July 2006, Vol. XXIII, No. 8
Substantial progress has been made in the last several decades in research providing evidence for the relationship between mental disorders in childhood and adulthood.1-4 As a consequence, child psychiatrists today have much better data on the longterm outcomes of childhood disorders, while adult psychiatrists now pay much greater attention to the developmental origins of the conditions they treat. A number of possible psychopathologic mediation mechanisms are currently being studied and discussed, including genetic mediation, kindling effects, environmental influences, coping mechanisms, and cognitive processing of experiences.4 In this context, researchers debate and discuss continuities and discontinuities in psychopathology through childhood, adolescence, and adult life.
Lee N. Robins, PhD, MA, gave us the first large-scale study on adults who displayed deviant behavior as children.5 She concluded that almost all adults with chronic antisocial behavior have shown conduct disorder symptoms as children, but only about one third of children with conduct disorder go on to have antisocial personality disorder as adults.6 However, as Zoccolillo and colleagues7 pointed out, this does not mean that two thirds of children with conduct disorder have a relatively good outcome. Antisocial personality disorder alone does not provide adequate coverage of the range of disabilities in adult life that can be considered the adult sequelae of childhood conduct disorder.
As research has repeatedly demonstrated, there is a strong correlation between childhood conduct disorder and a wide range of adverse psychosocial outcomesincluding crime, substance use, relationship problems, and poor mental health. Other possible outcomes include other personality disorders and Axis I disorders, particularly psychoactive substance use disorders and mood disorders.
This article reviews longitudinal studies that have reported on the link between adolescent conduct disorder and later adult personality disorders. It is meant to be representative of the current research status in the field but is by no means exhaustive. Although the DSM classifies conduct disorder and oppositional defiant disorder together with attention-deficit/hyperactivity disorder (ADHD) as disruptive behavior disorders, only studies discussing disruptive behavior disorders in general have been reviewed (ie, studies on ADHD only have been omitted). Follow-up studies on adolescents have been favored.
Review of major studies
I will briefly review several studies here. Of these, major research findings from 5 clinical studies7-11 and 2 community studies12,13 are summarized in the Table. The findings have been reported in a variety of ways; thus, they are not directly comparable except in cases in which researchers have chosen the same method of presentation.
Helgeland and colleagues8 conducted a follow-up of 130 former adolescent psychiatric inpatients 28 years after the index hospitalization. The data were divided into 2 main diagnostic groups: patients with emotional disorders in adolescence (n = 45) and patients with disruptive behavior disorder in adolescence (n = 85). Adolescents with disruptive behavior disorder were not more likely than adolescents without disruptive behavior disorder to have a personality disorder in adulthood. However, adolescents with disruptive behavior disorder were more likely to have a Cluster B personality disorder in adulthood than were adolescents with emotional disorders (odds ratio [OR], 8.62; 95% confidence interval [CI], 3.14 - 23.69), particularly antisocial personality disorder (OR, 75.92; 95% CI, 8.45 - 682.09) and borderline personality disorder (OR, 8.93; 95% CI, 1.87 - 42.70).
Regression analysis demonstrated that disruptive behavior disorder in adolescence was a significant and independent predictor of antisocial personality disorder in males but not in females. Antisocial personality disorder was more likely to develop in males with disruptive behavior disorder, whereas borderline personality disorder was more likely to develop in females with disruptive behavior disorder. On the other hand, an emotional disorder in adolescence was a significant and independent predictor of a Cluster C personality disorder in females but not in males. Results pertaining particularly to the disruptive behavior disorder subgroup are presented in the Table.
In a study of 158 former child psychiatric inpatients, investigators found that 62 (39.2%) had a personality disorder at follow-up, with borderline, paranoid, and obsessive-compulsive types most often encountered.9 The comorbidity rate was high, with 51 patients fulfilling criteria for more than 1 personality disorder. In the disruptive behavior disorder subpopulation (n = 75), prior disruptive behavior disorder increased the odds of adult narcissistic personality disorder more than 6-fold and the odds of antisocial personality disorder more than 5-fold.
Myers and colleagues10 investigated adolescents with concurrent conduct disorder and substance abuse and found that 61% met the criteria for antisocial personality disorder at follow-up. Early onset of conduct disorder, greater diversity of conduct disorder behaviors, and heavier drug use were significant predictors of adult antisocial personality disorder. The study did not investigate the remaining personality disorders.
