 |
 |
|
|
|
Although all personality disorders are conceived of as stable conditions (“enduring pattern[s] of inner experience and behavior”) (1, p 263), ASPD is unique because its diagnostic criteria require the presence of another disorder—CD. The DSM specifically stipulates that a diagnosis of ASPD requires the presence of CD before age 15 years. The DSM only mentions AB that meets the post–age 15 years criteria for ASPD (that is, the Criterion A symptoms in DSM-IV, also called AAB) in a discussion of differential diagnosis and in the “other conditions that may be a focus of clinical attention” section. The DSM states, “only when antisocial personality traits are inflexible, maladaptive, and persistent and cause significant functional impairment or subjective distress do they constitute Antisocial Personality Disorder” (1, p 649). There is no clear evidence, however, that AB with an onset age of 15 years or later lacks these characteristics. Further, the current adult criteria for ASPD are quite similar to those of the DSM-III, which were derived from Robins’ work (2) and based on a desire to define a group of adults who were clearly antisocial. The present study examined the demographic characteristics and patterns of psychiatric comorbidity among a nationally representative sample of individuals exhibiting AAB but not CD, compared with individuals exhibiting CD but not AAB, individuals exhibiting ASPD, and individuals not exhibiting an AB disorder. Results indicating that AAB in the absence of CD is either extremely rare or not associated with significant problems would imply that the relative neglect of these individuals by our current diagnostic system is appropriate. Results indicating that AAB is similar to full ASPD, however, would imply that the current diagnostic practice of marginalizing late-onset AB (and the resulting paucity of research on it) is problematic. Neither of the dominant models of AB—that of the DSM (described above) and Moffitt’s taxonomy (described below)—predict AB that begins in mid- to late adolescence. Moffitt’s model distinguishes between life-course-persistent and adolescence-limited AB: individuals with life-course-persistent AB begin exhibiting AB in childhood, and adolescence-limited antisocial individuals begin “tapering off” their AB in mid- to late adolescence (3). However, several studies note the existence people exhibiting AAB but not CD. Among a sample of drug abusers, 33% of men and 42% of women exhibited AAB but not CD (4). Similarly, among a group of intravenous drug users, Brooner and others reported that 24% exhibited AAB but not CD (5). Several community-based studies have also discussed people who suffer from AAB but not CD.Tweed and others, for example, found that approximately 50% of adults who demonstrated severe AB in adulthood demonstrated no significant AB during childhood (6), and Ridenour and others found that 2.3% of men and 0.9% of women meeting diagnostic criteria for AAB had no conduct problems in childhood or adolescence (7). Elkins and others reported that 28% of ever-antisocial men suffered from AAB, but no more than one symptom of CD prior to age 15 years (8). Most recently, Marmorstein and Iacono compared subjects exhibiting persisting AB (CD and symptoms of AAB) with subjects exhibiting desisting AB (CD only) and individuals exhibiting late-onset AB (symptoms of AAB but no CD) and found striking similarities in background characteristics (for example, rates of parental divorce) and risk for future substance dependence between the late-onset group (which comprised 20% of people who demonstrated substantial antisocial behavior at any point in their lives) and the persisting group (9). There was one striking difference between the groups: the late-onset group was predominantly female (72%), whereas the persisting and desisting groups were predominantly male (78% and 77%, respectively). Moffitt’s model of AB, which has received considerable empirical support, does not include a late-onset subtype (10). However, Silverthorn and Frick propose that the typical correlates of childhood-onset and adolescence-limited AB may not apply to girls (11). Instead, they propose that many girls exhibit an onset of AB during adolescence but experience many of the problems typically associated with the childhood-onset pathway in boys (they termed this pattern the “delayed onset” pathway). Given the predominantly female nature of late-onset AB (beginning after age 15 years) found by Marmorstein and Iacono, it is possible that sex-specific factors account for this pattern (9). Perhaps parents are more strict with girls than with boys, inhibiting the girls’ AB until they gain the freedom associated with mid- to late adolescence (for example, having friends with cars). Alternately, perhaps other psychiatric disorders place youth at risk for late-onset AB (for example, major depression is associated with AB) (12,13). It is possible that some girls experience an onset of depression in early adolescence that predisposes them to developing AB a few years later. In this context, the present study examined the demographic features and patterns of comorbidity of individuals with AAB but not CD, compared with individuals who had ASPD, individuals who had CD but not AAB, and a nonantisocial control group. Demographic similarities as well as similar patterns of comorbidity between people with full ASPD and individuals with AAB but not CD would support their prominent inclusion in a diagnostic scheme. In contrast, similarities between individuals in the nonantisocial control group and individuals with AAB without ASPD, or evidence that AAB without ASPD is extremely rare, would indicate that their problematic behaviour is fairly isolated and, therefore, perhaps not worthy of substantial attention. From previous studies, I expected that individuals with AAB but not CD would be quite similar to individuals with full ASPD (CD and AAB), with one exception: I predicted there would be more women in this group than in the CD-only and ASPD groups. I also considered the possibility that internalizing disorders, which are generally more prevalent among female subjects than male subjects, may also be overrepresented in this group (compared with the CD-only and ASPD groups). MethodParticipants Participants for this study were drawn from the NCS, a nationally representative household sample of 8098 people ranging in age from 15 to 54 years. Data were collected between 1990 and 1992, and 82.4% of those selected for the study participated. Of respondents, 52% were female. Kessler and others provide further detail regarding the design of the NCS, the sample, and the data (14-17). Only participants aged 18 years or over at the time of the assessment (n = 7612) were included in the study. This was for 2 reasons: 1) to give participants at least 3 years in which to develop AAB symptoms, which must occur after age 15 years; and 2) the DSM diagnosis of ASPD has a cutoff of age 18 years, making those under age 18 years ineligible for the diagnosis. Participants were divided into 4 groups: the nonantisocial control group (not meeting criteria for CD or AAB; n = 6496 or 85.3% of the sample), the ASPD group (meeting diagnostic criteria for both CD and AAB; n = 261 or 3.4% of the sample), the CD-only group (meeting diagnostic criteria for CD but not AAB; n = 665 or 8.7% of the sample), and the AAB-only group (meeting diagnostic criteria for AAB but not CD; n = 190 or 2.5% of the sample). The number of symptoms of each disorder in each group were as follows:
Measures Diagnoses. Diagnoses were determined from fully structured interviews administered by trained lay interviewers using a paper-and-pencil format. This structured interview (the Baseline NCS Interview Schedule) is a slight modification of the CIDI (18). Kessler and others detail the modifications made to this interview (17). Diagnoses of DSM-III-R (19) CD, AAB, and ASPD were computed from respondents’ reports on an interview asking about each symptom of these disorders. The number of endorsed symptoms was then totalled, and diagnoses were computed. Respondents were asked about the presence of CD symptoms prior to age 15 years as well as the presence of AAB symptoms after age 15 years. Of the DSM-III-R comorbid diagnoses, alcohol abuse, alcohol dependence, drug abuse, drug dependence, major depression, dysthymia, generalized anxiety disorder, social phobia, simple phobia, panic disorder, agoraphobia, and posttraumatic stress disorder were examined. Hierarchy requirements were not used, even when required by the DSM-III-R (for example, alcohol abuse was diagnosed even if alcohol dependence