Examining the relationship among substance abuse, negative emotionality and impulsivity across subtypes of antisocial and psychopathic substance abusers

Abstract

Growing evidence suggests that individuals with Antisocial Personality Disorder (ASPD) can be categorized into theoretically meaningful subtypes. This study builds on earlier cluster-analytic research (Poythress et al., 2010) that identified four subtypes of ASPD in a large sample of prison inmates and offenders ordered into mandatory substance abuse treatment. These four subtypes (primary, secondary, and “fearful” psychopathic and non-psychopathic ASPD) differed in theoretically important ways on various criterion measures. Of those participants in substance abuse treatment (N = 571), we compare the four clusters, as well as non-ASPD substance abusers, in terms of (a) the severity of their self-reported alcohol and drug problems and (b) whether the severity of their substance abuse is predicted by similar etiologically important correlates (i.e., negative emotionality, impulsivity). There were modest subgroup differences in abuse, although as expected secondary psychopaths reported more severe misuse than primary psychopaths. Associations between impulsivity and negative emotionality and drug use for the total sample were in the expected direction, though relatively modest in magnitude. Unexpectedly, these associations were weaker among psychopathic subtypes relative to the non-psychopathic subgroups. These findings suggest that the etiology of drug use may differ across subgroups of chronically antisocial individuals.

Introduction

Psychopathic personality (psychopathy) has been the focus of substantial attention from researchers over the course of the past several decades. Most modern conceptualizations of psychopathy date to the work of Cleckley (1941/1982), whose well-known criteria reflect prominent interpersonal (e.g., deceitful and insincere; unreliable; superficial charm), behavioral (e.g., irresponsibility; sexually promiscuity; impulsivity), and affective (e.g., grandiosity; lack of remorse or shame; incapacity for love) features thought to define the core of the disorder. Cleckley's criteria substantially underpin most contemporary measures of psychopathy (Lilienfeld, 1998), even though the current version of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV-TR) diagnosis of Antisocial Personality Disorder (ASPD) places little weight on the affective features that Cleckley believed were central to the disorder.

Not surprisingly, multiple social and behavioral problems have been linked to psychopathy, such as substance abuse and other forms of antisocial behavior. Investigations (e.g., Peters et al., 1998, Smith and Newman, 1990) of criminal justice samples have reported that offenders in general and those who are highly psychopathic in particular frequently meet criteria for a diagnosis of substance abuse or dependence. The reasons for the reasonably strong relationship between antisocial and psychopathic traits and substance abuse are not entirely clear, although various explanations have been proposed in the literature. For example, both negative emotionality and impulsivity have been investigated as risk factors for substance abuse, and both traits are more prevalent among persons involved in the criminal justice system. Before considering the potential role that these constructs might play in regards to the relationship between antisocial and psychopathic traits and substance abuse, it is necessary to first address the heterogeneous nature of ASPD.

Although intended to identify a relatively homogeneous diagnostic entity, it has long been argued that the DSM criteria for ASPD delineate a heterogeneous collection of individuals (Lilienfeld, 1994). Although psychopathy has traditionally been conceptualized as a unitary construct, more recent research has similarly viewed this disorder as multifaceted (Skeem, Poythress, Edens, Lilienfeld, & Cale, 2003). Various parsings of items composing specific assessment measures, such as the Psychopathy Checklist-Revised (PCL-R; Hare, 2003), have delineated narrower trait dimensions, such as two (Harpur, Hare, & Hakstian, 1989), three (Cooke & Michie, 2001), and four (Hare, 2003) factor models of the 20 items comprising the PCL-R item pool. Expanding beyond factor analyses of measures such as the PCL-R, recent investigations have focused on identifying specific subtypes of psychopathy. The utility of a subtyping approach, as suggested by Hicks, Vaidyanathan, and Patrick (2010), is that “it identifies individuals high on these independent, latent psychopathy trait dimensions (as opposed to correlated, measured factor scores), clarifying the differential correlates without the need for statistical adjustments” (p. 39).

Much of the subtyping research on psychopathy in recent years has been informed by Karpman, 1941, Karpman, 1948a, Karpman, 1948b distinction between “primary” and “secondary” psychopathy. Karpman conceptualized primary psychopaths as lacking anxiety, empathy, and guilt, and as more deliberate in their actions than secondary psychopaths. He regarded the latter to be more vulnerable to negative emotions, such as anxiety, and to exhibit higher levels of impulsivity, hostility, and aggression. Etiologically, Karpman theorized that secondary psychopathy develops primarily as a result of environmental factors, such as parental abuse or rejection and/or unresolved emotional conflict, whereas primary psychopathy develops from a postulated “constitutional” deficit, characterized by a lack of conscience (Karpman, 1948a, pp. 478, 485). Recent cluster analytic studies of male prisoner samples and male jail inmates (Hicks et al., 2004, Poythress et al., 2010, Skeem et al., 2007; see also Swogger & Kosson, 2007, for similar findings) have yielded substantial support for psychopathic subtypes broadly consistent with these primary and secondary conceptualizations.

