Personality Disorders: Theory, Research, and Treatment

© 2014 American Psychological Association

2014, Vol. 5, No. 4, 428 – 433   1949-2715/14/$12.00    http://dx.doi.org/10.1037/per0000065

Diana Diamond, City University of New York and Weill Cornell Medical College

Kenneth N. Levy, Pennsylvania State University and Weill Cornell Medical College

John F. Clarkin, Weill Cornell Medical College

Melitta Fischer-Kern, Medical University of Vienna

Nicole M. Cain, Long Island University –  Brooklyn

Stephan Doering,  Medical University of Vienna

Susanne Hörz, Ludwig-Maximilians-Universitaet Muenchen

Anna Buchheim,  University of Innsbruck

 

We investigated attachment representations and the capacity for mentalization in a sample of adult female borderline patients with and without comorbid narcissistic personality disorder (NPD). Participants were 22 borderline patients diagnosed with comorbid NPD (NPD/BPD) and 129 BPD patients without NPD (BPD) from 2 randomized clinical trials. Attachment and mentalization were assessed on the Adult Attachment Interview (AAI; George, Kaplan, & Main, 1996). Results showed that as expected, compared with the BPD group, the NPD/BPD group was significantly more likely to be categorized as either dismissing or cannot classify on the AAI, whereas the BPD group was more likely to be classified as either preoccupied or unresolved for loss and abuse than was the NPD/BPD group. Both groups of patients scored low on mentalizing, and there were no significant differences between the groups, indicating that both NPD/BPD and BPD individuals showed deficits in this capacity. The clinical implications of the group differences in AAI classification are discussed with a focus on how under- standing the attachment representations of NPD/BPD patients helps to illuminate their complex, con- tradictory mental states.

Clinicians report that patients with comorbid narcissistic (NPD) and borderline personality disorders (BPD) are the most challeng- ing to treat of patients in the personality disorder spectrum because they tend to provoke or alienate the therapist through their attempts to direct and devalue the relationship with the therapist and the treatment (Kernberg, 2007; Ronningstam, 2010). Several studies have shown particularly high rates of comorbidity of NPD with

BPD, with rates ranging from 17% to 80% (Levy et al., 2007). In the recent wave 2 NESARC Study, 38.9% of those with BPD had a comorbid NPD diagnosis (Stinson et al., 2008). A number of studies have found that NPD patients are among the highest risk for drop out from psychotherapy (Ellison, Levy, Cain, Ansell, & Pincus, 2013), with rates up to 64% (Hilsenroth, Holdwick, Castle- bury, & Blais, 1998). Grandiosity in particular predicted engagement in treatment (e.g., less service utilization; Ellison et al., 2013). In addition, compared with BPD patients without NPD, those with NPD/BPD show different patterns of Axis II comor- bidity, fewer comorbid Axis I diagnoses, and fewer hospitaliza- tions and inpatient days (Diamond et al., 2012; Hörz et al., 2014). To further understand the characteristics of patients with NPD/ BPD compared with those with BPD, we examined attachment and mentalization (through the assessment of reflective function [RF]; Fonagy, Steele, Steele & Target, 1998) in these two groups using data from two recent randomized clinical trials on changes in attachment, RF, and symptomatology in psychotherapy for BPD patients. Attachment and RF are theorized to be an important aspect in personality disorders, especially BPD and NPD (Fonagy, Gergely, Jurist & Target, 2002). In this study, attachment and RF were assessed from the Adult Attachment Interview (AAI; George, Kaplan & Main, 1996), a semistructured interview that provides a window into individuals’ current representational states with re- spect to early attachment relationships and experiences, and their associated modes of defense and affect regulation. In the AAI five-way classification system (Main & Goldwyn, 1998), the or- ganized categories are characterized by a consistent attachment strategy: Secure or valuing/autnomous; dismissing or devaluing/ idealizing; preoccupied or angrily or passively conflicted states of mind with respect to attachment relationships. By contrast, the two disorganized categories, cannot classify and unresolved, show lack of integration in the attachment system, evident in the cannot classify category by oscillations between multiple opposing strat- egies throughout the interview (most typically between dismissing and preoccupied strategies), or in the unresolved category by brief but drastic lapses in the monitoring of discourse and reasoning in response to specific questions about loss and abuse (Hesse, 2010). Previous research suggests that insecure and/or disorganized internal working models of attachment and deficits in RF underlie the difficulties with self-regulation and interpersonal functioning that make NPD/BPD patients so challenging to treat (Fonagy et al., 2002; Levy et al., 2007). BPD has been linked primarily with preoccupied and/or unresolved attachment representations on the AAI (George et al., 1996), whereas NPD has been associated primarily with dismissing and secondarily with preoccupied at- tachment representations (Bakermans-Kranenburg & van IJzen- doorn, 2009; Fonagy et al., 1996; Levy et al., 2006; Rosenstein & Horowitz, 1996). Thus, studies have shown that NPD has been associated with diverse and contradictory attachment representa- tions (e.g., dismissing and preoccupied attachment), which have been linked respectively to the NPD mental states of grandiosity and vulnerability (Cain et al., 2008; Meyers & Pilkonis, 2011). These findings, along with clinical observations that have linked NPD not only with grandiosity in which others are systematically devalued but also with oscillations between grandiosity and vul- nerability (Kernberg, 2007; Levy et al., 2007), suggest a lack of integration in attachment representations indicating that further research on the attachment correlates of NPD is warranted. Inter- estingly, previous investigations (Levy et al., 2006) suggested that the dismissing classification was associated with the cannot clas- sify category on the AAI, and the preoccupied classification with the unresolved category, which called for further investigation of these groups. Hence, in investigating the differences in attachment status across the NPD/BPD and BPD groups, we combined the dismissing and cannot classify categories, and the preoccupied and

