Full access

Reviews

Published Online: 25 October 2022

Borderline Personality Disorder: Updates in a Postpandemic World

Lois W. Choi-Kain, M.D., M.Ed. [email protected], Zeynep Sahin, Ph.D., and Jenna M. Traynor, Ph.D.Authors Info & Affiliations

Publication: FOCUS

Volume 20, Number 4

https://doi.org/10.1176/appi.focus.20220057

CME

PDF/EPUB

Abstract

Progress in understanding borderline personality disorder has unfolded in the last decade, landing in a new COVID-19–influenced world. Borderline personality disorder is now firmly established as a valid diagnosis, distinct from its co-occurring mood, anxiety, trauma-related, and behavioral disorders. Further, it is also understood as a reflection of general personality dysfunction, capturing essential features shared among all personality disorders. Neuroimaging research, representing the vast neurobiological advances made in the last decade, illustrates that the disorder shares frontolimbic dysfunction with many psychiatric diagnoses but has a distinct signature of interpersonal and emotional hypersensitivity. This signature is the conceptual basis of the psychotherapies and clinical management approaches proven effective for the disorder. Medications remain adjunctive and are contraindicated by some guidelines internationally. Less invasive brain-based therapeutics show promise. The most significant change in the treatment landscape is a focus on briefer, less intensive formats of generalist management. Shorter variants of therapies, such as dialectical behavior therapy and mentalization-based treatment, are in the process of being shown to be adequately effective. Earlier intervention and greater emphasis on functional improvement are needed to more effectively curb the disabilities and risks of borderline personality disorder for patients and their families. Remote interventions show promise in broadening access to care.

Over five decades, there has been steady progress in defining the boundaries around borderline personality disorder, discriminating it from other major psychiatric disease categories, delineating its naturalistic course, identifying its biological features, and developing specific interventions proven for its treatment. Although stigma continues, an empirical foundation guiding effective clinical management has been established. New challenges have emerged. This review will provide a clinically relevant summary of the last decade of progress and controversy since our prior review (1).

Every health professional will encounter patients with borderline personality disorder. The good news is its prognosis has transformed from one of therapeutic nihilism to one of realistic optimism, especially with the advent of multiple evidence-based treatments (2). But grave concerns remain. The disorder is a significant risk factor not only for repeated life-threatening self-injurious problems, but also for other psychiatric disorders, medical illness, disability, violent crime, and victimization (3, 4); 5%–10% of patients with borderline personality disorder die by suicide (5). Missed or delayed diagnosis is pervasive, limiting appropriate care and resulting in repeated emergency room visits, inpatient admissions, and polypharmacy (6–10). Access to effective treatment, such as dialectical behavior therapy (DBT) or mentalization-based treatment (MBT), is limited by the intensive, lengthy format of these therapies, which requires training far outside the norm for most training programs for health care providers (11). With high dropout rates approaching 30%, and treatment nonresponse rates estimated at over 40%, patients with borderline personality disorder will frequently lack adequate readiness or conditions to respond to specialized treatment (12, 13). Therefore, all clinical professionals need a credible foundation for providing sound care.

Our understanding of underlying mechanisms, transdiagnostic features shared with co-occurring disorders, and targets for biological interventions has matured. Research efforts aim to refine existing treatment options. Understanding what works for whom, when, and where has sharpened our reconceptualization of steps of care for this disabling, and sometimes fatal, disorder to replace our prior one-size-fits-all attitude toward care. This review will provide updates on basic facts, current controversies in diagnosis, neurobiological advances, and clinical developments to aid health professionals in understanding and treating patients experiencing borderline personality disorder.

Basic Facts: Definitions, Epidemiology, Longitudinal Course

Disturbances in managing emotion, behavior, interpersonal relatedness, and identity define the four quadrants of borderline personality disorder symptoms. These symptoms destabilize major domains of functioning, which cascade into interruption of developmental trajectories and life opportunities, as well as pathways toward poor overall physical and mental health. The disorder affects 1%–3% of the general adolescent and adult population (14, 15). Most health care professionals will encounter patients with the disorder because these patients are overrepresented in acute settings, accounting for 9% of all emergency department visits, 50% of suicidal patients in emergency rooms, and 20%–25% of psychiatric inpatients (7–10). Almost three-quarters of this population will engage in an average of three suicide attempts in their lifetime (16), and one in 10–20 patients will die from suicide (3–5). Identity disturbance, frantic efforts to avoid abandonment, and feelings of emptiness predict suicide attempts (3); however, these cardinal features of the disorder are not part of routine assessment in the evaluation of suicidal patients.