was also present). In addition, I computed 3 composite variables indicating the presence of one or more diagnoses in a particular category: any substance use disorder, any depressive disorder, and any anxiety disorder. These were used as additional possible comorbid “diagnoses” in order to examine relations between different trajectories of AB and other categories of psychopathology. Demographic Information. Information regarding sex, education level, and socioeconomic status was assessed via structured interview. Sex, a dichotomous poverty variable (living in poverty, compared with not living in poverty), personal income (yearly dollars), and number of years of formal education were the specific characteristic examined. Statistical Analyses To adjust for differences in probabilities of selection within and between households, differences in nonresponse, and poststratification, sample weights (a combination of the weights that were used to adjust for each of those factors, as discussed in other NCS papers, such as Kessler; 14) were applied to the data prior to conducting the analyses. For all dichotomous dependent variables (percentage female, percentage in poverty, and all cooccurring diagnoses), I conducted chi-square analyses to examine the associations of each dependent variable with the 4 trajectories of AB (control, CD-only, AAB-only, and ASPD) groups. When the initial chi-square was significant, I conducted pairwise chi-square analyses to determine which specific group differences were significant. For continuous dependent variables (annual personal income in dollars and education in years), I conducted 1-way ANOVAs to compare the levels of each among the 4 trajectory groups. When the initial F-value was significant, I used Bonferroni posthoc tests to determine which trajectory groups were significantly different. ResultsFrequency of Trajectories of AAB Most of the sample (over 85%) did not meet criteria for CD or AAB. Of the sample, 3.6% demonstrated ASPD (the presence of both CD and AAB). CD without AAB was present in 8.9% of the sample. AAB without CD was present in 2.3% of the sample. It should be noted that, when raw data rather than weighted data were examined, each prevalence figure changed by 0.2% or less. Cooccurring Diagnoses Table 1 presents the results of chi-square analyses examining the prevalence of different cooccurring disorders among the antisocial trajectory groups. In all cases, the overall chi-square was highly significant. The superscripts in Table 1 present the results of pairwise chi-square analyses comparing the prevalence of each cooccurring disorder among specific pairs of trajectory groups. These analyses are informative about specific group differences. In the text below, I also discuss comparisons between the AAB-only group and the other groups.
When the composite substance use disorder variable was examined, the AAB-only group had higher rates than the control and CD-only groups. The AAB-only group also had lower rates than the ASPD group. This pattern remained constant for both alcohol and drug dependence diagnoses. In contrast, for both alcohol and drug abuse diagnoses, the AAB-only and ASPD participants had similar (not statistically different) rates. Overall, the AAB-only participants had slightly lower rates of substance use disorders, on average, than ASPD participants. The difference in rates between these 2 groups averaged approximately 10%. When the composite depressive disorder variable was examined, rates of depressive disorders among AAB-only and ASPD participants were comparable with and higher than those among the CD-only and control groups. Patterns were similar for major depression and dysthymia. When the composite anxiety disorder variable was examined, rates of anxiety disorders among AAB-only and ASPD participants were comparable with and higher than those among the CD-only and control groups. This pattern remained constant for social phobia, agoraphobia, panic disorder, generalized anxiety disorder, and posttraumatic stress disorder. AAB-only participants had the highest rates of simple phobia—significantly higher than all other groups. Demographic Characteristics Table 2 presents the results of group comparisons on demographic characteristics. Comparisons between the AAB-only group and the other groups are discussed below.
