تکانشگری، پرخاشگری و اختلال شخصیتی DSM-IV

Abstract The aim of this study is to assess whether impulsive and aggressive traits can be placed on a continuum with DSM-IV Cluster B Personality Disorders (PDs) and to determine if different aspects of these personality traits are specifically associated with individual Cluster B PDs. The study group comprised 461 outpatients admitted consecutively to a clinic that specializes in the diagnosis and treatment of PDs. Principal component analyses clearly suggested a five-factor structure of both normal and psychopathological personality traits. Importantly, measures of impulsivity, aggressiveness and novelty seeking formed a part of the principal component that clustered all Cluster B PDs. Regression analyses indicated that impulsive traits were selectively associated with Borderline PD whereas different aspects of aggressiveness were useful in discriminating Narcissistic PD from Antisocial PD. Sensation seeking traits formed a part of Histrionic PD. These results indicate that impulsive/aggressive traits may be useful in explaining both why Cluster B PDs tend to covary, and why they frequently differ in clinical pictures and courses.

. Introduction During the last decade there has been an increasing interest in the role of impulsivity (Moeller et al., 2001) and aggressiveness (Coccaro, 1989 and Dolan et al., 2002) in psychiatric disorders with a particular focus on Personality Disorders (PDs). Although most research has considered impulsivity and aggressiveness as distinctive features of Borderline (BPD) (Links et al., 1999) and Antisocial (ASPD) PDs (Dolan et al., 2002, Barratt et al., 1997, Gardner et al., 1991, Lish et al., 1996 and Raine, 1993), recent studies have reported that impulsivity and aggressiveness may also characterize other Cluster B PDs (New et al., 2002). These findings suggest that impulsive and aggressive features may represent personality aspects relevant to several DSM-IV PDs that have been reported to show high rates of covariation ( Widiger et al., 1991, Ekselius et al., 1994 and Fossati et al., 2000). Thus, impulsive–aggressive features, as well as other action personality traits, such as Novelty Seeking ( Cloninger et al., 1993), could represent personality characteristics that underlie all Cluster B PDs and could explain their co variation. In other words, according to this hypothesis, Cluster B PDs could represent the maladaptive variant of extreme “action-oriented”, impulsive–aggressive personality dimensions. The need for a broader, multivariate perspective in the study of the relationship between DSM-IV PDs and personality traits is also suggested by the high co-occurrence rate, consistently reported in PD diagnoses (see, for instance, Widiger et al., 1991, Ekselius et al., 1994 and Widiger, 2005). This substantial co-occurrence in PD diagnoses seems to indicate an inadequacy in the DSM-IV categorical model of PD diagnoses and the need to explain these co-occurrences in terms of continuities both in PDs, and between PDs and adaptive personality traits ( Ball, 2001 and Widiger, 2005). The hypothesis of a continuum between impulsive–aggressive personality traits and Cluster B PDs raises several general questions, relevant to PD diagnoses and psychopathology. Firstly, this putative personality dimension should both converge with measures of Cluster B PDs, and diverge from measures of “feeling” personality traits, e.g., anxiety, and of other DSM-IV PDs. Moreover, the hypothesis of a continuum between adaptive and maladaptive personality traits raises questions as to the possible application of the Five Factor Model ( Widiger, 1991 and Widiger, 2005) for the description of personality psychopathology. Today, although specific models of the relationships between basic personality traits and selected PDs have been proposed and tested (Yeung et al., 1993 and Ball et al., 1997), few factor analytic studies have been carried out to directly attempt to identify common latent dimensions clustering measures of both adaptive and maladaptive personality traits. As a whole, these studies have suggested that Neuroticism, Introversion, low Agreeableness and Conscientiousness are related to dimensional