unresolved categories. We expected that compared with the BPD group the NPD/BPD group would be more likely to be classified either as dismissing or cannot classify, whereas the BPD group would be more likely to be classified either as preoccupied or unresolved on the AAI than the NPD/BPD group.

Related to insecure/disorganized attachment representational states are deficits in mentalization assessed in this study through the RF scale (Fonagy et al., 1998), an operationalized measure of the capacity to mentalize in attachment relationships. Mentalizing, also known as RF, is a social– cognitive capacity to reflect on one’s own and others mental states in the context of attachment experi- ences. In previous research, deficits in RF have been linked to the development of insecure and/or disorganized working models of attachment, whereas high RF has been found to be a buffer against the transgenerational transmission of attachment insecurity and possibly against the development of BPD in individuals who have experienced childhood trauma and abuse (Fonagy et al., 2002). Because in previous research NPD has been associated with lack of empathy and disinterest in understanding others, and BPD with deficits in reflective capacity linked to lack of resolution of trauma (Fonagy et al., 1998), we expected that both the NPD/BPD and BPD individuals would be characterized by low or deficient RF. Both the dismissing/cannot classify and preoccupied/unresolved attachment groups were expected to show low scores in RF. To date there have been no studies on attachment representations or RF of NPD/BPD patients and thus this study addresses a gap in the literature.

 

Footnote:

Diana Diamond, Department of Psychology, City University of New York and Department of Psychiatry, Weill Cornell Medical College; Kenneth N. Levy, Department of Psychology, Pennsylvania State Uni- versity and Department of Psychiatry, Weill Cornell Medical College; John F. Clarkin, Department of Psychiatry, Weill Cornell Medical College; Melitta Fischer-Kern, Department of Psychoanalysis and Psychotherapy, Medical University of Vienna; Nicole M. Cain, Depart- ment of Psychology, Long Island University – Brooklyn; Stephan  Doering, Department of Psychoanalysis and Psychotherapy, Medical University of Vienna; Susanne Hörz, Department of Psychology, Clin-  ical    Psychology,    and    Psychotherapy, Ludwig-Maximilians-Universi-

taet; Anna  Buchheim,  Institute  of  Psychology,  University  of  Innsbruck.