There were more women in both the CD-only and AAB-only groups than in the ASPD group. The highest percentage of women was in the control group. Examination of the percentage of each group living in poverty revealed that the control group had a significantly lower percentage than any of the antisocial groups (who all had higher and statistically equivalent rates). Examination of personal income revealed that control group members and CD-only group members had similar, relatively high levels of personal income, while AAB-only and ASPD participants had similar, relatively low levels of personal income. Members of the control group had the highest number of years of education. CD-only and AAB-only participants had intermediate and approximately equivalent numbers of years, while ASPD participants had the lowest (statistically different from all other groups) levels of education. Antisocial Trajectories Based on More Distinct Groups I considered the possibility that the aforementioned analyses were based on insufficiently distinct groups. For example, an AAB-only participant may have 2 symptoms of CD, which does not meet the diagnostic threshold for CD but would still not constitute a true “late-onset” trajectory of AB. Conversely, a CD-only participant may have 2 symptoms of AAB, which does not meet the diagnostic threshold for AAB but would still involve engagement in significant AB after age 15 years. Acting on this possibility, I split the sample into 4 more distinct groups according to symptom-count data: members of the control group had zero symptoms of either CD or AAB (n = 3100 or 40.7% of the sample), ASPD participants had full CD and AAB diagnoses (n = 261 or 3.4% of the sample), strict CD-only participants had a diagnosis of CD but zero symptoms of AAB (n = 155 or 2.0% of the sample), and strict AAB-only participants had a diagnosis of AAB but zero symptoms of CD (n = 57 or 0.7% of the sample). I then repeated all analyses using this grouping strategy. The results were strikingly similar to those presented above. Consistent with the main results, all initial c2 and F-values were significant. The pattern of group differences was quite similar as well. With respect to the pattern of differences and similarities between the ASPD and AAB-only groups, there were only 3 differences from the results presented above. There was no longer a significant difference between these 2 groups in their rates of drug dependence (47.7% among the ASPD group and 46.2% among the AAB-only group; c2 = 0.03, P > 0.05). There was also no longer a significant difference between these 2 groups in their likelihood of having any substance use disorder (82.6% of the ASPD group and 74.4% of the AAB-only group; c2 = 1.43, P > 0.05). In addition, there was no longer a significant difference between these 2 groups in their rates of simple phobia (17.9% of the ASPD group and 25.6% of the AAB-only group; c2 = 1.24, P > 0.05). The primary results of this study do not seem to be owing to AB patterns falling just above or just below diagnostic thresholds. When I used more clearly distinct groups, the number of people falling into the different antisocial trajectories decreased; however, the basic pattern of comorbidities and demographic characteristics remained similar. DiscussionThe results of this study indicate that a substantial number of individuals (approximately 2.3%) meet diagnostic criteria for AAB but not for CD or ASPD and that these individuals experience levels of comorbid diagnoses and disadvantage similar to those of individuals with full ASPD diagnoses. There were only a few significant differences between the ASPD and AAB-only groups: AAB-only participants had somewhat lower rates of substance (alcohol and drug) dependence, AAB-only participants had higher rates of simple phobia, AAB-only participants had more average years of education, and AAB-only participants were more likely to be female. For all other characteristics, AAB-only and ASPD participants were quite similar and typically significantly worse off than both members of the control group and CD-only participants. Individuals who begin to display significant AB in midadolescence or later experience significant disability similar to that experienced by those exhibiting full ASPD. Given that 2.3% of the population appears to demonstrate such a pattern, it is concerning that these individuals are excluded from the DSM diagnosis of ASPD. Owing to the DSM requirement for this diagnosis of CD prior to age 15 years, individuals demonstrating this pattern of AB are excluded from many research studies and may be overlooked in some clinical settings. This seems problematic, given that the levels of comorbid psychopathology and disadvantage they experience is on par with that associated with ASPD. Strikingly, when I applied more stringent trajectory criteria (requiring CD-only participants to have zero symptoms of AAB, instead of simply not meeting the AAB diagnostic criteria, and requiring AAB-only participants to have zero symptoms of CD, instead of simply not meeting the CD diagnostic criteria) even fewer significant differences between the ASPD and the AAB-only groups emerged. Thus the primary results of this study are not owing to participants who exhibit lifelong AB but to those who simply fall one symptom short of the diagnostic requirements of CD or AAB. Secondarily, I found that participants with only CD tended to demonstrate rates of comorbidity higher than those of members of the control group and lower than those of participants with either AAB-only or ASPD; similarly, demographic data indicated that in terms of income and education they also tended to be somewhat better off than the AAB-only and ASPD participants, but worse off or similar to members of the control group. Although AB that stops by midadolescence is associated with some degree of comorbid psychopathology and disadvantage, these problems are less severe than those associated with AB that occurs either in childhood through adulthood or in late adolescence and adulthood only. There were more women in the CD-only and AAB-only groups (29.9% and 30.1%, respectively) than in the ASPD group (18.4%). Thus female subjects are more likely to display transient AB, either in childhood through early adolescence or in late adolescence through adulthood, than they are to exhibit lifelong AB. This finding is not particularly supportive of Silverthorn and Frick’s hypothesis that females are more likely than male subjects to display a late-onset pattern of AB (11); instead, it merely indicates that female subjects are more likely to evidence either early- or late-onset AB, but are less likely than male subjects to engage in lifelong AB. The gender composition of these groups also differs from that reported by Marmorstein and Iacono, who found that female subjects comprised 72% of the late-onset (AAB-only) group but only 23% of the desisting (CD-only) group of a community-based sample (9). The reason for this difference is unclear. Marmorstein and Iacono examined antisocial symptoms that possibly occurred up to age 17 years. Perhaps many male subjects who display the AAB-only pattern begin to do so at an even later age and, therefore, were detected by the present study, since it includes participants up to age 54 years. Because participants in the Marmorstein and Iacono study had only 2 years in which to develop symptoms of AAB, they did not need to meet full criteria for AAB to be considered part of the “late-onset” group (a 2-AAB-symptom cutoff was used). Perhaps female subjects who have not had CD tend to demonstrate only 2 AAB symptoms after age 15 years and, therefore, do not meet criteria for the AAB-only group in the present study. Future research using other samples will be necessary to clarify this discrepancy. This study has several clinical implications. First, clinicians should be aware that some individuals begin to engage in significant AB during midadolescence (age 15 years) or later. Owing to this, clinicians should be mindful that individuals who do not meet diagnostic criteria for CD prior to age 15 years may still begin engaging in AB as older adolescents or adults. Those who begin engaging in significant AB during midadolescence or later are also as prone to poverty and low income, substance abuse, depressive disorders, and anxiety disorders as are those exhibiting full ASPD. Although the DSM differentiates ASPD from AAB by noting that ASPD “cause[s] significant functional impairment or subjective distress,” (p 649) thereby implying that AAB is not associated with these things, the results of this study imply that this distinction is not appropriate: AAB is associated with significant dysfunction and disability. This study has several strengths, including its use of a large, nationally representative sample and its careful assessment of a range of forms of psychopathology and demographic characteristics. Of course, it also has limitations. Although the age 15 years cutoff for assessing CD, compared with AAB, symptoms is consistent with both the DSM and other large studies, it is not clear whether this is the optimal age cutoff. The implications of an earlier or later cutoff are unknown. I did not examine the specific nature of the symptoms exhibited by participants in each group. It is possible, for example, that subjects in the CD-only and AAB-only groups exhibited different, or less severe, symptoms than people in the ASPD group. Additionally, because ages of onset and offset of CD and AAB symptoms were not assessed, I was unable to examine how these trajectories of AB might have changed over time. I was also unable to examine alternative ways of defining trajectories of