Adult antisocial syndrome with comorbid borderline pathology: Association with severe childhood conduct disorderRick Howard, PhD
University of Nottingham, Nottingham, UK, Peaks Academic Research Unit, Rampton Hospital, Retford, UKNick Huband, PhD
Division of Forensic Mental Health, Nottinghamshire Healthcare NHS Trust, Nottingham, UKConor Duggan, PhD, FRCPsych
University of Nottingham, Nottingham, UK, Division of Forensic Mental Health, Nottinghamshire Healthcare NHS Trust, Nottingham, UK
BACKGROUND: This study tested the hypothesis that adult antisocial syndrome co-concurrent with borderline personality disorder (AAS+BPD) would be associated with greater conduct disorder (CD) severity than AAS alone.
METHODS: Sixty-nine personality disordered individuals exhibited a sufficient number of adult antisocial traits to meet DSM-IV criterion A for antisocial personality disorder (AsPD). These were subdivided into those who did (AAS+BPD) or did not (AAS alone) meet DSM-IV criteria for a BPD diagnosis. We then compared the 2 groups on CD symptoms and historical, clinical, and self-report measures.
RESULTS: The mean number of CD criteria met and the total number of individual CD symptoms were significantly greater in the AAS+BPD group compared with the AAS alone group. The former also were more likely to be female, to have self-harmed, to show greater personality disorder comorbidity, and to self-report greater anger.
CONCLUSIONS: The functional link between CD and adult antisocial symptoms appears to be mediated, or at least moderated, by co-occurring borderline pathology.
KEYWORDS: adult antisocial syndrome; conduct disorder; borderline personality disorder
ANNALS OF CLINICAL PSYCHIATRY 2012;24(2):127-134
Antisocial personality disorder (AsPD) is unique among DSM-IV Axis II personality disorders in requiring a diagnosis of an Axis I disorder, namely conduct disorder (CD).1 Studies that have compared individuals who met the adult criteria for AsPD but not the CD criteria with those who met both sets of criteria have reported few, if any, differences between them in terms of sociodemographic, criminal history, Axis I, or current antisocial behavior variables, suggesting that CD makes little or no difference to a diagnosis of AsPD.2-4 However, 2 important caveats bear examination before this conclusion can be definitely drawn.
The first caveat concerns the relative severity of adult antisocial symptoms in adult antisocial syndrome (AAS) individuals with vs without antecedent CD. In Black and Braun’s3 study and the more recent study by Perdikouri et al,2 AAS individuals with a diagnosis of CD tended to show greater severity of adult antisocial symptoms compared with those without a CD diagnosis. Only 1 adult behavior problem (“conning others”) significantly differentiated the groups in Black and Braun’s3 study, whereas 2 of the adult AsPD items (“repeated acts that are grounds for arrest” and “irritability and aggressiveness”) significantly separated the groups in the study by Perdikouri et al.2 This suggests that subtle differences in adult antisocial psychopathology might not have been fully captured in the above studies. Consistent with this, a recent study by Walters and Knight5 challenged the conclusion that antecedent CD makes no difference to a diagnosis of AsPD. Male prison inmates who met both adult AsPD criteria and CD criteria, compared with inmates who met only the adult AsPD criteria, scored higher on measures of criminal thinking, antisocial attitudes, and misconduct in prison, indicating the existence of a clinically meaningful distinction between antisocial adults with and without antecedent childhood CD.
The second caveat concerns the issue of psychiatric comorbidity. Many personality disordered patients—particularly forensic patients—present with >1 personality disorder (PD) diagnosis.6 Walters and Knight did not consider this issue of PD comorbidity, so unfortunately it is not possible to know whether their AAS with CD group may have differed from their AAS alone group in terms of PD comorbidity. Perdikouri et al2 did consider PD comorbidity in their community sample, finding significantly greater comorbidity (≥3 PDs) in their AsPD group (AAS with CD) than in their AAS group. Inspection of their data relating to personality disorder diagnosis reveals a greater prevalence of antisocial/borderline comorbidity in the AsPD group (70%) compared with the AAS group (49%). Although this difference was not statistically significant, the tendency for AsPD/BPD comorbidity to be more prevalent when CD was present than when absent suggests the possibility of a functional link between CD and AsPD/BPD comorbidity, rather than—as assumed by DSM-IV1—between CD and AsPD per se.