representations of PDs (Livesley, unpublished manuscript, 1990; Schroeder et al., 1992). Despite their groundbreaking contribution, these factor analytic studies did not use DSM-based measures of PD features and relied only on self-report questionnaires. The evidence for latent dimensions connecting adaptive and maladaptive personality traits suggests the need for a new approach to the study of interpersonal characteristics — i.e., adult attachment — that are thought to play a role in the development of certain Cluster B PDs, namely Borderline (Fonagy et al., 1996 and Bateman and Fonagy, 2004) and Narcissistic PDs (Bateman and Fonagy, 2004). According to this view, adult attachment patterns should show continuities with both normal personality traits (e.g., impulsivity) and with their extreme personality variants (e.g., Borderline PD). A final relevant question concerning the adaptive–maladaptive interface in personality is the need to explain the clinical differences that occur in PDs that purportedly belong to the same latent dimension. For example, the Cluster B PDs of BPD and ASPD have frequently been reported to load on the same factor (Fossati et al., 2000 and Virkunnen et al., 1994). However, it is also well known that ASPD subjects frequently manifest difficulties in complying with social norms and may experience guilt feelings; indeed, these aspects make them quite different from BPD subjects who often display self-destructive behaviours and an erratic life style. Thus, in order to enhance our understanding of individual PDs we should change our focus from the exclusive study of broad personality domains to more specific facets of personality. This shift in approach suggests that concentrating exclusively on general domain level personality dimensions leads to a loss of relevant information and is inadequate for describing PDs, particularly in clinical samples (Butcher and Rouse, 1996 and De Clerq and DeFruyt, 2003). From this perspective, PDs often do not vary in terms of underlying dimensions and structures, but are different in terms of the specific expression of symptoms (De Clerq and DeFruyt, 2003). Starting from the above considerations, this study aimed to test the following hypotheses in an outpatient sample: 1. Do “action-oriented” personality traits, i.e., impulsivity, aggressiveness and novelty seeking, identify a latent dimension that is also relevant to Cluster B PDs but does not cluster other DSM-IV PDs? 2. Is it possible to differentiate this latent dimension from other dimensions, clustering other adaptive and maladaptive aspects of personality (e.g., “feeling” personality traits) and interpersonal relationships? 3. Is it appropriate to describe these converging/discriminating issues in the relationship between impulsivity, aggressiveness and DSM-IV PDs in the context of a five-factor model of personality and its psychopathology? 4. Within the same “action-oriented” latent dimension, is it possible to differentiate the individual Cluster B PDs in terms of their specific association with distinct combinations of impulsivity, aggressiveness and novelty seeking facets? In order to avoid the risk of imposing an a priori five-factor structure to the data, no measure based on the Big Five model of personality was used in this study. Instead we relied on Cloninger's seven-factor model of personality, both because of its comprehensive assessment of temperament and character features, and because of its well-known clinical usefulness in PD assessment (Cloninger et al., 1993). Furthermore, so as not to capitalize on method correlations, different methods for assessing adaptive and maladaptive personality features were used in the study. Since the existing literature indicates that categorical PDs show a high degree of diagnostic co-occurrence (Widiger, 2005), that many of the thresholds for categorical PD diagnoses are fairly arbitrary (Widiger, 1992) and that meaningful individual differences can be observed beyond the simple presence or absence of a categorical personality disorder diagnosis (Klein, 1993 and Widiger and Frances, 1985), dimensional (i.e., number of symptoms) DSM-IV PD diagnoses were retained for further analyses in this study.

