Introduction
Accumulating research on the clinical relevance of callous-unemotional (CU) traits in youth has prompted change in the DSM classification system, reflecting the importance of assessing such traits in children. Earlier DSM editions (i.e., DSM-III/IV; APA
1994,
2000) further specified conduct disorder (CD) diagnosis only by the age of onset, differentiating between childhood and adolescent onset. In the DSM-5 this specification has been supplemented with the specifier “with limited prosocial emotions” (LPE), assessing whether individuals demonstrate limited prosocial emotions, including reduced guilt and empathy (APA
2013, p. 470). The LPE specifier operationalizes personality features reflecting “callous-unemotional” (CU) traits. However, in a recent review, Salekin (
2017) noted that DSM-5 includes no similar specifiers covering additional dimensions of psychopathic traits in children, despite mounting research indicating that both grandiose-manipulative (also conceptualized as grandiose-deceitful, GD) and daring-impulsive (also conceptualized as impulsivity/need for stimulation, INS) traits can be reliably assessed and exhibit theoretically relevant outcomes in youth. According to Salekin (
2017), grandiose-manipulative traits are related to both increased aggressive behavior and decreased prosocial behavior, with some research indicating that this relationship is even stronger with GD traits than CU traits. Daring-impulsive traits, in contrast, appear to be more related to educational difficulties and general risk-taking behaviors, but have also been associated with elevated conduct problems (Salekin
2016).
Research has so far narrowly focused on associations between CU traits and CD. Although this has improved our knowledge, we need additional research covering the broader concept of psychopathy, including all its underlying dimensions (Salekin
2016). Other topics also warrant further research, such as interactions between CD and childhood-onset neurodevelopmental problems (NDPs) that could affect developmental pathways to aggressive antisocial behavior in adults. Previous research suggests that subgroups of youths with neurocognitive dysfunctions (e.g., deficient decision-making) also display CD (Blair et al.
2014). Several studies have demonstrated considerable overlap between the presence of CD and neurodevelopmental problems, such as attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD) (e.g., Kerekes et al.
2014; Lundström et al.
2011).
In reviewing the adult phenotype of antisocial personality disorder (ASPD), Hofvander et al. (
2009) demonstrated that one developmental trajectory of this behavioral condition could be regarded as progression from childhood ADHD to subsequent oppositional defiant disorder and eventual CD. According to Hofvander et al. (
2009), about half of all children on this trajectory would meet ASPD criteria as adults. It has also been suggested that varying degrees of CU and psychopathic traits can be present in individuals with ASD, although the data suggest that these tendencies likely express different affective/informational processes (Jones et al.
2010). Another theory is that such a combination of traits could reflect a “double hit”, that is CU traits being an additional impairment adding further dysfunction to ASDs (Rogers et al.
2006). A recent study of genetic and environmental influences demonstrated that ASDs and elevated CU traits, while superficially similar, are etiologically independent (O’Nions et al.
2015). In summary, the relationship between NDPs (i.e., ADHD and ASD), CD, and psychopathic personality traits (CU traits so far have been the primary research focus) is complex and requires further study.
Method
Participants were selected from the ongoing Child and Adolescent Twin Study in Sweden (CATSS), a nationwide longitudinal study of twins from the age of 9 years whose inclusion started in 2004 (see Anckarsäter et al.
2011, for a detailed presentation of CATSS). The overall response rate for this study has been high, with an 80% inclusion rate as of the beginning of 2010. Participants were 9- and 12-year-old twins whose parents completed all CPTI-SV items (Colins et al.
2014) and the Autism—Tics, AD/HD and other Comorbidities (A-TAC) inventory (Larson
2013). From the total dataset (
N = 22,202) a subsample of 8762 individuals met these criteria, of whom 4453 (50.8%) were boys and 4309 (49.2%) were girls. In this study sample, 4185 (47.8%) were 9 years old and 4577 (52.2%) were 12 years old at baseline. Zygosity was distributed as follows: 29% (2545) monozygotic (MZ), 69.2% (6063) dizygotic (DZ), and 1.8% (154) unknown zygosity twins. The exact numbers of individuals included in specific analyses differ slightly from one another (see Table
3), due to missing data in certain A-TAC modules.
Discussion
This study mainly investigated to what extent psychopathic traits, alone and in conjunction with NDPs (i.e., ADHD and ASD), were associated with being screened positive for CD. A main finding was that the CPTI-SV GD dimension was significantly associated with CD. Moreover, the other two psychopathy dimensions (i.e., the CPTI-SV CU and INS subscales) failed to display any significant associations with screening positive for CD in the presence of co-occurring NDPs. This counters the results of many previous studies indicating an association between CU traits and CD.
As expected in a population-based non-clinical sample of youth, elevated psychopathic traits were rare, resulting in a markedly skewed distribution of scores. An overwhelming majority of the children displayed very low or zero scores for the three CPTI-SV subscales, although scores on the INS subscale were relatively higher. Unlike previous research finding an age-related increase in the GD dimension for both boys and girls (Colins et al.
2014), this study did not find this effect in the CPTI-SV dimensions. This might be because our study population was somewhat older than those in several previous studies, and that these personality traits are more developed and manifest when children reach the preadolescent phase.
We found statistically significant correlations between CD and NDPs (i.e., ADHD and ASD) as assessed with A-TAC. This was expected, given the large sample size as well as previous research findings indicating both comorbidity and symptom overlap between behavioral and neuropsychiatric conditions in youth (Blair et al.
2014; Hart et al.