AB, such as Moffitt’s adolescence-limited, compared with life-course-persistent, typology. It was also impossible to examine, in cases of comorbidity, which disorder appeared first. For instance, it is possible that some of the AAB-only participants developed AAB secondary to substance abuse or dependence. In summary, over 2% of the population demonstrates a late-onset pattern of AB. Individuals exhibiting AAB after age 15 years but not CD prior to age 15 years are at levels of risk comparable to those of individuals suffering from full ASPD for dysfunction and disability and, therefore, are equally in need of research and clinical attention. Future research addressing other ways of subtyping developmental patterns of AB would be particularly useful. Funding and SupportData for this study were drawn from the NCS, which was funded by the National Institute of Mental Health (Grants R01MH/DA46376 and R01 MH49098), the National Institute of Drug Abuse (through a supplement to R01 MH/DA46376), and the WT Grant Foundation (Grant 90135190). AcknowledgementI thank Ronald Kessler and the other NCS investigators for their work on this study and the Inter-University Consortium for Political and Social Research for making this data set available for public use. References1. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; 1994. 2. Robins LN. Deviant children grown up. Baltimore (MD): Williams and Wilkins; 1966. 3. Moffitt TE. Adolescence-limited and life-course-persistent antisocial behaviour: a developmental taxonomy. Psychol Rev 1993;100:674–701. 4. Cottler LB, Price RK, Compton WM, Mager DE. Subtypes of AAB among drug abusers. J Nerv Ment Dis 1995;183:154–61. 5. Brooner RK, Schmidt CW, Felch LJ, Bigelow GE. Antisocial behaviour of intravenous drug abusers: Implications for diagnosis of antisocial personality disorder. Am J Psychiatry 1992;149:482–7. 6. Tweed JL, George LK, Blazer D, Swartz M, MacMillan J. Adult onset of severe and pervasive antisocial behaviour: a distinct syndrome? J Personal Disord 1994;8:192–202. 7. Ridenour TA, Cottler LB, Robins LN, Compton WM, Spitznagel EL, Cunningham-Williams RM. Test of the plausibility of adolescent substance use playing a causal role in developing adulthood antisocial behaviour. J Abnorm Psychol 2002;111:144–55. 8. Elkins IJ, Iacono WG, Doyle AE, McGue M. Characteristics associated with the persistence of antisocial behaviour: results from longitudinal research. Aggress Violent Behav 1997;2:101–24. 9. Marmorstein NR, Iacono WG. Longitudinal follow-up of adolescents with late-onset antisocial behaviour: a pathological yet overlooked group. J Am Acad Child Adolesc Psychiatry 2005;44:1284–91. 10. Moffitt TE, Caspi A, Harrington H, Milne BJ. Males on the life-course-persistent and adolescence-limited antisocial pathways: follow-up at age 26 years. Dev Psychopathol 2002;14:179–207. 11. Silverthorn P, Frick PJ. Developmental pathways to antisocial behaviour: the delayed-onset pathway in girls. Dev Psychopathol 1999;11:101–26. 12. Marmorstein NR, Iacono WG. Major depression and conduct disorder in a twin sample: gender, functioning, and risk for future psychopathology. J Am Acad Child Adolesc Psychiatry 2003;42:225–33. 13. McCracken JT, Cantwell DP, Hanna GL. Conduct disorder and depression. In: Koplewicz HS, Klass E, editors. Monographs in Clinical Pediatrics. Volume 6. Depression in children and adolescents. New York (NY): Harwood Academic Publishers; 1993. p 121–32. 14. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, and others. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in the United States: results from the National Comorbidity Survey. Arch Gen Psychiatry 1994;51:8–19. 15. Kessler, Ronald C. National Comorbidity Survey, 1990–1992 Ann Arbor (MI): University of Michigan, Survey Research Center. 2nd ICPSR ed. Inter-university Consortium for Political and Social Research; 2002. 16. Kessler, RC. The National Comorbidity Survey of the United States. Int Rev Psychiatry 1994;6:365–76. 17. Kessler RC, Wittchen HU, Abelson JM, McGonagle K, Schwarz N, Kendler KS, and others. Methodological studies of the Composite International Diagnostic Interview (CIDI) in the US National Comorbidity Survey (NCS). Int J Methods Psychiatr Res 1998;7:33–55. 18. Composite International Diagnostic Interviews (Version 1.0). Geneva (CH): World Health Organization; 1990. 19. Diagnostic and statistical manual of mental disorders. 3rd ed, revised. Washington (DC): American Psychiatric Association; 1987. Author(s)Manuscript received July 2005, revised, and accepted November 2005. 1. Assistant Professor, Department of Psychology, Rutgers University, Camden, New Jersey. Address for correspondence: Dr N R Marmorstein, Department of Psychology, 311 North 5th Street, Camden, NJ 08102 e-mail: marmorst@camden.rutgers.edu
1 | 2
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||