This research was supported by grants from the National Institute of Mental Health (MH012530; PI: Levy), International Psychoanalytic Association (PI: Levy), the Kohler Fund of Munich Borderline Person- ality Disorder Research Foundation (BPDRF; PI: Kernberg), and grants from the Jubiläumsfonds of the Austrian National Bank (10636; PI: Doering) and the German Psychoanalytic Society (DPG; PI: Doering). Correspondence concerning this article should be addressed to Diana Diamond,  50  Riverside  Drive,  Apartment  6A,  New  York,  NY 10024.

E-mail: ddiamonda@gmail.com

 

Method

 Procedures

Combined data from two recently completed randomized con- trolled trials of transference-focused psychotherapy (TFP) for women with borderline personality disorder (Clarkin, Levy, Len- zenweger, & Kernberg, 2007; Doering et al., 2010) are included in this report. Only the procedures and measures relevant to the current study are described below. A full description of the method and findings from the two randomized clinical trials can be found in Clarkin et al. (2007); Doering et al. (2010); Fischer-Kern et al. (2010); and Levy et al.  (2006).

 

Participants

A total of 151 outpatients were included in this study. Sixty participants were drawn from the Cornell–New York (NY) study (Clarkin et al., 2007; Levy et al., 2006) and 91 participants were drawn from the Vienna–Munich (V-M) study (Doering et al., 2010). All participants from both studies were clinically referred. Inclusion and exclusion criteria between the studies were similar, with the exception of the upper age range included, how each group handled the inclusion of mood disorders, and that the V-M sample excluded those with antisocial personality disorder.

Combined demographics are as  follows:  100%  women  ages 18 –50 (M = 28.90; SD = 7.45), 57% were single, 38.4% were in a relationship, and 4.6% were divorced/separated. In addition, 53.0% of participants were employed, 23.8% were enrolled in school or a training program, and 23.2% had no employment.

All referrals in both samples were assessed for axis I and II diagnoses, including Diagnostic and Statistical Manual of Mental Disorders, fourth edition/fifth edition (DSM–IV/5) criteria for BPD using reliably administered structured interviews. However, the Cornell-NY group employed the International Personality Disor- der Examination (IPDE; Loranger, 1999) as their structured diag- nostic interview for axis II and the V-M group used the Structured Clinical Interview for DSM–IV (SCID)-II. Although different structured interviews were used between the samples and the convergence between the two interviews is only fair in general, the convergence is good to strong for the diagnosis of NPD (Widiger & Lowe, 2010). Of the 151 participants, 22 met criteria for NPD/BPD (14.6% of the sample) whereas 129 met criteria for BPD. t tests and chi-square analyses revealed that the two samples did not differ in the total number of lifetime and current axis I disorders and axis II disorders, GAF scores, or the number of patients with NPD.

 

Measures

International Personality Disorder Examination (IPDE; Lo- ranger, 1999). The IPDE is a highly reliable and valid semi- structured interview based on the DSM–IV criterion for diagnosing personality disorders (Loranger, 1999). Raters were trained clini- cians with extensive experience in the diagnosis of severe Axis II pathology and trained directly by the developer of the measure (A. Loranger). Reliability was established before beginning the study and monitored throughout. See Critchfield, Levy, and Clarkin (2007) for more detail regarding diagnostic interviewers, inter- viewer credentials and training, and reliability procedures.

The Structured Clinical Interview for DSM–IV (SCID-I/ SCID-II, German version, Wittchen et al., 1997).  The SCID is a highly reliable and valid semistructured clinical interview ad- ministered by trained clinicians and designed to yield psychiatric diagnoses consistent with DSM–IV/DSM–IV–TR (American Psy- chiatric Association, 2000) diagnostic criteria. In the V-M study, four research assistants, who were masked for the therapy deliv- ered, conducted assessments before randomization and after one year of treatment. See Doering et al. (2010) for more detail regarding diagnostic procedures, interviewer training, and reliabil- ity procedures.