1995; Rasmussen and Gillberg
2000). The finding of weak to moderately strong correlations between psychopathic traits, NDPs, and CD further supports the “ESSENCE” theory, that children with one type of neurodevelopmental problem often display other neurodevelopmental problems (Gillberg
2010). This indicates a need for preventative programs.
In our sample, there was a particularly strong correlation between the CPTI-SV INS subscale and ADHD, which may reflect content overlap (i.e., high similarity between individual items) rather than true comorbidity. That scores on both these measures were correlated with CD is unsurprising, given that impulsivity increases the risk of norm-breaking behavior and could be the common denominator between INS and ADHD. On the other hand, the weak correlation between the CPTI-SV CU subscale and ASD is surprising: we expected to find a stronger correlation, because previous research has found problems with emotional processing in ASD that are similar to CU traits (Rogers et al.
2006). The weak correlation between the CPTI-SV CU subscale and CD was also unexpected given prior research. Considering that previous research has found that CU traits specify a more severe form of CD (Edens et al.
2007; Frick et al.
2014; Frick and Nigg
2012; Frick and White
2008; Herpers et al.
2012; Viding and McCrory
2018), we expected to find a stronger correlation between the two. That the CPTI-SV GD subscale and not the INS subscale displayed the strongest correlation with CD was also surprising, given that impulsivity is a prominent feature of CD. These results add a new context to the common understanding of what aspects are most strongly related to a diagnosis of CD, and the nature of these relationships merits further study.
In previous studies, ADHD emerged as a precursor to CD (Hofvander et al.
2009). This was also found by the univariable part of this study, while in the multivariable analysis of the full model (i.e., all CPTI-SV subscales and the two A-TAC domains of ADHD and the three of ASD), only the concentration and attention part of ADHD was associated with CD. Executive functions important for attentional abilities have been found to be prominent features of ADHD in combination with CD, and also related to teacher-rated symptom severity in groups of students with ADHD only or ADHD combined with CD (Barnett et al.
2009). The association found here between the concentration and attention domain of ATAC and CD might be explained by this particularly salient attentional dysfunction found in individuals with ADHD and CD.
The multivariable analysis also suggested that, while the LPE specifier for CD in DSM-5 (largely corresponding to the CU subscale of CPTI-SV) might, as previous research has found, be useful for identifying an especially problematic subgroup of children with CD, a GD interpersonal style may be equally useful in identifying children at risk of developing CD. Indeed, it may be as strong a risk factor for CD as is ADHD. This would mean that parents should be observant of children who lie, manipulate, and exploit others significantly more than other children do, even if these children do not display other forms of antisocial behavior. The discrepancies between the present results and previous research could be because our study population was older, as older children generally tend to display higher levels of GD traits (Colins et al.
2014). It is also possible that these traits affect the development of CD at different ages, suggesting an even more complicated relationship between psychopathic traits and CD. For example, the presence of CU traits could be a stronger risk factor for CD at younger ages, or at least a more useful indicator because these traits stand out more clearly in young children. The GD interpersonal style, in contrast, might become a risk factor at a greater age, or may simply emerge later and therefore not be visible for clinical evaluation until the child has reached late childhood. Our results are also interesting with regard to the question that Salekin (
2016), among others, has asked: Is there a core dimension of psychopathic traits in children and, if so, what is it? Elevated CU traits may be a root of psychopathy and expected to precede other peripheral or secondary features, but our results highlight the importance of GD as another potential core feature of psychopathy. Clearly, more research is needed in this area, but judging from the current results, it seems that clinicians interested in detecting and preventing CD may need to assess different dimensions of childhood psychopathic traits depending on the age of the child undergoing assessment, making the task even more complex.
Strengths and Limitations
This study has several strengths, including a large twin dataset representative of the general population (Andrew et al.
2001) and the inclusion of a child psychopathy measure. However, three specific limitations affect this study. First, ASD, ADHD, and CD were derived from parent ratings alone. Not having access to either teacher reports or clinical diagnoses and professional ratings could affect the validity of the ratings. Second, the threshold for CD might be considered low. The criteria for CD (i.e., the dependent variable) are arguably over-inclusive, given that they are based on the endorsement of only two types of antisocial behavior; in contrast, clinical diagnosis requires the endorsement of at least three separate symptoms. On the other hand, previous research has demonstrated that the threshold used has high specificity (Kerekes et al.
2014), which would reduce this problem. Third, psychopathy was measured based on parent reports alone, restricting the validity of this variable. Parents might find it difficult to rate their own children; notably, a substantial part of the original CATSS population could not be included in this study because their parents had not completed the full CPTI-SV. Those completing CPTI-SV might be biased relative to those who did not, possibly reducing the generalizability of the results and weakening the conclusions.
Conclusions
The current results lend further weight to the view that professionals working with children should broaden their scope in assessing psychopathy traits in youth with CD. Specifically, our results indicate that when considering whether a child or adolescent is at risk of developing an antisocial behavioral disorder (i.e., CD), one should look beyond the DSM-5 focus on CU traits and also examine GD traits, that is, behaviors such as lying, cheating, and manipulating others. Moreover, all three psychopathy dimensions could be used as specifiers of CD in clinical assessment, as suggested by Salekin (
2016). Our results are in line with such a strategy, demonstrating that GD traits and not only CU traits are associated with CD, at least in a somewhat older study population. Finally, these results also support the notion that various symptoms of mental health problems tend to aggregate in affected individuals, likely due to common genetic factors. We therefore need to be more accurate in assessments identifying the whole spectrum of mental problems, and thus the mental health needs of assessed children, to prevent them from traveling further toward maladjustment, marginalization, and mental illness.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.