. Results 3.1. Descriptive statistics Using SCID-II, 301 subjects (65.3%) received at least one DSM-IV PD diagnosis; the mean number of PD diagnoses was 1.02 (S.D. = 1.01), whereas the overall number of DSM-IV PD symptoms was on average 10 (S.D. = 5.5). Interestingly, subjects with no PD diagnoses showed a substantial, clinically relevant number of PD symptoms (mean = 7.00, S.D. = 3.5), although they did not meet the DSM-IV criteria for a specific PD diagnosis. Among subjects who received at least one DSM-IV PD diagnosis, the presence of multiple axis II diagnoses was frequent, since the average number of PD diagnoses was roughly 2 (S.D. = 0.9); the mean number of PD symptoms was 13 (S.D. = 4.9). The most frequently diagnosed DSM-IV PDs were Narcissistic (n = 74, 16.1%), Avoidant (n = 49, 10.6%), Passive–Aggressive (n = 69, 15%), Histrionic (n = 46, 10.0%), Obsessive–Compulsive (n = 45, 9.8%), Dependent (n = 44, 9.5%), Depressive (n = 37, 8.0%), Paranoid (n = 14, 3.0%), Schizotypal (n = 11, 2.4%), and Schizoid (n = 3, 0.7%) PDs. The base rates for DSM-IV BPD and ASPD diagnoses were 5.0% (n = 23) and 2.4% (n = 11), respectively. The correlation for the BIS-11, TCI-R, ASQ and AQ scales and the dimensionally assessed DSM-IV PDs in the full sample are listed in Table 1. 1 In agreement with previous observations ( Widiger, 2005) several significant correlations were observed among the DSM-IV PDs in this study. Interestingly, in this study not only BPD and ASPD, but also Passive–Aggressive and Narcissistic PDs, showed positive and significant correlation with BIS-11, AQ and Novelty Seeking total scores. In the opposite direction, Passive–Aggressive PD was observed to be negatively correlated with two character dimensions; Self-directedness and Cooperativeness. Furthermore, Obsessive–Compulsive PD showed moderate, negative correlations with the BIS-11 scale and the Novelty Seeking facet of TCI-R. Table 1. DSM-IV personality disorder, temperament and character inventory-revised, attachment style questionnaire, Barratt impulsiveness scale-11 and aggressiveness questionnaire scale correlations APD DPD OCPD PAPD DEPD PPD SZPD SPD HPD NPD BPD ASPD NS HA RD PE S C ST CO DIC RAS NFA PWR BIS-11 AQ APD 1.00 DPD 0.45 1.00 OCPD 0.38 0.27 1.00 PAPD − 0.23 − 0.09 − 0.23 1.00 DEPD 0.55 0.47 0.42 − 0.17 1.00 PPD − 0.01 − 0.09 − 0.01 0.05 − 0.09 1.00 SZPD 0.03 − 0.09 − 0.00 − 0.11 − 0.04 0.49 1.00 SPD 0.15 − 0.06 0.02 − 0.11 0.01 0.06 0.40 1.00 HPD − 0.29 − 0.07 − 0.28 0.25 − 0.24 0.01 − 0.13 − 0.10 1.00 NPD − 0.33 − 0.26 − 0.25 0.51 − 0.34 0.03 − 0.12 − 0.12 0.43 1.00 BPD − 0.16 − 0.04 − 0.14 0.24 − 0.12 0.06 − 0.07 − 0.09 0.13 0.15 1.00 ASPD − 0.01 − 0.08 − 0.10 .017 − .011 0.01 − 0.04 − 0.05 0.07 0.12 0.21 1.00 NS − 0.27 − 0.10 − 0.34 0.19 − .0.21 0.02 − 0.07 − 0.04 0.25 0.27 0.23 0.17 1.00 HA − 0.28 0.32 0.09 0.13 0.27 0.11 0.09 − 0.03 − 0.07 0.01 0.09 − 0.00 − 0.20 1.00 RD − 0.07 0.19 0.01 − 0.02 0.05 0.12 − 0.11 − 0.16 0.17 0.03 0.04 0.03 0.16 0.03 1.00 PE − 0.13 − 0.16 0.02 − 0.15 − 0.13 0.07 − 0.05 − 0.01 0.06 0.04 0.02 0.04 − 0.08 − 0.43 0.12 1.00 S − 0.12 − 0.26 0.05 − 0.24 − 0.14 0.10 − 0.06 0.07 − 0.12 0.17 0.21 − 0.03 − 0.17 − 0.56 0.01 0.34 1.00 C 0.04 0.00 0.06 − 0.25 0.03 0.16 − 0.09 − 0.02 − 0.07 0.30 0.11 − 0.08 − 0.13 − 0.13 0.31 0.24 0.43 1.00 ST − 0.10 − 0.04 − 0.04 − 0.06 − 0.09 0.03 0.16 0.06 0.07 0.04 0.04 0.05 0.10 − 0.15 0.15 0.34 0.06 0.08 1.00 CO − 