Another study followed former adolescent psychiatric inpatients with either disruptive (n = 80) or emotional (n = 65) disorders.11 Patients with disruptive disorder were more likely to have a personality disorder at follow-up than those with emotional disorders (OR, 4.65; 95% CI, 1.92 - 11.31); Cluster B personality disorder was most often reported in this group (OR, 8.80; 95% C, 2.49 - 31.04). Those with conduct disorder were more likely than those without conduct disorder to have a personality disorder at follow-up (OR, 4.30; 95% CI, 1.16 - 15.89); antisocial personality disorder was most often reported in this group (OR, 4.58; 95% CI, 1.04 - 20.21). Youths with ADHD, on the other hand, were more likely than youths without ADHD to have borderline personality disorder at follow-up (OR, 12.11; 95% CI, 1.36 - 107.57).
The clinical study by Zoccolillo and colleagues7 differs from the aforementioned studies in that it examined 171 individuals who spent most of their childhood in children's homes. Of these, 61 had childhood conduct disorder. Recording 4 areas of potential adult dysfunction (crime, marriage, social, and work), the researchers found that 86% of males and 73% of females with childhood conduct disorder showed multiple maladaptations in adulthood. Genderspecific prevalences of adult personality disorder are given in the Table.
In a community study, Kasen and colleagues12 investigated the association between childhood psychopathology and young adult personality disorder in a random sample of 551 youths. The prevalences of Cluster A, B, and C personality disorders were 9.6%, 16.7%, and 8.2%, respectively. Because of overlap among clusters, the total percentage of young adults with personality disorder was 23.6%. Previous disruptive disorders, anxiety disorders, and major depression all significantly increased the odds of young adult personality disorder. The relative risk of young adult personality disorder for youths in whom disruptive behavior disorder was previously diagnosed is cited in the Table.
Lewinsohn and colleagues13 examined the occurrence of elevated dimensional personality disorder scores at age 24 in a community sample; the subjects had been assessed twice as adolescents. The researchers found a much a lower prevalence of personality disorder than Kasen and colleagues had found12: 3.8% in those with a history of Axis I disorder versus 1.7% in those without an Axis I history. The occurrence of major depression, anxiety disorder, disruptive behavior disorder, and substance use disorder in childhood and adolescence was associated with elevated adult personality disorder dimensional scores. Although the researchers initially found an association between history of disruptive behavior disorder and 7 of the 10 specific personality disorder dimensional scores, only associations with 2 personality disorders in Cluster Bantisocial personality disorder and narcissistic personality disorderremained significant after adjusting for potential confounders, such as gender, race, education, and other Axis I disorders in childhood and adolescence.
Bernstein and colleagues14 investigated the childhood antecedents of personality disorder diagnosed in adolescence in a community sample (n = 641). They found that childhood conduct problems remained an independent predictor of adolescent personality disorder in all 3 clusters (A, B, and C), even when other childhood problems were included in the same regression model. The researchers concluded that personality disorder can be traced back to childhood disturbances.
Data from the National Institute of Mental Health Epidemiologic Catchment Area Project showed that prevalence of adult disorders was correlated with the number of conduct problems in childhood.15 Dichotomizing the adult disorders into externalizing (adult antisocial behavior and substance use) disorders and nonexternalizing (all other) disorders, Robins and Price15 found that the effect of conduct problems on nonexternalizing disorders was largely mediated through externalizing disorders, particularly in males. This implies, according to those investigators, that conduct disorder in younger cohorts may in part be responsible for the rising rate of other disorders.
In a study of childhood risk factors for young adult personality disorder, Cohen16 found an overall correlation of 0.36 between adolescent disruptive behavior disorder and young adult personality disorder, with higher correlations in females (0.43) than in males (0.29). Altogether, 30% of all personality disorder cases were potentially attributable to childhood disruptive behavior disorder symptoms.
Burket and Myers17 studied DSM-IIIR personality disorder comorbidity in 25 adolescents hospitalized with conduct disorder. DSM-III-R allows the use of personality disorder diagnoses in adolescents; the exception is antisocial personality disorder, which is not applicable to adolescents under age 18. These investigators found high prevalences of personality disorders in these adolescents: passive aggressive personality disorder, 56%; borderline personality disorder, 32%; and paranoid, histrionic, sadistic, and self-defeating personality disorder, 28% each.
