicides were nonsuppressors. Norman et al. (1990) compared 13 depressed inpatients who subsequently completed suicide with 25 attemptors of suicide and 28 non-attemptors from the same inpatient population. While the DST nonsuppression rate was similar for the latter two groups, it was significantly higher for those who competed suicide. A meta-analysis by Lester (1992) supported the conclusion that the DST nonsuppression was more prevalent among those who completed suicide. A more recent study by Coryell and Schlesser (2001) demonstrated dramatic predictive ability of the DST. Seventy-eight inpatients with major depressive disorder or schizoaffective disorder were under assessment between 1978 and 1981 and followed for up to 15 years. Of the 78 patients, 32 had abnormal DST results upon admission to the hospital. Of the 32 patients, 26.8 percent eventually completed suicide; in comparison, only 2.9 percent of those with normal DST responses completed suicide.
The mechanism by which the HPA axis influences suicidal behavior is not yet established. Various researchers investigating the pathophysiology of suicide have summarized findings that integrate HPA hyperfunction with disturbances in serotonin function (Lopez et al., 1997; Yehuda et al., 1988). As described below, serotonin function also appears associated with suicide. Evidence suggests a reciprocal relationship between the serotonergic system and the HPA axis. Activation of serotonergic pathways or administration of agents that increase the activation of serotonin receptors elicit increases in plasma cortisol (Calogero et al., 1990; Dinan, 1996; Fuller, 1990; Matheson et al., 1997a; 1997b; Meltzer et al., 1984; Owens et al., 1990). Conversely, serotonin receptors are inhibited by glucocorti-