ارتباط عاطفی، شناختی و فیزیولوژیکی استرس ناشی از سوء استفاده در اختلال شخصیت مرزی و ضد اجتماعی

Abstract Childhood abuse is an important precursor of borderline personality disorder (BPD) and antisocial personality disorder (ASPD). The current study compared the emotional reactivity to abuse-related stress of these patients on a direct and an indirect level. Changes in self-reported affect and schema modes, psychophysiology and reaction time based cognitive associations were assessed following confrontation with an abuse-related film fragment in patients with BPD (n = 45), ASPD (n = 21), Cluster C personality disorder (n = 46) and non-patient controls (n = 36). Results indicated a hyperresponsivity of BPD-patients on self-reported negative affect and schema modes, on some psychophysiological indices and on implicit cognitive associations. The ASPD-group was comparable to the BPD group on implicit cognitions but did not show self-reported and physiological hyper-reactivity. These findings suggest that BPD and ASPD-patients are alike in their implicit cognitive abuse-related stress reactivity, but can be differentiated in their self-reported and physiological response patterns.

Introduction One of the main common etiological precursors of borderline personality disorder (BPD) and antisocial personality disorder (ASPD) is childhood trauma. Both BPD and ASPD have been linked to a broad range of abusive events but sexual and emotional abuse seem to stand out in BPD-patient's history (Battle et al., 2004, Johnson et al., 1999 and Lobbestael, Bernstein, in press) and physical abuse in ASPD-patients (Bernstein et al., 1998, Bierer et al., 2003, Lobbestael, Bernstein, in press, Lobbestael, Harkema, et al., 2009 and Lobbestael, Lobbes, et al., in press). Although quite a lot is known about maltreatment precursors of BPD and ASPD, the influence of abuse-related stressful stimuli on these patients has hardly been studied. Schmahl, Elzinga, et al. (2004) compared the physiological reactivity to abandonment and abuse-related scripts of BPD-patients with post-traumatic stress disorder patients but did not found any difference between these groups. Brain imaging studies indicated that over- versus under-reactivity in different areas of dorsolateral and medial prefrontal cortex correlated to traumatic memory in BPD-patients (Schmahl, Vermetten et al., 2004), while abandonment memories caused a greater increase in bilateral dorsolateral prefrontal cortex activity in women with BPD compared to women without BPD (Schmahl et al., 2003). To our best knowledge, the impact of abuse-related stress on ASPD-patients has not been studied yet. In fact, little is known about the biological, emotional and cognitive sequelae of BPD and ASPD in general, including how these disorders differ from each other and from other forms of psychopathology. The current study tests whether salient and prolonged reminders of mixed childhood abuse are responsible for emotional, psychophysiological, and cognitive dysregulation in BPD and ASPD. Comparing abuse-related reactivity between BPD and ASPD-patients could contribute to the understanding of the differences between these two closely related personality disorders. Furthermore, the current study can contribute to the ongoing debate on whether BPD is characterized by a BPD-theme specific (i.e., abuse) emotional hyperresponsity (Arntz, 2005). In addition, the use of prolonged presented abuse-related stimuli more closely resembles real-life encounters and therefore can provide a more realistic insight into emotional reactivity. The aim of this study was to investigate the effects of abuse-related stimuli in BPD and ASPD-patients, both at a direct and an indirect level. On a direct level, the effect of abuse-related stress was evaluated by means of self-reported affect and schema modes. Schema modes originate from Schema Therapy (Young, Klosko, & Weishaar, 2003) and represent the momentary emotional and cognitive states and coping responses that are active at a given point in time. Modes can be seen as the different aspects