Comorbidity between antisocial and borderline PD is of particular interest from a forensic point of view because it has been linked to risk of violence in both forensic samples7 and community samples.8,9 In the Howard et al study,9 patients with AsPD/BPD comorbidity showed high levels of anger and impulsivity, suggesting deficient emotional self-regulation as a mechanism through which comorbid antisocial and borderline PD is linked to violence.
The supposition that CD plays a role in the etiology of adult antisocial behavior is supported by Robins’ seminal study10 of antecedents of adult antisocial behavior. More recent studies have linked childhood-onset social and behavioral problems to pervasive violence in adulthood11 and to adult AsPD after controlling for child psychiatric comorbidity and other confounding factors.12 Similar to its adult counterpart, AsPD,13 CD is strongly and significantly associated with early-onset (age ≤18) substance abuse, which in turn is highly predictive of aggression and psychopathy in adulthood.14
In short, CD (and the early-onset substance abuse associated with it) cannot be ignored as a key antecedent of severe adult antisocial behavior. The question is: what is the relevant pattern of PD comorbidity that is linked to severe CD? Although the assumption underlying the DSM-IV criteria for AsPD is that adult antisociality is functionally linked to CD, considerations outlined above suggest that AAS with comorbid BPD might be more clearly linked to CD.
The present study therefore tested the hypothesis that AAS, when comorbid with BPD, would be associated with greater CD severity and higher anger than when this comorbidity was absent. We re-visited the sample examined by Perdikouri et al2 to test our hypothesis. We predicted that in that community sample of individuals with confirmed PD, those having AAS with comorbid borderline PD would differ (score higher) on CD symptom severity than those with AAS alone (but with whatever other PD comorbidity), but that the groups would not differ (or differ only marginally) on the adult criteria for AsPD. Collection of childhood CD data relied—necessarily, given the nature of this (adult) sample—on retrospective self-report and therefore was likely to include both false-positive and false-negative errors.15 This caveat should be borne in mind when interpreting the results presented below.
The sample we examined was composed of volunteers in a clinical trial designed to explore the efficacy of a time-limited, group-based intervention for individuals with personality disorder. Details of the trial are recorded elsewhere16 but are summarized here. All participants resided in the East Midlands region of England and were recruited as follows. Local services were asked to identify persons who might wish to volunteer to participate and were thought likely to meet the trial’s inclusion criteria (at least 1 personality disorder diagnosed using DSM-IV; a level of literacy and cognitive functioning sufficient to allow engagement in the type of intervention being investigated; age 18 to 65; willing to be assigned to treatment or to waiting-list control). Exclusion criteria were the presence of functional psychosis or organic disability sufficient to impair understanding of the purpose of the study, engagement in the assessment interview, or completion of psychometric instruments.
Each of the 255 volunteers who attended for assessment was examined using the interview version of the WHO17 International Personality Disorder Examination (IPDE). Each examination was carried out by 1 of the trial’s 6 assessors, all of whom were experienced with working with individuals with personality disorder and trained in executing this instrument. Inter-rater reliability was checked by 1 of the authors observing a small number of assessments at random and independently scoring volunteers’ responses to each question.
Of the 241 volunteers who fulfilled the DSM-IV criteria for at least 1 personality disorder, 69 met ≥3 adult criteria for AsPD. We divided these individuals into 2 groups: those who met adult criteria for AsPD as well as criteria for a BPD diagnosis (AAS+BPD), and those who met adult criteria for AsPD but did not meet the borderline PD criteria (AAS alone). The 2 groups were then compared, using information acquired during the assessment interview and historical data obtained by reviewing each participant’s case notes and medical records. Although their inclusion was not critical to the central hypothesis, we also examined data for a third group of individuals who met criteria for BPD only. The study received approval from the relevant medical research ethical committee, and all volunteers provided written consent for their information to be accessed.
