Our results are consistent with studies that have identified CPA as a distal risk factor for a variety of negative outcomes. In a sample of criminal offenders, we found that CPA was associated with both lifetime frequency of aggression and lifetime history of a suicide attempt. Our finding of an association between aggression and lifetime suicide attempts is also consistent with prior studies (Brezo et al., 2006
), but constitutes, to our knowledge, the first demonstration of this relationship among criminal offenders. Notably, we found that lifetime aggression is a mediator of the relationship between CPA and lifetime suicide attempts. With the finding that frequency of aggression, in part, accounts for this relationship, we have begun to clarify one of the mechanisms by which CPA can influence suicidal behavior in criminal offenders.
The results of our mediation analysis are consistent with a model whereby the experience of CPA leads to a vulnerability to aggression that confers risk for a suicide attempt. These findings fit with prior neurobiological research. There is evidence that a positive pattern of early experiences is fundamental for healthy brain development (Perry and Pollard, 1998
). Conversely, animal and human research suggests that early traumatic experiences and parent-child negativity may compound genetic vulnerability to impaired serotonergic functioning (Matsumoto et al., 2005
; Brent and Mann, 2006
; Macri et al., 2007
; Crowell et al., 2008
). Attenuated serotonergic activity may underlie impulsivity (Braquehais et al., 2010
) which, in turn, leads to a greater propensity toward acting on aggressive impulses (Mann, 2003
) that are directed toward the self or others. Our findings provide additional evidence that a tendency toward aggression is one key link between CPA and suicide attempts.
Among abused individuals, the presence or absence of certain psychosocial variables (e.g., hopelessness, family cohesion) may determine whether underlying aggressive impulses are directed toward the self, in the form of suicidal behavior, or toward others (Plutchik et al., 1989
). Yet, other-directed aggression itself has serious consequences at the social level that may also contribute to suicidal behavior. Aggressive children and adolescents may experience higher levels of peer rejection than their non-aggressive peers (Kerestes and Milanovic, 2006
), potentially leading to increased levels of psychopathology (Masten, 2005
). Similarly, in adulthood aggression may contribute to continued social difficulties and may disrupt romantic relationships, leading to an increased likelihood of suicidal behavior (Yen et al., 2005
). Moreover, aggression may result in legal difficulties, adding to overall levels of stress and creating additional risk (Yen et al., 2005
). Future studies of CPA, aggression, and suicidal behavior would benefit from the incorporation of variables related to stressful life events, social functioning, and psychopathology in order to more comprehensively map the network of relationships that contribute to the CPA-suicidal behavior relationship.
Our results underscore the importance of routinely assessing suicide risk among individuals who exhibit aggression. Moreover, our findings raise the possibility that successful treatment of aggressive behavior among offenders with histories of CPA may also have a positive impact on suicide attempt risk. Pharmacological agents used to treat aggression and impulsivity (e.g., selective serotonin reuptake inhibitors, mood stabilizers) also impact suicide risk (Siever, 2008
). Moreover, adaptations of Dialectical Behavior Therapy (Linehan, 1993
) have been used to treat impulsivity, suicidal behavior, and aggressive behavior in forensic settings with some success (Berzins & Trestman, 2004
; Evershed et al., 2006
). More research is necessary in this area in order to better understand whether decreasing aggression among offenders reduces suicide risk, and to elaborate the mechanisms by which this might occur. One possibility is that reductions in impulsivity that result from treatment reduce the propensity to act on aggressive impulses toward the self or others.
While our findings indicate that aggression is one mechanism by which CPA influences suicide attempt risk, a next step is to clarify the type of aggression that mediates this relationship. The above discussion applies primarily to impulsive aggression, which occurs in response to a perceived threat or provocation and involves affective arousal and rash responding (Berkowitz, 1993
). However, a significant amount of research has been conducted on proactive aggression, which is premeditated and is used as a means to obtain a subsidiary goal (Berkowitz, 1993
). Childhood abuse is more highly associated with impulsive aggression than proactive aggression (Connor et al., 2004
; Murray-Close, 2010
), and most research on the link between aggression and suicidal behavior has focused on impulsive aggression (Turecki, 2005
). However, a limitation of our study is that our measure of lifetime aggression did not distinguish between impulsive and proactive aggression. Whereas impulsive aggression is the modal form of aggression in both non-forensic and forensic populations (Stanford et al., 2003
; Kockler et al., 2006
), offender samples may contain relatively high proportions of proactively aggressive individuals (Kockler et al., 2006
; Swogger et al., 2010
). Thus, we cannot rule out that tendencies toward proactive aggression played a role in our findings. This possibility is underscored by a recent study of substance dependent individuals that found a relationship between proactive aggression and suicide attempts (Conner et al., 2009
), indicating that the link between aggression and suicide attempts may not be limited to impulsive aggression. Future research examining the link between CPA and suicide attempts should take into account heterogeneity in aggression. Such a distinction has the potential to further increase the specificity of suicide risk assessments and further elucidate mechanisms that link CPA to suicidal behavior.
Several additional limitations of our study are noteworthy. First, CPA was assessed using a single question, potentially limiting reliability. Notably, however, the use of a small number of questions rather than a large protocol has been recommended to maximize specificity in childhood maltreatment research (Fang and Corso, 2008
), and single-item screens similar to that used in the present study have demonstrated adequate sensitivity (Thombs et al., 2006
). Second, the retrospective assessment of childhood abuse is not optimal given the potential for underreporting due to memory or social desirability biases (Fang and Corso, 2008
). Although there is evidence that CPA can be assessed retrospectively by self-report with reasonable accuracy (Widom and Shepard, 1996
), longitudinal studies would enable greater confidence with regard to the accuracy of data on CPA. Moreover, the incorporation of objective measures of CPA, along with self-report measures, in future studies would further enhance coverage of the construct. Finally, whereas we examined differences between suicide attempters and non-attempters, additional information might be gleaned from the use of frequency of suicide attempts as an outcome variable in future studies in order to further delineate those most at risk for suicide attempts.
In summary, we found that lifetime aggression is a mediator of the relationship between CPA and suicide attempts in a sample of criminal offenders. This suggests that the assessment of trait aggression, and intervention to reduce aggression among offenders with histories of CPA, may be important in the prevention of suicide attempts. Future studies using offender and non-offender samples will help to clarify the robustness of our findings and their generalizability to other high-risk populations.