0.28 − 0.19 − 0.12 − 0.10 − 0.18 0.09 − 0.06 − 0.02 0.06 0.04 0.14 0.06 0.14 − 0.47 0.34 0.22 0.31 0.21 0.15 1.00 DIC 0.16 − 0.02 0.10 0.14 0.05 0.13 0.12 0.08 − 0.10 0.01 0.01 0.00 − 0.16 0.17 − 0.42 0.06 − 0.15 − 0.17 0.03 − 0.46 1.00 RAS 0.10 − 0.02 0.01 0.00 − 0.08 0.13 0.11 0.09 − 0.05 0.06 0.05 0.09 − 0.09 0.01 − 0.39 0.05 − 0.20 − 0.33 0.06 − 0.19 0.24 1.00 NFA 0.33 0.32 0.12 0.07 0.27 0.04 0.07 − 0.09 0.00 0.05 0.15 − 0.02 − 0.13 0.52 0.00 − 0.16 − 0.57 − 0.13 0.04 − 0.51 0.33 0.25 1.00 PWR 0.15 0.25 0.07 0.12 0.20 0.11 0.07 − 0.08 0.03 0.08 0.19 0.00 − 0.01 0.38 0.16 − 0.04 − 0.41 − 0.22 0.10 − 0.33 0.21 0.05 0.59 1.00 BIS-11 − 0.11 0.09 − 0.24 0.27 − 0.09 0.06 0.01 − 0.03 0.16 0.21 0.30 0.12 0.52 0.18 − 0.06 − 0.33 − 0.41 − 0.28 0.07 − 0.15 0.06 0.11 0.16 0.21 1.00 AQ − 0.15 − 0.07 − 0.11 0.34 − 0.08 0.15 0.06 − 0.04 0.01 0.27 0.16 0.18 0.24 0.20 − 0.07 − 0.06 − 0.38 − 0.46 0.20 − 0.21 0.25 0.16 0.25 0.36 0.35 1.00 Note. The DSM-IV Personality Disorders are listed in SCID-II order; Depressive and Passive–Aggressive (Negativistic) Personality Disorders were included in the DSM-IV for further study. PD = Personality Disorder. Correlation (Pearson r) coefficients greater than 0.18 in absolute value are significant at Bonferroni-corrected nominal significance level (i.e., P < 0.00015). APD: Avoidant PD; DPD: Dependent PD; OCPD: Obsessive–Compulsive PD; PAPD: Passive–Aggressive PD; DEPD: Depressive PD; PPD: Paranoid PD; SZPD: Schizotypal PD; SPD: Schizoid PD; HPD: Histrionic PD; NPD: Narcissistic PD; BPD: Borderline PD; ASPD: Antisocial PD; NS: Novelty Seeking; HA: Harm Avoidance; RD: Reward Dependence; PE: Persistence; S: Self-directedness; C: Cooperativeness; ST: Self-transcendence; CO: Confidence; DIC: Discomfort with Closeness; RAS: Relationships as Secondary; PWR: Preoccupation with Relationships. BIS-11: Barratt Impulsiveness Scale-11 total score; AQ: Aggressiveness Questionnaire total score. Table options As shown in Table 1, all Cluster C PDs were significantly negatively correlated with Confidence and Novelty Seeking. In contrast, Cluster C PDs correlated positively with Harm Avoidance and two variables of adult attachment style; Need For Approval and Preoccupation With Relationships. Interestingly, the same pattern of correlation was also observed for Depressive PD. As a whole, these correlation coefficients were in the small-to-moderate range; these findings were somewhat expected, since fallible measurements based on different methods, i.e., self-report questionnaires and semi-structured interviews, were used to assess personality features and PD characteristics, respectively. In other words, these correlations were weakened by the effect of measurement errors, but were not inflated by capitalizing on using the same method of assessment; thus, in a sense, they represented a lower-bound estimate of the “true” associations between PDs and personality dimensions. 