Comparisons also were made on forensic data derived from the Offenders’ Index ,18 which is maintained by the English Home Office (Research, Development and Statistics Department) to provide an official record of all individuals convicted of a standard list offense in England and Wales since 1963. Standard list offenses include all indictable offenses and a few non-indictable offenses; they are generally the more serious types of offenses.
Additional comparisons were made using scores from psychometric measures available from volunteers who became participants in the trial. These measures were completed after assessment but before commencing any problem-solving therapy and comprised the Social Functioning Questionnaire19; the Social Problem-Solving Inventory-Revised,20 the State-Trait Anger Expression Inventory-2,21 the Barratt Impulsiveness Scale,22 the Dissociative Experiences Scale23; and the Experience of Shame Scale.24
Statistical analysis was carried out using Statistical Package for the Social Sciences software for Windows (version 15.0). All results were taken as significant at the level of P < .05, employing 2-tailed tests unless specified otherwise. Categorical comparisons were made using χ2 tests with the Yates’ correction. Where appropriate, specific dimensional comparisons were made using independent sample t tests after applying Levene’s test of homogeneity of variance.
Two groups were selected for comparison: 40 individuals who met criteria for AAS and BPD, and 29 who met criteria for AAS alone. Although the main statistical comparison was between these 2 groups, data for a third group of 34 individuals who met criteria for BPD only are presented in the tables below.
Social and demographic data
The 2 groups were not significantly different in terms of age, marital status, or occupation (TABLE 1). However, the AAS+BPD group contained a significantly smaller proportion of men than women (47.5% vs 89.6%; P=.0003).
Social and demographic data for 2 groups (AAS+BPD and AAS alone), plus comparators with BPD alone
|Sex, n (%)
|Age, mean yr (SD)
|Marital status n (%)
|Divorced, widowed, separated
|Living with partner
|Occupation n (%)
Reliability of IPDE diagnosis
Interviewer-observer agreement was derived from 16 double-rated interviews using the IPDE instrument. There was no disagreement in the assignment of research diagnosis of personality disorder in any of the cases jointly assessed in this way. Inter-rater reliability at the item level was calculated on the basis of 3×3 contingency tables; Cohen’s kappa coefficient ranged from 0.69 to 0.88 (mean 0.83; SD=0.05).
We found evidence of greater PD comorbidity in the AAS+BPD group. A significantly greater proportion of individuals in the AAS+BPD group met the diagnostic criteria for ≥3 PDs (40.0% vs 10.3%; P=.007) and for having PD in 2 clusters (40.0% vs 6.9%; P=.005). A minimum of 3 out of 15 CD items plus another 3 out of 7 adult items scored “definitely present” are required to meet DSM-IV criteria for an AsPD diagnosis.1 On this basis, 21/40 (52.5%) of the AAS+BPD group met full criteria for AsPD, compared with 9/29 (31%) of those with AAS alone.
Comparison of CD symptoms between groups
The mean number of CD criteria met was significantly higher in the AAS+BPD group (2.93; SD=2.92) than in the AAS group (1.45; SD=1.74; t=2.61; P=.011). The mean CD dimensional score based on aggregating the individual CD item scores was also significantly higher in the AAS+BPD group (7.93; SD=5.82) than in the AAS group (4.76; SD=4.65; t=2.42; P=.018).
Comparison of adult antisocial symptoms
The mean number of adult criteria met did not differ significantly between the AAS+BPD group (4.18; SD=0.98) and the AAS group (3.72; SD=0.62; t=1.90; NS). The mean adult antisocial dimensional score based on aggregating the individual adult symptom scores likewise did not significantly distinguish the groups: AAS+BPD (9.57; SD=1.74), AAS (8.76; SD=1.90; t=1.85; NS).
Self-reported psychometric data
Self-reported psychometric data are shown in TABLE 2. The AAS+BPD group scored significantly higher than the AAS group on trait anger (P=.011), outwardly directed anger expression (P=.001), and anger expression index (P=.026), although no significant differences were detected on anger control (inwards or outwards) or on inwardly directed anger expression.