Adult Attachment Interview (AAI; George et  al.,  1996). The AAI is a 20-question semistructured clinical interview asked in set order with standard probes, intended to elicit thoughts, feelings, and memories about early attachment relationships, in- cluding experiences of separation, rejection, loss, and abuse. The interview allows numerous opportunities for the interviewee to elaborate on, contradict, or fail to support previous statements. The AAI scoring system is designed to quantify the individual’s current state of mind with respect to attachment relationships with signif- icant caregivers.

The AAI was transcribed verbatim. Trained coders  blind  to time and treatment condition scored the transcripts with  sub- scale ratings, with specific patterns used to assign the individual  to one of five primary attachment classifications: Secure/auton- omous (F), dismissive (Ds), preoccupied  (E),  unresolved  (U), and cannot classify (CC) described earlier. If the interview  is given a primary disorganized classification, then it is also  assigned to a secondary organized category (e.g., F, Ds, or E).

In this study only the primary AAI classification was used. Previous research has shown remarkable stability and predictive validity of the AAI (see Hesse, 2010 for a review). Addition-   ally, the distribution of the AAI classifications has been found     in several meta-analyses to be remarkably similar across West-  ern industrialized countries and relatively independent of cross- cultural variation (Bakermans-Kranenburg & van IJzendoorn, 2009). As described earlier, in our subsequent analyses com- paring attachment classification of NPD/BPD and BPD groups,  we combined the dismissing and cannot classify categories, and the preoccupied and unresolved categories in line with our hypotheses.

In both samples, the AAI was administered and scored  by raters who have completed a two-week training workshop con- ducted by Mary Main and Erik Hesse, and had achieved reli- ability on an extensive set of training transcripts. Raters were blind to all identifying characteristics of the participants, in- cluding attachment status and the nature and purpose of  the  study. In the Cornell-NY sample, after training was completed   and reliability was established, the coders coded  a  subset  of  each other’s transcripts (n = 22). Raters agreed on 86% of the  categorical classifications (k = .80, t = 6.11,  p  <  .001). Similarly, in the V-M sample, two judges  coded  a  subset  of each other’s transcripts (n = 36). Raters agreed on 89% of the  categorical  classifications  (k  = .84,  t  = 8.254,  p  <  .001).

Reflective Functioning Scale (RF; Fonagy et al., 1998). The AAI was also scored by trained blind coders with the RF scale,   an 11-point scale that assesses individual differences in the capacity to mentalize in the context of attachment relationships. The RF scale ranges from -1 (negative RF in which interviews are antireflective, totally barren of mentalization, or grossly distorting of the mental states of others) to 9 (exceptional RF in which interviews show unusually complex, elaborate, or origi-   nal reasoning about mental states). A global RF score is given based on questions of the AAI that probe for reflection. Coders  for the Cornell-NY sample were trained by the second author (K.N.L.), who had received training from the developers of the coding manual (P. Fonagy and M. Target). Reliability was obtained between the coders and one of the developers of the coding manual on practice sets.  After  training  was  completed and reliability was established, the two coders coded a subset of each other’s transcripts (n = 28, ICC = 0.86). As with the AAI coding, coders were blind to both time and treatment condition (see Levy et al., 2006). All RF coders for the V-M sample were trained by the developers of the manual (P. Fonagy) and un- derwent a reliability test as part of the training. The interrater reliability calculation for the current study was based on 20 transcripts  (n  = 20;  ICC = 0.79).

In a larger study of multiple samples, Taubner et al. (2013) showed good interrater reliability for  the  global  RF  score,  which is relatively stable across time and is significantly lower    in  individuals  with  severe  psychopathology.

 

Procedures

In the current study we included only those participants who received the AAI (George et al., 1996), and all patient data used in this study were obtained at the beginning of treatment. The com- bined sample of 151 patients was divided into two groups: one including 129 participants who met criteria for BPD, and another that included 22 who met criteria for both BPD and NPD accord- ing to the DSM–IV– based criteria of the IPDE and the SCID-II.