3.2. Factor analysis results Although the first eight eigenvalues were greater than 1.00, the scree plot clearly suggested a five-factor solution. As shown in Table 2, quasi-inferential parallel analysis clearly supported a five-factor solution; indeed, only the first five real data eigenvalues exceeded 95% of the corresponding random data eigenvalues. Thus, only the first five principal components were retained for further analyses. Table 2. Quasi-inferential parallel analysis results Order Observed eigenvalue Random data null quantiles 50% 75% 90% 95% 97.5% 99% 99.5% 99.9% 1 4.32 1.46 1.48 1.51 1.53 1.54 1.56 1.58 1.59 2 3.76 1.39 1.41 1.43 1.44 1.45 1.47 1.48 1.52 3 2.31 1.34 1.35 1.37 1.38 1.39 1.40 1.41 1.52 4 1.67 1.29 1.31 1.32 1.33 1.33 1.34 1.35 1.39 5 1.54 1.25 1.26 1.28 1.28 1.29 1.30 1.30 1.31 6 1.21 1.21 1.22 1.24 1.24 1.25 1.25 1.26 1.27 7 1.05 1.18 1.19 1.20 1.21 1.21 1.22 1.22 1.23 Table options Promax and Varimax rotations yielded matching solutions, with factor score correlations ranging from .99 to 1.00. Interestingly, even in raising elements of the target matrix to the fourth power, Promax rotation yielded factors that were virtually orthogonal (median factor inter-correlation = 0.06; min. r = − 0.17; max r = 0.19). Thus, only Varimax-rotated factor loadings are listed in Table 3. Table 3. Factor analysis results: varimax-rotated factor loadings (n = 641) Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 h2 Avoidant personality disorder 0.450 − 0.582 0.083 0.003 − 0.130 0.565 Dependent personality disorder 0.595 − 0.311 − 0.208 − 0.154 − 0.103 0.528 Obsessive–compulsive personality disorder 0.235 − 0.562 0.094 − 0.119 0.169 0.422 Passive–aggressive personality disorder 0.115 0.600 0.148 − 0.166 − 0.130 0.440 Depressive personality disorder 0.503 − 0.513 − 0.095 − 0.110 − 0.101 0.547 Paranoid personality disorder 0.054 0.058 0.154 0.639 0.008 0.439 Schizotypal personality disorder 0.043 − 0.082 0.068 0.847 0.08 0.738 Schizoid personality disorder − 0.104 − 0.180 0.069 0.564 − 0.033 0.368 Histrionic personality disorder − 0.026 0.524 − 0.117 − 0.226 0.100 0.350 Narcissistic personality disorder − 0.109 0.671 0.185 − 0.183 0.05 0.533 Borderline personality disorder 0.203 0.431 0.013 − 0.111 0.05 0.242 Antisocial personality disorder 0.001 0.307 0.025 − 0.020 0.09 0.104 Novelty seeking (TCI-R) − 0.025 0.631 − 0.298 0.104 − 0.087 0.506 Harm avoidance (TCI-R) 0.695 − 0.040 0.096 0.078 − 0.315 0.599 Reward Dependence (TCI-R) 0.210 0.076 − 0.786 − 0.118 0.271 0.755 Persistence (TCI-R) − 0.276 − 0.097 0.053 − 0.154 0.810 0.768 Self-directness (TCI-R) − 0.720 − 0.345 − 0.142 − 0.101 0.162 0.694 Cooperativeness (TCI-R) − 0.187 − 0.370 − 0.522 − 0.123 0.245 0.520 Self-transcendence (TCI-R) 0.085 0.158 − 0.111 0.255 0.697 0.596 Confidence (ASQ) − 0.506 0.050 − 0.545 0.018 0.199 0.595 Discomfort with closeness (ASQ) 0.212 − 0.272 0.684 0.059 0.167 0.549 Relationships as secondary (ASQ) 0.074 0.043 0.602 0.136 0.09 0.395 Need for approval (ASQ) 0.783 − 0.012 0.245 − 0.033 0.09 0.683 Preoccupation with relationships (ASQ) 0.732 0.183 0.081 0.028 0.234 0.631 BIS-11 0.300 0.581 − 0.017 0.184 − 0.262 0.531 AQ 0.341 0.519 0.290 0.150 0.160 0.518 Note. h2 = Communalities; The DSM-IV Personality Disorders are listed in SCID-II order; Depressive and Passive–Aggressive (Negativistic) Personality Disorders were included in the DSM-IV for further study. TCI-R: Temperament and Character Inventory-Revised; ASQ: Attachment Style Questionnaire; BIS-11: Barratt Impulsiveness Scale-11; AQ: Aggressiveness Questionnaire. Table options The five-factor solution explained 52.4% of the variance of the observed variables, and was significantly replicated across sub-samples based on gender (median congruence coefficient = 0.94) and axis I diagnosis (median congruence coefficient = 0.95), respectively. Re-analyzing the data after dropping the two