Subtypes of Borderline Personality Disorder
Today, I address a fascinating article on BPD subtypes. Wolf, Scharoba, Noack, Keller & Weidner (2023) published “Subtypes of Borderline Personality Disorder in a Day-Clinic Setting—Clinical and therapeutic differences” in Personality Disorders: Theory, Research, and Treatment. Here’s the highly edited article with some bold:
This preliminary study aims at extending existing empirical evidence on subtypes of borderline personality disorders (BPDs) by identifying subtypes among patients with BPD, comparing their characteristics to the trait domains of the dimensional model of the International Classification of Diseases, 11th Revision (ICD-11; World Health Organization [WHO], 2022), and examining differences in sociodemographic, clinical, and therapeutic outcome variables. Data of N = 109 patients were subjected to cluster analysis based on the International Personality Disorder Examination variables for BPD and analyzed regarding differences in clinical and therapeutic variables. Clustering suggested a three-cluster solution, namely, internalizing (n = 35), externalizing (n = 28), and mixed subtype (n = 46). Subtypes showed differences in clinical variables and therapeutic outcomes with the internalizing showing more affective disorders and the mixed subtype showing the lowest therapeutic change in borderline-specific symptoms. Together, the present results correspond to the model of the ICD-11 (WHO, 2022). Clinical and treatment implications are being discussed.
Being the most prevalent personality disorder (PD) in the healthcare sector with rates ranging from 15% to 30%, borderline personality disorder (BPD) imposes high demands on healthcare systems. It is characterized by persistent instability in various areas of life, namely affectivity, interpersonal relationships, identity, and impulse control. In addition, patients with BPD show high comorbidities with other PDs as well as with many comorbidities, particularly mood, anxiety, substance use, and eating disorders. Currently, several therapeutic programs are available for the treatment of BPD: dialectical behavior therapy (DBT; mentalization-based therapy, schema-focused therapy, and transference-focused psychotherapy. However, not all patients with BPD benefit from treatment in the same way: after 2 years, 35%–60% of patients still met the Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria of BPD, declining further after 8 years to approximately 20%. Although these remission rates appear to be relatively high, some affected individuals regress more slowly than others. This raises the question of which factors influence therapy outcomes and whether certain characteristics of individuals with BPD are one of these factors. BPD is a very heterogeneous diagnosis: there are 256 different symptom constellations in the DSM and for five out of nine criteria, two patients may present with only one overlapping symptom.
The introduction of the ICD-11 marks a radical change in the classification history of PD from a categorical to a dimensional approach. Besides many advantages, new challenges for clinicians arise in the adequate description and corresponding therapy assignment of PD. The ICD-11 is conceptually aligned with the DSM-5 alternative model of PD. All PD are conceptualized dimensionally with the global PD and PD severity classification comprising the actual diagnoses. The question arises whether the borderline pattern specifier provides useful or distinct information that goes beyond the severity classification and trait domains. Aspects of severity, that is, self (e.g., identity disturbance) and interpersonal functioning (e.g., turbulent or conflictual relationships), along with emotional, cognitive, and behavioral manifestations (e.g., emotion dysregulation, impulsiveness), comprise most BPD features.
We included patients who underwent a DBT-informed day-clinic treatment between January 2009 and June 2017. Duration of treatment ranged between 2 and 16 weeks (16–107 days, M = 74.43 days, SD = 22.01). The treatment concept integrates modules of DBT alongside a variety of complementary therapeutic elements. A cluster analysis was performed using the patients’ data on subscales of the International Personality Disorder Examination (IPDE) measuring BPD (IPDE variables).
Borderline symptomatology was assessed at admission and discharge using the German version of the Borderline Symptom List 95. The Borderline Symptom List 23 (BSL-23) is the short version of the BSL-95 which aggregates to one global index and has an internal consistency of 0.97. The Brief Symptom Inventory 18 (BSI-18; Franke, 2000), a valid, reliable (Cronbach’s alpha .63–.93), and application—economical self-assessment questionnaire assessing general psychological with the Global Severity Index burden as well as the severity of syndromes of somatization, depression, and anxiety within the last 7 days.
The Beck Depression Inventory (BDI) assesses the severity from minimal or no depressive symptoms to severe symptoms.
Inclusion criteria for this study were, having a formal diagnosis of BPD, being 18 years of age or older, and having adequate mastery of the German language. The age of patients with BPD ranged from 18 to 49 years (M = 29.1, SD = 6.5), 84.4% (n = 92) of patients were female, and 57.7% were in a stable relationship.
The 10 subscales of the IPDE served as input variables for the cluster analysis, adjusted for the overall severity of borderline pathology (i.e., each case’s total dimensional score of IPDE) because the primary interest lay in the patterning across personality traits.
The present study aimed to identify subtypes among patients with BPD, analyze differences in clinical and therapeutic outcome variables, and compare characteristics of subtypes to the trait domains of ICD-11. The outcome of this preliminary study suggests a dimensional model, differentiating subtypes by personality traits, which corresponds roughly with the model of ICD-11 (WHO, 2022).
Cluster analysis results in three distinct subtypes: internalizing, externalizing, and mixed subtypes. All clusters share generally high values in mood instability and self-harm but seem to differ in variables that may be the cause of those symptoms. The internalizing may suffer from mood instability and self-harm due to high identity disturbance. The externalizing, on the other hand, may experience lower levels of identity disturbance precisely because they externalize relationship and mood instability (as a coping mechanism) leading to more frequent anger and a tendency to quarrelsomeness (devaluation). The mixed subtype equally reports impairments in all areas and may be the most impaired, showing the highest rates in definite BPD.