of one's personality, and each individual has several modes. Schema modes can be triggered by emotional events and an individual may shift from one schema mode into another. This way, the schema mode concept describes the rapid shifting in emotion and behaviour demonstrated by patients suffering from severe personality disorders. Modes can be maladaptive or adaptive. The more severe the personality pathology, the stronger the presence of the maladaptive modes and the higher the number of modes a person has and the more mode switches occur. The maladaptive modes are divided into four categories; dysfunctional child modes that result out of unmet core childhood needs; dysfunctional coping modes that correspondent to an overuse of the fight, flight or freeze coping styles and dysfunctional parent modes that reflect negative behaviour of the patient's parent(s) towards the patient as a child that the patient has internalized. The adaptive modes reflect healthy thoughts, feelings and behaviours. A self-report study has demonstrated BPD to be characterized by 9 maladaptive modes, while ASPD was linked to the Enraged Child and the Bully and Attack modes (Lobbestael, van Vreeswijk, & Arntz, 2008). But, assessment by therapists suggests that more schema modes are prominent in ASPD-patients than they report themselves (Lobbestael, Arntz, Löbbes, et al., in press). Since affect assessment by means of self-report is vulnerable to bias due to lack of self-knowledge or avoidance of negative thoughts and feelings (Wilson & Dunn, 2004), and lying and conning are central diagnostic features of ASPD (APA, 2005), the use of indirect measures is advisable. Therefore, the current study incorporated two types of indirect measures to assess the impact of abuse-related stimuli; psychophysiological indices and a reaction time based paradigm to measure the implicit association between the self and abuse. The latter paradigm was operationalized with the Single Category Implicit Association Test (SC-IAT, Karpinski & Steinman, 2006), a variant of the classical Implicit Association Task (IAT, Greenwald, McGhee, & Schwartz, 1998) in which associations with single targets like the self-concept can be measured without the need for an opposite category. In sum, the central question of this study was: Do BPD and ASPD-patients differ in their reaction to abuse-related stimuli on self-reported affect, self-reported schema modes, psychophysiological reactivity, and on implicit abuse-related self-image? We expected BPD-patients to display a stronger intensity of affective experiences in reaction to the abuse-related stimuli relative to control patients and non-patients. Due to similarities between BPD and ASPD (American Psychiatric Association, 2005 and Paris, 1997) and the common etiological influence of abuse, it was hypothesized that BPD and ASPD-patients would show a similar response pattern to abuse-related stimuli on levels the person cannot easily control. Since antisocials tend to underreport the impact of negative events, it was expected that the ASPD-group would indicate a lower increase in self-reported indices of negative affect and schema modes than the BPD group. In contrast, we expected a similar abuse-related reactivity in BPD and ASPD-patients at an indirect level (i.e., the physiological indices and the implicit association task). Additionally, this study assessed the severity of childhood trauma in BPD and ASPD-patients and tested the effect of trauma severity on the changes in the direct or indirect abuse-related reactivity. Finally, the level of psychopathy was assessed in the ASPD-group and the predictive value of psychopathy on abuse-related stress was tested. We compared, according to DSM-IV (APA, 2000), a BPD group, a ASPD-group (with varying degrees of psychopathy), and a Cluster C PD (ClC-PD) patient group, next to a non-patient control group (NpC). By using NpCs as a control group, it can be tested whether results indicate a general Axis II response pattern, while inclusion of ClC-PD patients permits drawing PD cluster-specific conclusions.