PDs that were included in the DSM-IV for further studies, i.e., Depressive and Passive–Aggressive (Negativistic) PDs, did not alter the factor solution (factor score correlations: min r = 0.98, max. r = 0.99). The factor solution remained stable even when the data was re-analyzed after deleting one variable at a time (median factor score correlation = 0.99). Importantly, all Cluster B PDs showed their highest loadings on Factor 2, which was also characterized by substantial positive loadings of Novelty Seeking, BIS-11, and AQ scales and by negative loadings of Obsessive Compulsive, Depressive, and Avoidant PDs. Factor 2 defined a latent dimension that was characterized by inhibition and excessive control at one extreme and by impulsive behaviour at the other extreme; this would be consistent with the Big Five personality dimension of Conscientiousness (with reversed sign). NS, BIS-11, and AQ scales also loaded on Factor 2. Factor 1 was characterized by high loadings of attachment, personality and personality disorder variables; in particular, Avoidant PD, Dependent PD, Depressive PD, Harm Avoidance, Need for Approval, and Preoccupation with Relationships were positively related to Factor 1, whereas Self-Directedness and Confidence (in self and others) were negatively related to this factor. Interestingly, high scores on Preoccupation with Relationships and Need for Approval and low scores on Confidence characterize the so-called anxious attachment pattern. As a whole, Factor 1 clustered variables related to negative emotionality (anxious and depressive personality traits), as well as to insecurity in self and others; this factor closely resembles the Big Five dimension of Neuroticism. Factor 3 was not characterized by any substantial PD loadings; rather, it clustered Discomfort with Closeness and Relationships as Secondary (to achievement) at the positive pole, and Reward Dependence, Cooperativeness, and Confidence at the negative pole. Thus, Factor 3 identified a personality dimension that was characterized by lack of empathy, hostility, intolerance, avoidance of intimacy and self-absorption and was highly consistent with the Big Five dimension of (low) Agreeableness. Factor 4 grouped the DSM-IV Cluster A PDs. These PDs share common characteristics of social isolation, eccentricity, and withdrawal from relationships; these characteristics suggest that Cluster A PDs may represent extreme variants of the Big Five dimension of (low) Extraversion. Finally, Factor 5 was characterized by positive, substantial loadings of Persistence and Self-transcendence. Subjects high on this factor could be described as creative, selfless, fulfilled, persistent and stable, despite experiencing frustration and fatigue, which are in turn considered as a personal challenge. Factor 5 was somewhat similar to the Big Five Openness to Experience dimension. An interesting finding emerged when we tried to evaluate the replicability of this five-factor structure in the subgroups, composed by subjects with any PD diagnosis (n = 301) and without PD diagnosis (n = 160), respectively. Only Factor 1 (congruence coefficient = 0.91) and Factor 3 (congruence coefficient = 0.95) emerged as matching factors across the two subgroups. Congruence coefficient values for Factor 2 and Factor 5 were 0.71 and 0.79, respectively, indicating that the structures identified in the two sub-groups showed some similarities, but were far from being replicated. Finally, Factor 4 was not replicated (congruence coefficient = 0.17). 