Although borderline personality disorder is typically first diagnosed in adulthood, its evolution begins early. Negative emotionality and impulsivity predict harsh parenting styles, which in turn, potentiate risk for later diagnosis of the disorder in the presence of exposure to other social adversities, such as bullying (17, 18). Community-based prospective studies trace fluctuations in borderline personality disorder symptoms starting in childhood, peaking in adolescence, and normatively declining in early adulthood (19). Youths with higher levels of symptoms are at increased risk for retaining the diagnosis and developing other psychiatric disorders in adulthood (20, 21). They are more likely to experience significant social and functional impairment, with lower academic and occupational achievement, higher use of medical and psychiatric services, and less stable social support, even when symptoms decline (22, 23). Early identification of those at-risk can curb this trajectory. However, a widespread myth that the disorder cannot be diagnosed among those younger than age 18 dominates practice, contrary to criteria published as early as in the DSM-III and growing evidence verifying validity of the diagnosis in adolescence (22, 23).

Remission among adults with borderline personality disorder is common, with low rates of relapse (24, 25) (Figure 1). However, recovery (i.e., remission plus having full-time work or study and at least one emotionally sustaining relationship outside family), is more unstable (24, 25). The McLean Study of Adult Development (25) reported that virtually all (99%) participants with borderline personality disorder had achieved at least 2 years of remission by the 16-year follow up, with approximately one-third experiencing relapse, 78% attaining 8 years of remission, and only 10% relapsing. Sixty percent had achieved 2 years of recovery, but nearly half had lost it, demonstrating greater instability of recovery compared with remission. The Collaborative Longitudinal Study of Personality (CLPS) (24, 26) also found remission in 85% and low rates of relapse, but also more limited gains in functional recovery across 10 years. Lower Global Assessment of Functioning scores and more borderline personality disorder criteria met at baseline predicted worse outcomes, whereas good functioning, lower level of borderline or other personality disorder traits, and absence of major depressive disorder predicted remission (26, 27).

FIGURE 1. Longitudinal course of borderline personality disorder^a
^aSummarized from the McLean Study of Adult Development (25) and the Collaborative Longitudinal Personality Disorders Study (24).

These studies have also revealed elevated treatment use, complex psychiatric and medical comorbidity, functional impairment, and ongoing suicide risk among patients with borderline personality disorder compared with those with major depressive disorder or other personality disorders (3, 4, 28–32). Costs to society related to borderline personality disorder exceed that of most psychiatric conditions, including depression and anxiety disorders, and are comparable with those of schizophrenia (10). Symptoms manifesting in adolescence increase risk of later unemployment 22-fold and of later disability 15-fold, contributing to high indirect societal costs (33–37) (Table 1).

TABLE 1. Borderline personality disorder: basic facts and tools^a

Variable or assessment	Measure or population
Epidemiology % in adults (% in adolescents) (14, 15, 23, 32)
General population	2.7% (1%–3%)
Psychiatric inpatients	20%–25% (33%–49%)
Psychiatric outpatients	8%–18% (11%–22%)
Emergency room visits with suicidal ideation	56% (88%)
Suicidal behavior (3, 5, 16)
Average number of attempts	3
Suicide	3%–10% (50×greater than among general population)
Proportion of all suicides with borderline personality disorder	18% of all suicides; 33% of all youth suicides
Cost to society (10, 33)
Indirect costs of lost productivity (Europe)	$12,484–$50,000
Risk for disability with borderline personality symptoms<19	15-fold
Risk for unemployment with borderline personality symptoms<19	22-fold
Psychosocial functioning (24–37)
Functional impairment and changes in symptom levels	Good premorbid function provides positive prognosis; remission predicts good functioning
Poorer interpersonal functioning	Improvement in close relationships occurs with reduction in borderline personality disorder symptoms
Assessments
Semi-structured interviews or clinician rated
Structured Clinical Interview for DSM-5 Personality Disorders	All personality disorders
Diagnostic Interview for DSM-IV Personality Disorders	All personality disorders
International Personality Disorders Examination	All personality disorders in DSM-IV and ICD-10
Revised Diagnostic Interview for Borderlines	Borderline personality disorder only
Diagnostic self-reports
Personality Diagnostic Questionnaire-4	All personality disorders
Personality Assessment Inventory	All personality disorders
Borderline Symptom List	Borderline personality disorder only
Five-Factor Borderline Inventory	Borderline personality disorder only
Screening self-reports for borderline personality disorder
McLean Screening Instrument for borderline personality disorder
Borderline Personality Questionnaire
Zanarini Rating Scale for BPD
Dimensional personality functioning assessment
Personality Inventory for DSM-5	Personality traits
Level of Personality Functioning Scale Self-Report	Personality functioning

From Tomko et al. (14), Kaess et al. (15), Choi-Kain and Sharp (23), and Gunderson et al. (32).