Results Self-reported negative affect At baseline, there was no main effect of gender on the SRNA, p = .40. The patient groups indicated a higher baseline level of negative affect than the NpC group while that of the ClC-PD and BPD groups were higher than that of the ASPD-group ( Table 2 and Table 3). Results on the change scores of the SRNA revealed no significant main effect of gender, p = .10, or of age, p = .26. There was a significant change in all groups of the SRNA indicating that negative affect increased significantly after stress induction ( Table 4). The effect sizes were large for BPD and ClC-PD groups and medium for ASPD and NpC groups ( Table 4). Group difference analyses revealed that the BPD group displayed a stronger increase in SRNA than the ASPD and the NpC groups ( Table 5). Table 2. Medians, quartiles (25th percentile and 75th percentile) of the baseline levels of all dependent variables. Dependent variables BPD ASPD ClC-PD NpC Median 25per 75per Median 25per 75per Median 25per 75per Median 25per 75per SRNAa 41.00 21.00 54.50 18.00 12.00 27.50 43.50 22.50 64.25 9.00 6.00 14.00 Malb 45.49 39.44 52.01 27.36 17.40 35.03 43.61 34.15 51.97 14.29 8.72 21.07 Adapc 49.33 40.38 62.79 65.00 54.92 78.25 47.33 36.17 61.88 83.83 77.83 92.83 HRd 78.76 69.89 84.36 73.96 67.20 82.76 76.86 71.21 84.50 74.65 67.30 82.64 SBPe 113.75 108.38 119.88 118.88 111.63 129.06 119.50 109.17 134.38 121.25 115.75 131.50 DBPf 76.00 73.50 81.75 77.00 62.38 83.50 80.33 73.88 88.00 79.75 73.50 85.50 SCRg 1.90 .90 4.77 7.71 3.70 10.42 3.88 .50 6.21 4.71 2.72 7.52 SCLh 2.51 1.54 4.55 3.06 1.67 4.37 2.83 1.80 3.86 3.65 2.23 5.30 Frowning 19.42 10.79 554.59 9.75 5.82 33.83 28.79 11.23 80.99 19.27 10.96 131.03 SC-IATi 80.92 1.43 187.13 185.32 −7.66 255.16 30.69 −34.55 72.32 72.05 30.18 161.83 a self-reported negative affect. b maladaptive schema modes. c adaptive schema modes. d heart rate. e systolic blood pressure. f diastolic blood pressure. g skin conductance response. h skin conductance level. i Single-category implicit association task. Table options Table 3. Group contrasts (Wald values) of all baseline scores of all groups. Dependent variables BPD vs ASPD BPD vs ClC-PD BPD vs NpC ASPD vs ClC-PD ASPD vs NpC ClC-PD vs NpC SRNAa 10.71** .70 49.24** −15.11** 12.75** 57.22** Malb 23.53** −.92 66.82** −18.12** 12.85** 59.24** Adapc 7.69* −.72 45.74** −11.63** 13.67** 53.76** HRd 1.14 .00 1.46 −1.22 .00 1.59 SBPe −2.89 −1.56 −6.23* −.53 −.16 −1.88 DBPf .28 −1.31 −.60 −1.99 −1.24 −.09 SCRg −13.18** −1.95 −7.95* 6.76* 1.23 −2.62 SCLh .12 .71 −3.84 −1.06 −3.87 −1.54 Frowning 8.12* .10 .56 −7.12* −4.43* .23 SC-IATi −1.15 6.09* −.02 9.02* .82 −5.91* *p < .05; **p < .001. a self-reported negative affect. b maladaptive schema modes. c adaptive schema modes. d heart rate. e systolic blood pressure. f diastolic blood pressure. g skin conductance response. h skin conductance level. i Single-category implicit association task. Table options Table 4. Medians, quartiles (25th percentile, and 75th percentile), and effect sizes of the change scores of all dependent variables for all groups. Dependent variables BPD ASPD ClC-PD NpC Change score Change score Change score Change score Medianj 25per 75per ESk Medianj 25per 75per ESk Medianj 25per 75per ESk Medianj 25per 75per ESk SRNAa 17.50 8.00 35.00 .88* 5.00 −.50 27.50 .51* 14.00 2.50 25.00 .80** 5.00 1.00 18.00 .68** Malb 3.38 −.41 9.46 .47* −.17 −4.98 5.62 .00 .45 −1.99 2.73 .08 −.33 −1.64 2.46 .00 Adapc −4.08 −13.88 4.33 .34* −2.50 −16.17 6.08 .20 −1.50 −7.42 3.25 .27 −.67 −7.83 