Available research demonstrates a high prevalence of personality disorder in adults who had conduct disorder or disruptive behavior disorder as children. Disruptive behavior disorderor, more specifically, conduct disorderin childhood and adolescence significantly increases the risk of adult personality disorder, with one third to two thirds of these young persons having at least 1 personality disorder (most often antisocial personality disorder) at adult follow-up. In addition, conduct disorder in younger cohorts may be at least partly responsible for rising rates of later nonexternalizing disorders.15 These findings support the view of conduct disorder as a complex illness with extensive implications for later adult mental health.17 On a more positive note, primary prevention and early intervention in conduct disorder may be expected to reduce the rates of a broad range of later disorders.
The clinical studies highlighted in this discussion recruited mainly inpatients, indicating that the patients investigated were gravely afflicted. There seems to be a paucity of studies reporting on outcome in adolescents with less serious conduct disorder/disruptive behavior disorder. One would expect adolescent outpatients to exhibit less serious adult psychopathology than inpatients do, including lower prevalences of personality disorder.
The results regarding community samples are uncertain, because the community studies presented here12,13 differed considerably in their estimates of the prevalence of subsequent adult personality disorder. There may be a number of reasons for this apparent disparity, including differences in such factors as inclusion criteria, attrition mechanisms, follow-up time, and DSM version used. These issues should be addressed in future research.
Lewinsohn and colleagues13 found that although a history of disruptive behavior disorder initially was associated with 7 of the specific personality disorder dimensional scores, only the associations with antisocial personality disorder and narcissistic personality disorder remained significant after adjusting for potential confounders. These authors pointed to an important implication of this interesting negative finding: they proposed that the association between disruptive behavior disorder and most Axis II psychopathology may in part be accounted for by comorbid Axis I disorders.
The persistent finding of gender differences by most researchers in the adult outcome of adolescent conduct disorder and disruptive behavior disorder8,11,13,14,16 indicates that similar childhood and adolescent problems may lead to dissimilar adult personality disorders. This suggests that conduct disorder and disruptive behavior disorder proceed along different developmental pathways in males and females, resulting in different adult outcomes despite similar adolescent psychiatric disorders.
Although DSM-IV cautions against routine diagnosis of personality disorder in adolescents, a growing body of research over the last decade has suggested that personality disorders may be recognizable during adolescence.17,18,19 Westen and colleagues20 have, with some exceptions, found that personality pathology in adolescence resembles that in adulthood and is diagnosable in adolescents aged 14 through 18. Investigating incarcerated juveniles aged 11 through 17, Eppright and colleagues21 found a considerable number of inmates with concurrent conduct disorder and personality disorder; antisocial personality disorder was most common in males and borderline personality disorder was most common in females.
The differential diagnosis between, or indeed the comorbidity of, conduct disorder with adolescent bipolar disorder should be considered.22 Although important distinctions are apparent, specifically at the neural level, the obvious commonalities of behavioral deficits in the 2 disorders may reflect some underlying relationship. These issues need clarification in future research.
Future studies on the continuities from adolescent conduct disorder to adult social and mental health functioning should not confine themselves to antisocial personality disorder, as has so often been the case in past studies. Part of the apparent discontinuity between behavior in childhood and adulthood may be an artifact of using only antisocial personality disorder as the adult outcome measure.6 In addition to the remaining personality disorders, Axis I disorders should be included. It seems that only some of the maladaptive behaviors in adult life following adolescent conduct disorder are properly captured by current Axis I and II diagnostic criteria.6
Moreover, the heterogeneity of conduct disorder has been debated. The proposed taxonomy of adolescence-limited and life-course-persistent antisocial behavior types presented by Moffitt23 has gained much support, as has the tripartite model of delinquent development presented by Loeber and Hay.24 Different subtypes could be expected to have different adult outcomes; these issues should be explored.
Kasen and colleagues12 suggest that adolescent Axis I disorders may set in motion a chain of maladaptive behaviors and environmental responses that foster more persistent psychopathology over time. In turn, these disturbances have serious implications for adult mental health and social functioning. Because Axis I disorders in children may be more amenable to treatment than adult personality disorder, early intervention among at-risk children would be desirable. This may be particularly true in conduct disorder. Early identification and successful intervention may help reduce the risk for subsequent debilitating adult personality disorder and other serious malfunctions.
Dr Kjelsberg is a senior researcher at the Centre for Research and Education in Forensic Psychiatry in the psychiatric division of Ulleval University Hospital, Oslo, Norway. She reports no conflicts of interest regarding the subject matter of this article.