 

Statistical Procedures

To examine the association between categorical attachment classification for the NPD/BPD and BPD groups, we employed chi-square tests of association. Fisher’s exact test p values were calculated because of the relatively small cell sizes. Cramer’s V is reported as an estimate of effect size/strength  of  association  (0.10 = small, 0.30 = medium, 0.50 = large). We also conducted independent samples t tests to examine mean differences between the NPD/BPD and BPD groups on RF. Cohen’s d is reported for all t tests as an estimate of effect size (0.10 = small, 0.30 = medium,  0.50  = large).

 

Results

All AAI interviews were rated at pretreatment, using the five- way AAI classification system that includes organized [Secure (F), Dismissing (Ds), Preoccupied (E)], as well as disorganized [Un- resolved (U), Cannot Classify (CC)] categories. Given our hypoth- esis that the NPD/BPD patients would be more likely to be classified in the dismissing or cannot classify attachment catego- ries and that the BPD patients would be more likely to be classified in the preoccupied or unresolved attachment categories, we col- lapsed the attachment classifications into dismissing plus cannot classify versus all other AAI classifications for the NPD/BPD group, and unresolved plus preoccupied versus all other classifi- cations for the BPD group. Figure 1 shows the percentage of patients with NPD/BPD and BPD, respectively, classified in the dismissing/cannot classify and unresolved/preoccupied attachment groups. As displayed in Figure 1 and as expected, we found that the NPD/BPD patients were nearly twice as likely as the BPD patients to receive a primary AAI classification of dismissing or cannot classify, that is, 55% versus 30% (x2(1) = 5.34, p = .028). Also as expected, we found that the BPD diagnostic group was significantly  more  likely  to  have  unresolved  and  preoccupied

attachment classifications than was the NPD/BPD group, that is, 65% versus 36% (x2(1) = 6.53, p = .017). Combining the attach- ment classifications served two purposes in that (a) it allowed us to create clinically meaningful groups that preserved our original hypotheses about the essential attachment representations of the two groups under investigation, while (b) allowing for a more effective test of those hypotheses using the larger samples sizes and increased power afforded by such aggregation.

We also investigated differences between the two groups on RF. Mean comparisons of RF scores indicated no significant differ- ences between the NPD/BPD and BPD groups, both of which were in the low range as expected, M = 2.52 (1.38), and 2.85 (1.12), respectively,  t(149)  =  -1.21;  p  =  .228;  d  =  .26.  In  previous studies a score of 3 is considered to be indicative of low RF involving naïve/simplistic or overanalytic reasoning about mental states (Fonagy et al., 1998).

 

Discussion

Our results highlight the fruitfulness of using an attachment framework, and the AAI classification system in particular to illuminate the characteristics of attachment representations of in- dividuals with personality disorders. As expected, we found that NPD/BPD individuals were more likely to be categorized as either dismissing or as cannot classify (oscillating throughout the inter- view among several opposing attachment strategies) compared with the BPD individuals. On the other hand, compared with the NPD/BPD group, the BPD group was more likely to be classified as preoccupied, that is, angrily or passively entangled in past attachment relationships, or as unresolved with focal but drastic collapse in the monitoring of discourse or reasoning on specific questions related to childhood loss and abuse, involving “entrance into peculiar compartmentalized or even partially dissociated states of mind” (Hesse, 2010, p. 570). The low RF ratings in both groups, evidence of metacognitive deficits involving impairments in the capacity to reason about one’s own or others’ behavior in terms of mental states, suggests that both groups may show equal impairments in the capacity to mentalize in the face of attachment trauma that have been found in multiple studies to be a prevalent feature of the BPD patients’ history (Buchheim & George, 2011). However, the high rates of unresolved plus preoccupied AAI classification in the BPD group (65% for BPD vs. 36% for NPD/ BPD) suggest that further research is warranted to explore whether in the context of BPD, NPD patients may be better defended and/or more resilient, particularly in the face of loss and trauma. Further research should also investigate whether the two groups differ in type of self-reported trauma.