3.33 .20 HRd −1.99 −4.42 −.70 .95* −3.72 −6.10 −1.60 1.21* −2.28 −4.00 −2.28 .89* −3.34 −6.67 −1.89 1.13** SBPe 1.25 −1.62 6.63 .34* −1.88 −5.63 4.69 .09 2.25 −.75 6.88 .54** −.75 −4.00 4.00 .03 DBPf 2.50 −.13 4.46 .62* .87 −2.13 2.75 .22 2.00 .13 5.00 .61** 1.50 −.50 4.75 .44* SCRg .63 −.20 1.83 .56* −.77 −2.12 1.10 .40 2.25 −.75 6.88 .12 −.19 −1.77 1.08 .18 SCLh .21 −.07 .69 .28* .01 −.19 .44 .14 .12 −.12 .36 .32* .04 −.49 .32 .03 Frowning −8.31 −37.14 8.71 .39* 10.34 −10.42 51.91 .36 −3.31 −29.43 31.17 .06 9.26 −24.31 94.70 .22 SC-IATi −106.19 −262.55 60.67 .42* −35.34 −142.04 28.91 .39* 13.51 −88.52 81.77 .05 −11.61 −91.72 76.70 .08 *p < .05; **p < .001. The p-levels are associated with the robust pre-post changes of the within-sample tests. a self-reported negative affect. b maladaptive schema modes. c adaptive schema modes. d heart rate. e systolic blood pressure. f diastolic blood pressure. g skin conductance response. h skin conductance level. i Single-category implicit association task. j positive scores indicate increases in dependent variables, and negative scores indicate decreases in dependent variables. k robust standard deviation (√sw2)/.642 (;sw2 = winsorized variance). Table options Table 5. Group contrasts (Wald scores) on the change scores of all dependent variables. Dependent variables BPD vs ASPD BPD vs ClC-PD BPD vs NpC ASPD vs ClC-PD ASPD vs Nap ClC-PD vs NpC Wald p Wald p Wald p Wald p Wald p Wald p SRNAa 5.03* .03 2.48 .12 9.27* .002 −1.02 .31 .14 .71 2.67 .10 Malb 5.18* .02 4.44* .04 4.48* .03 −.38 .54 .21 .65 .03 .87 Adapc .10 .75 −.62 .43 −.57 .45 −.10 .76 −.10 .75 .00 .99 HRd 3.13 .08 −.10 .76 6.69* .01 −2.30 .13 .19 .66 5.46* .02 SBPe 2.78 .10 −.35 .56 3.70* .05 −4.40* .04 −.002 .96 6.09* .01 DBPf 1.95 .16 −.006 .94 .43 .51 −2.12 .15 −.67 .41 .53 .47 SCRg 7.68* .006 1.42 .23 5.69* .02 −3.50 .06 .42 .52 1.82 .18 SCLh .80 .37 −.22 .64 2.90 .09 −.28 .60 .35 .55 1.70 .19 Frowning −4.39* .04 −.93 .34 −5.95* .02 1.83 .18 .00 .99 −2.50 .11 SC-IATi .65 .42 7.21* .007 4.72* .03 1.88 .17 1.06 .30 −.12 .73 *p < .05; **p < .001. a self-reported negative affect. b maladaptive schema modes. c adaptive schema modes. d heart rate. e systolic blood pressure. f diastolic blood pressure. g skin conductance response. h skin conductance level. i Single-category implicit association task. Table options Schema modes At baseline, there was no main effect of gender on the level of the adaptive or maladaptive modes, p's > .27. The patient groups indicated a higher baseline level of maladaptive modes than the NpC group, and BPD and ClC-PD higher than the ASPD-group. The opposite pattern was found for the adaptive modes ( Table 2 and Table 3). Results on the change scores of the maladaptive and adaptive modes revealed no significant main effect of gender, p's > .32, or of age, p's > .31. There was a significant increase of the maladaptive modes, and a significant decrease of adaptive modes in the BPD group ( Table 4). Group differences showed that the BPD group indicated a stronger increase in maladaptive modes than the other groups. There were no group differences regarding the change in adaptive modes ( Table 5). Physiological measures At baseline, there was no main effect of gender on DBP, SCR, or SCL p's > .06. There was a main effect of gender on HR, SBP and frowning which were higher in men, Wald's > 7.33, p's < .004. There were no group differences in baseline HR, DBP and SCL levels. The BPD group displayed a lower baseline level of SBP compared to the NpC group, and a lower SCR baseline