Subtypes showed at least moderate PD severity but only the mixed subtype showed higher PD severity, whereas the internalizing and externalizing subtypes did not differ in PD severity. However, all subtypes differed in aspects that fall under PD severity (self- and interpersonal functioning): while the internalizing and mixed report disturbance in both (relationship instability and identity disturbance), the externalizing reports mainly disturbance in interpersonal functioning (relationship instability). Moreover, subtypes could be distinguished based on the presence of ICD-11 trait domains. The externalizing subtypes’ most prominent trait domains seem to be dissociality and negative affectivity while the internalizing exhibits lower disinhibition and no dissociality, and instead displays predominantly negative affectivity, which is consistent with the fact that it shows the highest rate of affective disorders. This corresponds with Gamache et al. (2021) who highlight that mainly impulsivity/aggression and depressivity seem to distinguish BPD. The mixed subtype equally reports impairments in all areas, meeting all three trait domains (NA, D, A). This subtype may be the most impaired subtype, showing the highest rates in definite BPD and PD severity, and might qualify for the borderline pattern specifier to justify resource allocation.
Although the aspects of severity classification compromised most BPD symptoms, a clear differentiation of subtypes based on PD severity and trait domains was difficult: internalizing and externalizing subtypes showed differences within aspects that fall under PD severity (self- and interpersonal functioning) but did not differ significantly in PD severity. Furthermore, subtypes reported the same trait domains but different features within the trait domains. For example, all three subtypes exhibit disinhibition in different manners: the internalizing reported self-harm but no unexpected action or difficulties in maintaining action, whereas the externalizing reported self-harm and unexpected action but no difficulties in maintaining action. Together with the reported superiority of the AMPD in accounting for BPD (McCabe & Widiger, 2020), this supports the eventual inclusion of narrower facets in future editions of the ICD-11 including a distinction between self and interpersonal functioning. One example is the Five-Factor Personality Inventory for ICD-11 of Oltmanns and Widiger (2020), which includes scales for affective dysregulation, rapidly shifting emotions, separation insecurity, dysregulated anger, or suicidality which would allow the assignment of specific interventions (e.g., high-stress tolerance for patients with severe self-injury or suicidality; emotional regulation for patients with mood instability).
Subtypes respond differently regarding treatment outcome with the mixed subtype showing a lower therapeutic change in borderline-specific symptom burden which corresponds to the findings of Conway et al. (2016) who found that the severity of PD predicts therapy outcome but is in contrast with the results of Eurelings-Bontekoe et al. (2012). The question arises as, to why the mixed subtype shows the lowest therapeutic change. Reasons could be, first, the high impairment that requires a more intense treatment program. Second, it may be that severely impaired patients with BPD have difficulties differentiating between areas of impairment. Third, the focus of therapists and patients may diverge (e.g., care/attention vs. promoting autonomy/independence) and, fourth, the patients’ motivation is not sufficient for the requirements of the program. The externalizing, on the other hand, showed higher therapeutic change. Since DBT treatment especially targets impulsivity with high-stress-tolerance training (Bohus & Wolf-Arehult, 2013) and social skills and is effective on impulsive behaviors (Jamilian et al., 2014) this may explain the higher therapeutic change in this cluster because it showed higher unexpected action, anger, and quarrelsomeness but seems to have a relatively stable feeling of identity which may be helpful for mentalization during the therapeutic process (de Meulemeester et al., 2017).
Assignment of single modules of the DBT manual depending on the borderline subtype could be beneficial for treatment outcomes. Showing high values in identity disturbance (S), mood instability (NA), emptiness (NA), and self-harm (D), the internalizing may benefit stronger from modules like self-esteem or emotional regulation. Since it has the highest rate of affective disorders and the lowest pre-post BDI, further interventions should be directed at negative affectivity, for example, through complementary therapy of depression. Module assignments such as emotional regulation and depression therapy can be derived from the assignment of ICD-11 trait domains of negative affectivity and disinhibition, but fewer modules targeting identity diffusion, represented by PD severity. The externalizing, showing higher values on anger (A), quarrelsomeness (A), and unexpected actions (D), may additionally benefit from modules such as social skills training and high-stress tolerance, which can be inferred by coding the trait domains of dissociality and disinhibition. The question arises to what extent its PD severity allows assignment to specific modules for interpersonal disturbance and whether this impairment is sufficiently addressed by modules such as social skills training. The mixed subtype may be assigned to all DBT modules since they report both internalizing and externalizing impairments, respectively, all three trait domains. Since it also shows the highest PD severity, this subtype may need a more intensive or longer therapeutic program, which may be justified by coding a borderline pattern specifier in the diagnostic process.
Together, this study supports a model for subtyping BPD that differentiates patients in terms of personality traits and severity corresponding to the most recent versions of DSM and ICD (Bach et al., 2018). This study used an existing valid and standardized interview procedure to identify subtypes and examine therapeutic and clinical differences to assign therapeutic implications. It became clear that the assignment of severity classification and trait domains has some limitations for assigning appropriate treatment modules in terms of differentiating impairments within ICD-11 traits which support the eventual inclusion of narrower facets in future editions of the ICD-11. Results may be a stepping stone toward individual treatment planning to optimize current standards in therapy by assigning eligible specific interventions depending on the subtype.
This work is valuable for anyone working with clients with BPD. It also illustrates the value of the ICD-11.