Although these findings suggest that insecure and/or disor- ganized attachment representations may be risk factors for the  development of different types of affective and personality disorders (Buchheim & George, 2011), they do not  suggest  causal influences between different types of insecure/disorga- nized attachment and the two clinical groups  under  investiga- tion. Indeed, our findings simply affirm that there is an asso- ciation between dismissing  and  cannot  classify  attachment states of mind in NPD/BPD patients and preoccupied and unresolved attachment states of mind for BPD patients, which should be investigated further in future studies with larger  samples  and  more  complex  longitudinal designs.

Our findings have important clinical implications because the AAI classifications provide indices of not only security/insecurity, but also of modes of discourse, defense, and affect regulation that help illuminate the complex, contradictory clinical presentation of NPD/BPD patients. The typical narcissistic defenses and transfer- ences, including the pervasive devaluation or brittle idealization of others, omnipotence and derogation of attachment related experi- ences and feelings (Clemence, Perry, & Plakun, 2009), and rigid and/or sparse discourse, may be understood as dismissing at- tachment mechanisms in which the focus is “continuously away from past attachment relationships and their influences” (Hesse, 2010, p. 556). Indeed the majority of AAI classifications within the group of NPD/BPD individuals (55%) were likely  to  be  either dismissing (32%) or cannot classify (23%). This suggests that although dismissing mechanisms may predominate in some patients, in others they may alternate  with  preoccupied  strate- gies in which the focus is “persistently although confusedly so strongly oriented toward attachment relationships and experi- ences”  (Hesse,  2010,  p. 556).

Our findings on the attachment and RF correlates of NPD/BPD and BPD groups should be considered in the light of the study’s limitations. First, this study was based on a sample of female borderline patients with and without comorbid NPD, so that find- ings may not be applicable to male NPD patients with or without  a comorbid BPD diagnosis. In addition, there is no NPD-only comparison group, so that it is difficult to determine what findings are unique to the comorbidity and what are driven by NPD alone. Previous studies indicating the high rates of comorbidity of NPD in particular suggest that an NPD only group might be difficult to identify (Levy et al., 2007). Another limitation of the study is that we assessed NPD using measures that are based on the DSM–IV/5 criteria, which is more heavily weighted toward assessing grandi- ose as opposed to vulnerable presentations of NPD (Ronningstam, 2010). However, a recent study using daily diary cards indicated that pathological narcissism predicted fluctuations within the in- dividual of grandiose and vulnerable narcissistic strategies to nav- igate social interaction over the course of a week, and that these fluctuations were related to the perceived communal (friendly) responses of the other during the social interaction (Roche, Pincus,

Conroy, Hyde, & Ram, 2013). Hence it is possible that those identified by the grandiose criteria privileged in the DSM–IV/5 are likely to have vulnerable manifestations of the disorder as well because NPD individuals are highly reactive to social interaction. Finally, the unequal numbers in the NPD/BPD and BPD groups and small sample size limit the study’s power to detect differences. Despite these limitations, our findings contribute to understand- ing the different representational processes, modes of affect regu- lation and defense that characterize the diverse, contradictory attachment strategies associated with NPD/BPD patients and help to illuminate the shifts in mental states between dismissing and preoccupied, grandiose and vulnerable as the attachment system is activated in the clinical  situation.

 

Figure  1.   Number  and  percentage  of  patients  with  NPD/BPD  and  BPD,  respectively,  classified  in  the

Dismissing/Cannot Classify and Unresolved/Preoccupied Attachment groups.

 

References

American Psychiatric Association. (2000). Diagnostic and statistical man- ual of mental disorders (fourth edition, text revision). Washington, DC: Author.