level than the ASPD and NpC groups. The ASPD-group had a higher baseline SCR level compared to the ClC-PD group. The baseline frowning level of the ASPD-group was lower compared to those of the other groups ( Table 2 and Table 3). Results on the change scores of the physiological indices revealed no significant main effect of gender, p's > .12. The change scores did not reveal a significant effect of age for DBP, SCR, SCL and frowning, p's > .10, while age did significantly influence the change scores of HR, p = .004, and SBP, p < .001. HR significantly decreased in all groups (all large effects). SBP significantly increased in the BPD and ClC-PD groups (small and medium effects, respectively), and DBP significantly increased in the BPD and ClC-PD groups (medium effects) and in the NpC group (small effect). SCR significantly increased in the BPD group (medium effect), while frowning significantly decreased in the BPD group (small effect). Finally, SCL significantly increased in the BPD and ClC-PD groups (small effects, Table 4). Both BPD and ClC-PD groups showed a smaller decrease in HR compared to the NpC group. The BPD group showed a stronger increase in SBP compared to the NpC group, while the ClC-PD group showed a higher increase in SBP compared to the ASPD and NpC groups. There were no group differences in the change scores of DBP and SCL. The BPD group showed a stronger increase in SCR compared to the ASPD and the NpC groups. The ASPD and ClC-PD groups displayed an increase in frowning activity and the BPD group a decrease, which significantly differed from each other ( Table 5). SC-IAT At baseline, there was a main effect of gender on the SC-IAT, Wald = 11.64, p = .001, indicating that men had a stronger SC-IAT effect (i.e. association between self and abuse) at baseline than women. BPD, ASPD and NpC groups had a higher baseline SC-IAT score than the ClC-PD group ( Table 2 and Table 3). The effect size indicated a significant change of the SC-IAT effect from the baseline to the stress induction phase in the BPD and ASPD-groups, both small effects ( Table 4). Results on the change scores of the SC-IAT revealed no significant main effects of gender, p = .07, or age, p = .71. The BPD-group displayed a stronger change towards a self-abuse association than the ClC-PD and the NpC groups ( Table 5). Childhood trauma There was no main effect of gender on the severity of abuse, Wald = 3.05, p = .08. The patient groups had higher childhood abusive scores than the NpC group, and the BPD group had higher scores than the ClC-PD group (see Table 6). Table 6. Medians, quartiles (25th percentile, and 75th percentile), and contrasts between the groups of childhood trauma. Total trauma BPD ASPD ClC-PD NpC Median 44.42 24.15 29.69 4.25 25th percentile 27.28 12.26 15.43 1.28 75th percentile 69.04 54.34 53.42 13.42 BPD vs ASPD BPD vs ClC BPD vs NpC ASPD vs ClC ASPD vsNpC ClC vs NpC Wald 3.06 4.42* 44.55** −.01 17.75** 26.78** p .08 .04 <.001 .92 <.001 <.001 Note: *p < .05; **p < .001. Table options Influence of trauma history on abuse-related reactivity The severity of childhood trauma was significantly positively correlated with the changes on SRNA, Spearman's rho = .29, p < .001, the maladaptive and the adaptive modes, Spearman's rho = .27 and −.20, p = .002 and .02 respectively, SBP, Spearman's rho = .22, p = .009 and DBP, Spearman's rho = .19, p = .03. Influence of psychopathy Attempts to predict abuse-related reactivity from the PCL-r total, factor 1 or 2 or facet 1–4 did not reveal any significant results.