Self-reported psychometric data for 2 groups (AAS+BPD and AAS alone), plus comparators with BPD alone
mean score (SD)
(AAS+BPD vs AAS alone)
mean score (SD)
|Social Functioning Questionnaire
|Social Problem-Solving Inventory, Revised
|Barratt Impulsiveness Scale
|Experience of Shame Scale
|Dissociative Experiences Scale
|State-Trait Anger Expression Inventory-2a
| Anger expression out
| Anger expression in
| Anger control out
| Anger control in
| Anger expression index
Comparison on other characteristics
The 2 groups also were compared on a range of other variables of interest (TABLE 3) and showed a statistically significant difference on 2 measures. The AAS+BPD group had a greater proportion of individuals with at least 1 episode of self-harm (90.6% vs 59.3%; P=.012) and a higher mean number of custodial sentences (0.6 vs 0.1; P=.019).
Other characteristics of 2 groups (AAS+BPD and AAS alone), plus comparators with BPD alone
|Notes record information suggesting, n (%)
|Address changes in past 5 years, mean (SD)
|Notes record psychiatric hospitalization, n (%)
|Compulsory admission, ever
|Use of servicesc in previous 6 months n (%)
|Accident and Emergency attendance, any reason
|Psychiatric hospital admission, any
|Contact with community mental health staff, any
|Extent of service usec in previous 6 months, mean (SD)
|Accident and Emergency attendances for any reason
|Contacts with mental health staff
|Notes record violence, n (%)
|To partner/family members
|Criminal conviction, ever, n (%)
|Total convictions,d mean (SD)
|Custodial sentence, ever, n (%)
|Number custodial sentences,d mean (SD)
The key hypothesis—that adult antisocial behavior comorbid with borderline PD would be associated with greater severity of CD symptoms—received support both in terms of the greater number of CD criteria met by the AAS+BPD group and their greater overall CD severity, relative to the AAS only group. Despite this, adult antisocial symptom severity was comparable in the 2 groups. This dissociation between child and adult antisocial symptom severity implies that CD is more closely linked to AAS when this co-occurs with borderline pathology than when it occurs without such pathology. This has implications for the current (DSM-5) revision of the criteria for antisocial personality disorder. Implicit in the DSM-IV criteria1 for AsPD was the notion that disordered conduct in childhood is functionally linked to antisocial behavior in adulthood. The current results suggest that this relationship is at least moderated by the presence (vs absence) of co-occurring borderline pathology.
Our results further indicated considerable overlap between a diagnosis of AsPD and comorbid borderline pathology. This overlap has been implicitly recognized in the trait assessment proposed for the forthcoming DSM-5; specifically, that hostility and aggression—within the trait domain of “antagonism”—and impulsivity—within the domain of “disinhibition”—should be considered traits that are common to both the proposed “antisocial/psychopathy” and “borderline” categories. Although it is reasonable to suppose that it may be this commonality that is functionally linked to CD, this remains to be verified.
The finding that severity of childhood CD and AAS symptom severity can be dissociated by the presence vs absence of co-occurring borderline pathology is consistent with the idea that different developmental trajectories lead to a common endpoint of adult antisocial behavior, only 1 of which is mediated by conduct disorder. It has been suggested25 that the CD-mediated trajectory is likely linked to early-onset substance abuse, and this has been supported by Gustavson and colleagues’ finding that CD is associated closely with adolescent drug and alcohol abuse.14 These authors found early-onset substance abuse to be associated with both AsPD and BPD and to predict aggression and psychopathy in adulthood. It therefore seems likely that CD and heavy substance use in early/mid-adolescence have mutually potentiating effects on the development of adult antisocial behavior associated with AsPD/BPD comorbidity, but further research will be required to confirm this.
Individuals who showed adult antisocial syndrome comorbid with borderline PD, when compared with those showing AAS only, were more likely to be female, to have self-harmed, to show greater PD comorbidity, and to self-report greater anger. However, it is apparent from TABLE 1 that although men and women were roughly equally represented in the AAS with comorbid BPD group, men were overrepresented in the AAS-only group and women predominated in the BPD-only group. The finding of higher anger scores in AAS with comorbid BPD replicates previous results9 and suggests that poor emotional self-regulation is characteristic of this pattern of PD comorbidity. However, it is apparent from TABLE 2 and TABLE 3 that anger and self-harm scores were comparably high in both the AAS/BPD comorbid and BPD-only groups, suggesting that the deficient emotional self-regulation seen in the AAS/BPD comorbid group is attributable to their borderline pathology. In contrast, it is interesting to note that, despite their high anger scores, individuals in the BPD-only group were not severely antisocial: they scored low both in terms of the number of CD criteria met (mean 0.76) and the number of adult antisocial criteria met (mean 1.59). This suggests that the high score obtained by the AAS/BPD comorbid group for the number of childhood CD criteria met (mean 2.93) reflects something unique about this pattern of comorbidity.