3.3. Regression analysis results The results of the regression analyses of Cluster B PDs on the individual facets of Novelty Seeking, BIS-11 and AQ scales are listed in Table 4; Passive Aggressive (Negativistic) PD was also included in these analyses, since it loaded on the same factor as Cluster B PDs. As a whole, only a few significant effects of axis I disorders on the dependent variables were observed. Namely, a Substance Abuse Disorder diagnosis was a significant predictor of Histrionic (β = − 0.13, t = − 3.02, P < 0.005), Narcissistic (β = 0.10, t = 2.67, P < 0.01) and Antisocial PDs (β = 0.21, t = 4.55, P < 0.001), whereas a Mood Disorder diagnosis was negatively associated with Narcissistic PD (β = − 0.08, t = − 2.11, P < 0.05) and positively linked to Passive–Aggressive PD (β = 0.09, t = − 2.41, P < 0.05). Table 4. Regression analysis results Dependent variable HPD NPD BPD ASPD PAPD Predictors β β β β β Exploratory excitability (NS1) 0.15a Impulsiveness (NS2) Extravagance (NS3) 0.20a 0.16 0.18 Disorderliness (NS4) 0.19b Attention (BIS-11) 0.19c Motor impulsiveness (BIS-11) 0.18c 0.13 Non-planning impulsiveness (BIS-11) Physical aggressiveness (AQ) 0.13d Verbal aggressiveness (AQ) Anger (AQ) 0.21b 0.18 Hostility (AQ) 0.19e a Significant when controlling for the effect of NPD, BPD, ASPD, PAPD, any axis I diagnosis, and individual DSM-IV axis I disorders. b Significant when controlling for the effect of HPD, BPD, ASPD and PAPD, any axis I diagnosis, and individual DSM-IV axis I disorders. c Significant when controlling for the effect of HPD, NPD, ASPD, and PAPD, any axis I diagnosis, and individual DSM-IV axis I disorders. d Significant when controlling for the effect of HPD, NPD, BPD, and PAPD, any axis I diagnosis, and individual DSM-IV axis I disorders. e Significant when controlling for the effect of HPD, NPD, BPD, and ASPD, any axis I diagnosis, and individual DSM-IV axis I disorders. Table options Although Histrionic, Narcissistic, Borderline, Antisocial, and Passive–Aggressive PDs appeared as extreme variants of a common latent personality dimension (closely akin to [low] Conscientiousness in the Big Five), the different facets of the temperament variables that characterized the same dimension showed distinct relationships with each PD. For instance, as a whole NS was related to Narcissistic and Histrionic, but did not markedly characterize Borderline and Antisocial PDs. In particular, Exploratory Excitability and Extravagance features of Novelty Seeking indicated Histrionic PD features, after removing the effect of co variation with other Cluster B PDs; interestingly, Narcissistic PD was characterized not only by Novelty Seeking features, namely, Disorderliness, but also by anger, an emotional component of aggression. Although both Anger and Hostility were associated with Passive–Aggressive (Negativistic) PD, only the latter remained a significant indicator of Passive–Aggressive PD, once the effect of the potential confounders was checked for. Both Motor and Attention Impulsiveness were uniquely associated with Borderline PD, whereas only Physical Aggression was a significant predictor of Antisocial PD features, once the effect of the other PDs involved in the regression analyses and axis I disorders were held constant.