Longitudinal studies have also elucidated the interactive course of borderline personality disorder with common co-occurring disorders. Borderline personality disorder is associated with complex patterns of comorbidity and ongoing vulnerability for poor mental health (34). The majority of individuals with the disorder in CLPS also met criteria for depression, with longer time to remission from and shorter time to relapse to a major depressive episode (35). In contrast, bipolar disorder co-occurred with borderline personality disorder in only 19.4%, but the courses of these conditions did not seem to significantly affect each other (29, 36). Co-occurrence with anxiety disorders had a bidirectional impact on course: borderline personality disorder exerted negative effects on the course of generalized anxiety disorder (GAD), social phobia, and posttraumatic stress disorder (PTSD), but only PTSD influenced borderline personality disorder, predicting its relapse. Improvement in borderline personality disorder predicted PTSD remission and those with borderline personality disorder had more new onsets of GAD and panic disorder, as well as relapses to obsessive-compulsive disorder (OCD) and panic disorder (27, 28) than other groups. New onsets of substance use disorder occurred more frequently among patients with borderline personality disorder than among those with major depressive disorder and other personality disorders (37).

Longitudinal findings have revised borderline personality disorder’s reputation of treatment resistance with evidence of high rates of remission. However, the findings also have demonstrated less common and more fragile recovery. Prospective longitudinal research also indicates the need for early intervention, with warnings of elevated risk for suicide, costly disability, and risk for persistent psychiatric complexity. Treatment can capitalize on the natural tendency for remission and should focus more on functional improvement to curb the ongoing psychiatric disabilities associated with the disorder.

Controversies: Dimensional Versus Categorical Diagnosis

Dimensional models of diagnosis address limitations of traditional categorical modes (38, 39). For example, the categorical threshold of a minimum of five criteria required for diagnosis is arbitrary. Fewer than five symptoms of borderline personality disorder continue to impair functioning (25). Even a single symptom is associated with significant increases in comorbid psychopathology, suicidal ideation and attempts, psychiatric hospitalizations, and significantly reduced functioning (40). The disorder almost always co-occurs with other diagnoses (41) (Table 2), suggesting it may be best captured by common symptom dimensions, rather than simple stand-alone diagnosis.

TABLE 2. Likelihood of lifetime axis I and II disorders among patients with borderline personality disorder and among a non–borderline personality disorder community sample

Psychiatric disorder	Odds ratio (controlled for sociodemographic characteristics)	99% CI	Odds ratio (unadjusted)	99% CI
Any substance use disorder	3.2*	2.73–3.97	4.50	3.57–5.71
Any drug use disorder	4.0*	3.30–4.72	5.78	4.67–7.14
Any anxiety disorder	7.7*	6.51–9.19	14.29*	10.87–18.87
Any mood disorder	9.1*	7.71–10.78	14.93*	11.63–19.61
Major depressive disorder	2.5*	2.07–2.90	11.76*	9.35–14.93
Dysthymia	3.2*	2.51–4.16	8.33*	6.62–10.53
Bipolar I	9.9*	8.11–12.01	16.39*	13.33–20.41
Bipolar II	4.3*	3.00–6.03	3.70*	2.71–5.05
Any other personality disorder	12.5*	10.54–14.86	15.87*	12.82–19.61
Paranoid	5.8*	4.69–7.25	12.20*	8.93–16.39
Schizoid	4.5*	3.55–5.76	14.29*	8.40–24.39
Schizotypal^a	26.5*	21.53–32.68	111.11*	66.67–200.00
Antisocial	3.5*	2.71–4.40	6.33*	4.76–8.40
Histrionic	6.5*	4.90–8.52	14.49*	6.85–31.25
Narcissistic	14.5*	12.12–17.42	55.56*	40.00–83.33
Avoidant	7.2*	5.45–9.56	11.63*	7.87–17.24
Dependent	8.2*	4.65–14.46	20.41*	9.71–41.67
Obsessive-compulsive	3.7	3.06–4.52	7.75*	5.65–10.64

Authors of report note this figure may be inflated due to the methodologic factors (14).

*p<0.01; adapted from Tomko et al. (14) and Grant et al. (34).

The ICD-11 (42) has shifted to a dimensional classification system. The DSM-5 (43) retains a categorical focus but includes the Alternative Model of Personality Disorders (AMPD), which, like the ICD-11, emphasizes disruptions of self- and interpersonal functioning as central to personality pathology (44). A third dimensional approach to personality diagnosis, the Hierarchical Taxonomy of Psychopathology (HiTOP), proposes a set of empirically derived hierarchical traits that are shared across spectrums of psychiatric conditions (45).