Bakermans-Kranenburg, M. J., & van IJzendoorn, M. (2009). The first 10,000 Adult Attachment Interviews: Distributions of adult attachment representations in clinical and non-clinical groups. Attachment & Human Development, 11, 223–263. doi:10.1080/14616730902814762

Buchheim, A., & George, C. (2011). Attachment disorganization in bor- derline personality disorder and anxiety disorder. In J. Solomon & C. George (Eds.), Disorganized attachment and caregiving (pp. 343–382). New York, NY: The Guilford  Press.

Cain, N. M., Pincus, A. L., & Ansell, E. B. (2008). Narcissism at the crossroads: Phenotypic description of pathological narcissism across clinical theory, social/personality psychology, and psychiatric diagnosis. Clinical Psychology Review, 28, 638 – 656. doi:10.1016/j.cpr.2007.09.006

Clarkin, J. F., Levy, K. N., Lenzenweger, M. F., & Kernberg, O. F. (2007). Evaluating three treatments for borderline personality disorder: A mul- tiwave study. The American Journal of Psychiatry, 164, 922–928. doi: 10.1176/appi.ajp.164.6.922

Clemence, A., Perry, J., & Plakun, E. M. (2009). Narcissistic and border- line personality disorders in a sample of treatment refractory patients. Psychiatric Annals, 39, 175–184.  doi:10.3928/00485713-20090401-05

Critchfield, K. L., Levy, K. N., & Clarkin, J. F. (2007). The personality disorders institute/borderline personality disorder research foundation randomized control trial for borderline personality disorder: Reliability of Axis I and II diagnoses. Psychiatric Quarterly, 78, 15–24. doi: 10.1007/s11126-006-9023-x

Diamond, D., Clarkin, J., Horz, S., Levy, K., Fisher-Kern, M., Cain, N.,… Buchheim, A. (2012, September). Attachment and reflective function in patients with co-morbid NPD/BPD. Paper presented at the 2nd Interna- tional Congress on Borderline Personality Disorder and Allied disorders, Amsterdam, The Netherlands.

Doering, S., Hörz, S., Rentrop, M., Fischer-Kern, M., Schuster, P., Be- necke, C., . . . Buchheim, P. (2010). Transference-focused psychother- apy v. treatment by community psychotherapists for borderline person- ality disorder: Randomised controlled trial. The British Journal of Psychiatry, 196, 389 –395. doi:10.1192/bjp.bp.109.070177

Ellison, W. D., Levy, K. N., Cain, N. M., Ansell, E., & Pincus, A. L. (2013). The impact of pathological narcissism on psychotherapy utili- zation, initial symptom severity, and early-treatment symptom change: A naturalistic investigation. Journal of Personality Assessment, 95, 291–300. doi:10.1080/00223891.2012.742904

Fischer-Kern, M., Buchheim, A., Hörz, S., Schuster, P., Doering, S., Kapusta, N. D., . . . Fonagy, P. (2010). The relation of personality organization,  reflective  functioning,  and  psychiatric  classification  in borderline personality disorder. Psychoanalytic Psychology, 27, 395– 409. doi:10.1037/a0020862

Fonagy, P., Gergely, G., Jurist, E. L., & Target, M. (2002). Affect regula- tion, mentalization, and the development of the self. New York, NY: Other Press.

Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., & Gerber, A. (1996). The relation of attachment status, psychiatric classi- fication, and response to psychotherapy. Journal of Consulting and Clinical Psychology, 64, 22–31.  doi:10.1037/0022-006X.64.1.22

Fonagy, P., Steele, M., Steele, H., & Target, M. (1998). Reflective-function manual: Version 5.0 for application to the adult attachment interview. Unpublished manual. University College, London.

George, C., Kaplan, N., & Main, M. (1996). The Berkeley Adult Attach- ment Interview. Unpublished manuscript. Department of Psychology, University of California, Berkeley.

Hesse, E. (2010). The Adult Attachment Interview: Protocol, method of analysis, and empirical studies. In J. Cassidy & P. Shaver (Eds.), Handbook of attachment: Theory, research and clinical applications (pp. 552–598). New York, NY: The Guilford Press.