As arguably the most thorough and exhaustive empirical analysis of antisocial and psychopathic subtypes (as well as providing the basis for the current study), the findings of Poythress et al. (2010) warrant a brief review. Their sampling procedure included an assessment of over 1,400 offenders either receiving court-ordered substance abuse treatment or currently incarcerated in prisons across multiple jurisdictions throughout the U.S.—approximately half of whom met DSM-IV-TR criteria for APSD. Based on theory (e.g., Karpman, 1941, Karpman, 1948a, Karpman, 1948b, Lykken, 1995) and previous research, Poythress et al. expected the emergence of primary and secondary variants of psychopathy within the subsample of offenders who met ASPD criteria (those participants who did not meet ASPD criteria were not included in the cluster analysis). Given that many offenders who meet criteria for ASPD do not exhibit substantial psychopathic traits (Hare, 2003), Poythress et al. also hypothesized that they would identify an ASPD only group within their data set.

The clustering variables selected by Poythress et al. (2010) assessed the following constructs thought to be most theoretically relevant to parsing the heterogeneity of ASPD: psychopathic traits, fearless temperament, sensitivity to reward stimuli, assessment of early abuse, and trait anxiety. These constructs were operationalized via the Interpersonal, Affective, and Behavioral facets of the PCL-R, the Multidimensional Personality Questionnaire Harm Avoidance scale (MPQ-HA), Behavioral Activation System (BAS) subscales (RR, DR, Fun), the Childhood Abuse and Trauma Scale (CATS), and the Personality Assessment Inventory (PAI; Morey, 2007) Anxiety (ANX) scale, respectively.

Three of the four subtypes identified via model-based clustering corroborated distinctions among primary, secondary, and non-psychopathic ASPD offenders (see Fig. 1). The primary psychopathy profile was marked by somewhat higher PCL-R scores on the core interpersonal and affective features rather than on behavioral features, and the lowest scores on fearful temperament and anxiety. The secondary group exhibited the highest mean anxiety of all groups, and had the highest score on the CATS, indicating strong endorsement of early stressful family events. As expected, the non-psychopathic ASPD cluster group did not exhibit markedly elevated scores on any of the clustering variables. A fourth unanticipated group also emerged from the analyses. These individuals appeared relatively psychopathic in terms of PCL-R facets but displayed an ostensibly fearful temperament in regards to significantly elevated MPQ-HA scores, thus provisionally being labeled “fearful psychopaths.”

Examination of external correlates of these clusters provided preliminary evidence of their construct validity (Poythress et al., 2010). In particular, theoretically informative differences emerged between primary and secondary groups across multiple domains. In particular, relative to the primary psychopathic group, significantly higher scores were observed among the secondary psychopaths for self-reported impulsivity and internalizing symptoms of psychopathology (e.g., depression).

Notably, Poythress et al. did not report any analyses specifically related to potential cluster differences in alcohol or drug use or abuse. Drug and alcohol problems were, however, included as variables in a broader “externalizing behavior” construct. As predicted, the secondary psychopathic subgroup obtained significantly higher mean scores on this externalizing dimension than did the primary psychopathic cluster.

From the descriptions provided above, antisocial and psychopathic subtypes might be expected to differ both in terms of their level of substance abuse and the putative causal factors responsible for such abuse. For example, Lykken (1995) hypothesized that primary psychopaths are less prone to the use of alcohol and other sedating drugs than secondary psychopaths. Gudonis et al. (2009) similarly contended that negative emotionality plays a critical role in how psychopathic subtypes relate to substance use disorders, with a strong relationship between negative affect and substance abuse being particularly evident among secondary psychopaths relative to primary psychopaths. Somewhat consistent with these hypotheses, research suggests stronger associations between externalizing symptoms, including substance abuse, and features of secondary psychopathy than between externalizing symptoms and features of primary psychopathy (Patrick et al., 2005, Smith and Newman, 1990, Sylvers et al., in press). Similarly, Lynam, Whiteside, and Jones (1999) reported stronger associations between secondary psychopathy scores and substance use behaviors relative to primary psychopathy.

In addition to negative emotionality, impulsivity may play a causal role in substance abuse (Dawe, Guilo, & Loxton, 2004), particularly in offender populations. Although Cleckley did not consider primary psychopaths to be particularly impulsive per se, many contemporary models of ASPD and psychopathy (e.g., American Psychiatric Association, 1994, Hare, 2003, Lilienfeld and Widows, 2005) identify impulsivity as a prominent feature of the disorder. In terms of primary and secondary variants, Karpman (1948a) viewed secondary psychopaths as more impulsive relative to the more calculating primary psychopaths. In the Poythress et al. (2010) study noted earlier, the secondary psychopathy group, relative to the primary psychopathy group, reported significantly higher levels of impulsivity. Similarly, other investigations, such as Smith and Newman (1990), have demonstrated stronger relationships with substance abuse problems for individuals with features of secondary psychopathy rather than primary psychopathy.

In the present study we sought to expand upon the original cluster-level comparisons reported by Poythress et al. (2010) by focusing more specifically upon substance use and misuse among that subset of participants receiving court-mandated substance abuse treatment. Based on the preceding literature review we expected that the following findings would emerge. First, we hypothesized that secondary psychopaths would demonstrate higher levels of self-reported substance-related impairment than primary psychopaths. Second, in our sample at large, we predicted that the severity of substance abuse problems would be at least moderately associated with indicators of negative emotionality and impulsivity. Based on theory and prior subtyping research, however, we anticipated that these associations would be most pronounced among secondary psychopaths but non-existent (or at least weakest) among primary psychopaths. Analyses comparing the other subtypes, given the dearth of theoretical or empirical guidance to support specific predictions, were primarily exploratory in nature.