Overall the criminal history of the comorbid group did not appear remarkably different from that of AAS, although their having received a greater number of custodial sentences indicated that they may have been more criminally inclined. There was no evidence that violence was more common in the AAS/BPD comorbid group. However, the assessment of violence was rudimentary, being based on case-note information and limited to domestic violence and violence to property. Acts of violence—other than in a domestic context—were relatively uncommon in this non-forensic sample. Further investigation of the relationship between AAS/BPD comorbidity and violence—both in a larger community sample and in forensic samples—is warranted, given the over-representation of individuals with AsPD/BPD comorbidity in forensic and particularly high-secure settings.7
Three caveats are in order when interpreting these results. First, this relatively small sample limits the generalizability of our findings, which need to be replicated in a larger sample. Second, this cross-sectional study with a retrospective assessment of CD symptoms relied on interviewees being truthful in their responses and accurate in their recollections. As previously noted, self-report can result in false-positive and false-negative errors, particularly for recalled childhood behaviors.15 Finally, it might be argued that because the AAS+BPD group showed greater PD comorbidity generally than the AAS group—they were, for example, significantly more likely to have ≥3 PDs—it is possible that severity of CD symptoms is associated simply with the extent of PD comorbidity generally, rather than with AAS+BPD specifically. This argument, however, lacks plausibility when one considers that co-occurring antisocial and borderline PD is itself associated with greater PD comorbidity. For example, in a study by Newhill and colleagues,26 a subgroup of community resident borderline PD patients was identified who showed psychopathic/antisocial features and who, therefore, closely resembled our AAS+BPD group. Interestingly, and in line with the current results, the psychopathic/antisocial BPD subgroup showed a history of childhood-onset CD, which typically is associated with greater severity of CD symptoms. This subgroup also showed a high co-occurrence of paranoid, narcissistic, and sadistic PDs, suggesting that antisocial/borderline comorbidity may represent a particularly severe and “toxic” constellation of PD traits (paranoid and narcissistic as well as hostile, aggressive, and impulsive). A constellation of traits such as this is typically found in personality-disordered, high-secure forensic patients (“secondary psychopaths”) characterized by the combination of high “hostile impulsivity” and high “anxiety/social withdrawal.”27
The functional link between CD and AAS appears to be mediated, or at least moderated, by co-occurring borderline pathology. Larger-scale studies will be required to confirm this.
DISCLOSURE: The authors report no financial relationship with any company whose products are mentioned in this article or with manufacturers of competing products.
ACKNOWLEDGEMENTS: The UK National Programme for Forensic Mental Health Research and Development and the English Home Office are gratefully acknowledged for jointly funding this work (research grant reference number MRD 12/28).
- Diagnostic and statistical manual of mental disorders. 4th ed text rev. Washington DC: American Psychiatric Association; 2000.
- Perdikouri M, Rathbone G, Huband N, et al. A comparison of adults with antisocial personality traits with and without childhood conduct disorder. Ann Clin Psychiatry. 2007;19:17–23.
- Black DW, Braun D. Antisocial patients: a comparison of those with and those without childhood conduct disorder. Ann Clin Psychiatry. 1998;10:53–57.
- Marmorstein NR. Adult antisocial behavior without conduct disorder: demographic characteristics and risk for cooccurring psychopathology. Can J Psychiatry. 2006;51:226–233.
- Walters GD, Knight RA. Antisocial personality disorder with and without antecedent conduct disorder: does it make a difference? J Pers Disord. 2010;24:258–271.
- Zimmerman M, Rothschild L, Chelminski I. The prevalence of DSM-IV personality disorders in psychiatric outpatients. Am J Psychiatry. 2005;162:1911–1918.