Hilsenroth, M. J., Holdwick, D. J., Castlebury, F. D., & Blais, M. A. (1998). The effects of DSM–IV cluster B personality disorder symptoms on the termination and continuation of psychotherapy. Psychotherapy: Theory, Research, Practice, Training, 35, 163–176. doi:10.1037/ h0087845

Hörz, S., Diamond, D., Clarkin, J. F., Levy, K. N., Rentrop, M., Fischer- Kern, M., . . . Doering, S. (2014). Co-morbid narcissistic personality disorder in patients with borderline personality disorder. Manuscript submitted for publication.

Kernberg, O. F. (2007). The almost untreatable narcissistic patient. Journal of the American Psychoanalytic Association, 55, 503–539. doi:10.1177/ 00030651070550020701

Levy, K. N., Meehan, K. B., Kelly, K. M., Reynoso, J. S., Weber, M., Clarkin, J. F., & Kernberg, O. F. (2006). Change in attachment patterns and reflective function in a randomized control trial of transference- focused psychotherapy for borderline personality disorder. Journal of Consulting and Clinical Psychology, 74, 1027–1040. doi:10.1037/0022- 006X.74.6.1027

Levy, K. N., Reynoso, J. S., Wasserman, R. H., & Clarkin, J. F. (2007).

Narcissistic personality disorder. In W. T. O’Donohue, K. A. Fowler, & O. Lilienfeld (Eds.), Personality disorders: Toward the DSM-V (pp. 233–278). Thousand Oaks, CA: Sage. doi:10.4135/9781483328980.n9

Loranger, A. W. (1999). International Personality Disorder Examination (IPDE): DSM–IV and ICD-10 modules. Odessa, FL: Psychological Assessment Resources.

Main, M., & Goldwyn, R. (1998). Adult attachment scoring and classifi- cations system. Unpublished scoring manual. Department of Psychol- ogy, University of California,  Berkeley.

Meyers, B., & Pilkonis, P. A. (2011). Attachment theory and narcissistic personality disorder. In K. Campbell & J. Miller (Eds.), The handbook of narcissism and narcissistic personality disorder: Theoretical approaches, em- pirical findings,  and treatment  (pp. 434–444). New York, NY: Wiley.

Roche, M. J., Pincus, A. L., Conroy, D. E., Hyde, A. L., & Ram, N. (2013). Pathological narcissism and interpersonal behavior in daily life. Person- ality Disorders: Theory, Research, and Treatment, 4, 315–323. doi:10.1037/a0030798

Ronningstam, E. (2010). Narcissistic personality disorder: A current re- view. Current Psychiatry Reports, 12, 68 –75. doi:10.1007/s11920-009- 0084-z

Rosenstein, D. S., & Horowitz, H. A. (1996). Adolescent attachment and psychopathology. Journal of Consulting and Clinical Psychology, 64, 244 –253. doi:10.1037/0022-006X.64.2.244

Stinson, F. S., Dawson, D. A., Goldstein, R. B., Chou, S. P., Huang, B., Smith, S., . . . Grant, B. F. (2008). Prevalence, correlates, disability and comorbidity of DSM–IV narcissistic personality disorder: Results from Wave 2 National Epidemiologic Survey on Alcohol and Related Con- ditions. Journal of Clinical Psychiatry, 69, 1033–1045. doi:10.4088/JCP.v69n0701

Taubner, S., Horz, S., Fischer-Kern, M., Doering, S., Buchheim, A., & Zimmerman, J. (2013). Internal structure of the reflective functioning scale. Psychological Assessment, 25,  127–135.

Widiger, T. A., & Lowe, J. R. (2010). Personality disorders. In M. Antony & D. Barlow (Eds.), Handbook of assessment and treatment planning for psychological disorders (2nd ed., pp. 571– 605). New York, NY: Guil- ford Press.

Wittchen, H. U., Wunderlich, U., Gruschwitz, S., & Zaudig, M. (1997). SCID-I Structured Clinical Interview for DSM–IV (German version). Göttingen, Germany: Hogrefe.