1. Rutter M. Relationships between child and adultpsychiatric disorders. Some research considerations.Acta Psychiatr Scand. 1972;48:3-21.
2. Rutter M. Psychopathology and development. I.Childhood antecedents of adult psychiatric disorder.Aust N Z J Psychiatry. 1984;18:225-234.
3. Rutter M. Relationships between mental disordersin childhood and adulthood. Acta PsychiatrScand. 1995;91:73-85.
4. Rutter M, Kim-Cohen J, Maughan B. Continuitiesand discontinuities in psychopathology betweenchildhood and adult life. J Child Psychol Psychiatry.2006;47:276-295.
5. Robins LN. Deviant children grown up. Baltimore:Williams & Wilkins; 1966.
6. Robins LN. Sturdy childhood predictors of antisocialbehaviour: replications from longitudinal studies.Psychol Med. 1978;8:611-622.
7. Zoccolillo M, Pickles A, Quinton D, Rutter M. Theoutcome of childhood conduct disorder: implicationsfor defining adult personality disorder and conductdisorder. Psychol Med. 1992;22:971-986.
8. Helgeland MI, Kjelsberg E, Torgersen S. Continuitiesbetween emotional and disruptive behavior disordersin adolescence and personality disorders inadulthood. Am J Psychiatry. 2005;162:1941-1947.
9. Ramklint M, von Knorring AL, von Knorring L,Ekselius L. Child and adolescent psychiatric disorderspredicting adult personality disorder: a followupstudy. Nord J Psychiatry. 2003;57:23-28.
10. Myers MG, Stewart DG, Brown SA. Progressionfrom conduct disorder to antisocial personality disorderfollowing treatment for adolescent substanceabuse. Am J Psychiatry. 1998;155:479-485.
11. Rey JM, Morris-Yates A, Singh M, et al.Continuities between psychiatric disorders in adolescentsand personality disorders in young adults. AmJ Psychiatry. 1995;152:895-900.
12. Kasen S, Cohen P, Skodol AE, et al. Influence ofchild and adolescent psychiatric disorders on youngadult personality disorder. Am J Psychiatry. 1999;156:1529-1535.
13.Lewinsohn PM, Rohde P, Seeley JR, Klein DN.Axis II psychopathology as a function of Axis I disordersin childhood and adolescence. J Am Acad ChildAdolesc Psychiatry. 1997;36:1752-1759.
14. Bernstein DP, Cohen P, Skodal A, et al. Childhoodantecedents of adolescent personality disorders. AmJ Psychiatry. 1996;153:907-913.
15. Robins LN, Price RK. Adult disorders predictedby childhood conduct problems: results from theNIMH Epidemiologic Catchment Area project. Psychiatry.1991;54:116-132.
16. Cohen P. Childhood risks for young adult symptomsof personality disorder: method and substance.Multivariate Behav Res. 1996;31:121-148.
17. Burket RC, Myers WC. Axis I and personalitycomorbidity in adolescents with conduct disorder.Bull Am Acad Psychiatry Law. 1995;23:73-82.
18. Grilo CM, McGlashan TH, Quinlan DM, et al. Frequencyof personality disorders in two age cohortsof psychiatric inpatients. Am J Psychiatry. 1998;155:140-142.
19. Johnson JG, Cohen P, Smailes E, et al. Adolescentpersonality disorders associated with violence andcriminal behavior during adolescence and early adulthood.Am J Psychiatry. 2000;157:1406-1412.
20. Westen D, Shedler J, Durrett C, et al. Personalitydiagnoses in adolescence: DSM-IV axis II diagnosesand an empirically derived alternative. Am J Psychiatry.2003;160:952-966.
21. Eppright TD, Kashani JH, Robison BD, Reid JC.Comorbidity of conduct disorder and personalitydisorders in an incarcerated juvenile population. AmJ Psychiatry. 1993;150:1233-1236.
22. Kyte ZA, Carlson GA, Goodyer IM. Clinical andneuropsychological characteristics of child andadolescent bipolar disorder. Psychol Med. 2006;Mar28:1-15 (E-pub ahead of print).
23. Moffitt TE. Adolescence-limited and life-coursepersistentantisocial behavior: a developmental taxonomy.Psychol Rev. 1993;100:674-701.
24. Loeber R, Hay D. Key issues in the developmentof aggression and violence from childhood to earlyadulthood. Annu Rev Psychol. 1997;48:371-410.