- Duggan C, Howard RC. The ‘functional link’ between personality disorder and violence: a critical appraisal. In: McMurran M Howard RC, eds. Personality, personality disorder and violence. Chichester, United Kingdom: Wiley-Blackwell; 2009:19–37.
- Newhill CE, Eack SM, Mulvey EP. Violent behavior in borderline personality. J Pers Disord. 2009;23:541–554.
- Howard RC, Huband N, Duggan C, et al. Exploring the link between personality disorder and criminality in a community sample. J Pers Disord. 2008;22:589–603.
- Robins LN. Deviant children grown up. Baltimore MD: Williams and Wilkins; 1966.
- Soderstrom H, Sjodin AK, Carlstedt A, et al. Adult psychopathic personality with childhood-onset hyperactivity and conduct disorder: a central problem constellation in forensic psychiatry. Psychiatry Res. 2004;121:271–280.
- Fergusson DM, Boden JM, Horwood J. Classification of behavior disorders in adolescence: scaling methods predictive validity, and gender differences. J Abnorm Psychol. 2010;119:699–712.
- Bakken K, Landheim AS, Vaglum P. Early and late onset groups of substance misusers: differences in primary and secondary psychiatric disorders. Journal of Substance Use. 2004;9:224–235.
- Gustavson C, Ståhlberg O, Sjödin, AK, et al. Age at onset of substance abuse: a crucial covariate of psychopathic traits and aggression in adult offenders. Psychiatry Res. 2007;153:195–198.
- Rueter MA, Chao W, Conger RD. The effect of systematic variation in retrospective conduct disorder reports on antisocial personality disorder diagnoses. J Consult Clin Psychol. 2000;68:307–312.
- Huband N, McMurran M, Evans C, et al. Social problem solving plus psychoeducation for adults with personality disorder. Br J Psychiatry. 2007;190:307–313.
- Loranger AW. International Personality Disorder Examination. Geneva Switzerland: World Health Organization; 1995.
- Friendship C, Thornton D, Erikson M, et al. Reconviction: a critique and comparison of two main data sources in England and Wales. Legal and Criminal Psychology. 2001;6:121–129.
- Tyrer P, Nur U, Crawford M, et al. The Social Functioning Questionnaire: a rapid and robust measure of perceived functioning. Int J Soc Psychiatry. 2005;51:265–275.
- D’Zurilla TJ, Nezu AM, Maydeu-Olivares A. Manual for the Social Problem Solving Inventory-Revised. North Tonawanda NY: Multi-Health Systems Inc.; 2000.
- Spielberger CD. STAXI-2: State-Trait Anger Expression Inventory-2. Odessa FL: Psychological Assessment Resources Inc.; 1999.
- Barratt ES. Anxiety and impulsiveness relating to psychomotor efficiency. Percept Mot Skills. 1959;9:191–198.
- Bernstein EM, Putnam FW. Development reliability and validation of a dissociation scale. J Nerv Ment Dis. 1986;178:448–454.
- Andrews B, Quian M, Valentine JD. Predicting depressive symptoms with a new measure of shame: the Experience of Shame Scale. Br J Clin Psychol. 2002;41:29–42.
- Howard R. How is personality disorder linked to dangerousness? A putative role for early-onset alcohol abuse. Med Hypotheses. 2006;67:702–708.
- Newhill CE, Vaughn MG, DeLisi M. Psychopathy scores reveal heterogeneity among patients with borderline personality disorder. Journal of Forensic Psychiatry and Psychology. 2009;21:202–220.
- Blackburn R. Subtypes of psychopath. In: McMurran M Howard RC, eds. Personality, personality disorder and violence. Chichester, United Kingdom: Wiley-Blackwell; 2009:113–132.
CORRESPONDENCE: Richard Howard, PhD, Forensic Mental Health, Institute of Mental Health, University of Nottingham Innovation Park, Sir Colin Campbell Building, Triumph Road, Nottingham NG7 2TU United Kingdom E-MAIL: firstname.lastname@example.org
Annals of Clinical Psychiatry ©2012 Quadrant HealthCom Inc.