A brief overview of three prominent dimensional models is provided (Table 3). Despite divergence in symptom domains included across these three models, they generally converge with the five-factor model (46). All of these models propose at least five dimensional trait domains by which individual symptoms are more specifically described. The ICD-11 additionally emphasizes level of risk of harm to self or others as a central component of personality disorder severity (47). HiTOP proposes a general psychopathology factor (i.e., the “p” factor) reflecting the individual’s overall predisposition to experience psychiatric problems (48).

TABLE 3. Conceptualization and measurement of overall personality impairment severity and stylistic trait domains across the ICD-11, DSM-5 Alternative Model of Personality Disorders (AMPD), and Hierarchical Taxonomy of Psychopathology (HiTOP) models of dimensional psychopathology

Variable	ICD-11	DSM-5 AMPD	HiTOP
Components of global personality disorder severity	Global severity: self- and interpersonal functioning, risk of harm to self and/or others	Criterion A: self- and interpersonal functioning	Superspectra: general psychopathology severity (p factor)
Specific and/or stylistic trait domains	Trait qualifiers: negative affectivity, detachment, dissociality, disinhibition, anankastia	Criterion B trait domains: negative affectivity, detachment, antagonism, disinhibition, psychoticism	Spectra: internalizing, detachment, antagonistic externalizing, disinhibited externalizing, thought disorder, somatoform (provisional)

The ICD-11 has retained a borderline pattern qualifier, defined by the DSM categorical definition, among its classification options, because that conceptualization of the disorder is clinically familiar and compatible with existing interventions and a long body of literature. Simultaneously, the disorder is being dimensionally reconceptualized (49, 50). Substantial symptom overlap across internalizing and externalizing domains (51) suggests the disorder may be more reflective of general psychopathology that is shared across personality disorders (52). Supporting this suggestion, cross-sectional data from two nationally representative samples of U.S. adults (53) have shown high shared variance between the general psychopathology factor and a latent borderline personality disorder factor. Taken together, these findings provide empirical rationale for a reconsideration of borderline personality disorder’s position in diagnostic systems.

However, if the disorder is best captured by a single factor model, an entirely dimensional approach to classification may miss distinct phenotypes that have been identified by previous factor analyses. Although some of these phenotypes are more externalizing or internalizing (e.g., a “highly externalizing, risk taking, reckless” vs. “a mix of affective instability, identity disturbance, and dissociation”) (54), others appear more nuanced (e.g., angry and aggressive vs. angry and mistrustful) or prototypical (55), suggesting that a hybrid dimensional and categorical model is useful. Similarly, 10-year longitudinal data (56) from CLPS have shown that borderline and narcissistic traits in particular are prospectively predictive of suicidal behavior patterns; therefore, dimensional approaches may exclude pertinent categorical phenomena that are relevant to risk assessment.

Clinical translation of dimensional classification remains in its infancy. Preliminary frameworks for integrating the DSM-5 AMPD and HiTOP into practice suggest that assessment of self- and interpersonal impairment (i.e., criterion A severity) may be used to gauge levels of risk and the intensity of treatment that may be required before moving on to assessment of stylistic traits (57, 58). Existing interventions for the disorder already target dimensional trait domains, such as DBT, which addresses negative affectivity and disinhibition with emotion regulation and distress tolerance skills. A published framework (58) for integrating the DSM-5 AMPD into clinical practice acknowledges that targeting criterion B traits would involve clinical approaches that are both common (e.g., attention to therapy alliance) and specific (e.g., cognitive restructuring, psychodynamic interpretation). Therefore, dimensional reformulation of the disorder may not require dramatic alteration of existing treatments. Symptom profiles associated with treatment outcomes should be identified, addressing the unevenness of treatment response that has been observed when conceptualizing the disorder as a categorical entity (13, 59).

Discernment regarding which interventions are best suited to more severe levels of personality disorder is needed. Dimensional models do not conceptualize self-harm and suicidal behavior as distinct phenomena. Rather, these symptoms are subsumed under definitions of broad trait domains (e.g., the DSM-5 AMPD includes self-harm in its description of negative affectivity; HiTOP includes suicidal behavior in the internalizing distress subfactor). Because chronic self-harm and suicidality are often prominent features persisting throughout treatment (60), guidelines for management of self-destructive symptoms are needed.

Studies of dimensional personality classification would benefit from measurement of treatment outcomes in order to address existing gaps in clinical applicability. Examination of self-harm and suicide-related outcomes is crucial. Longitudinal clinical trial designs that assess whether dimensional classification improves prediction of treatment outcomes are needed to determine clinical superiority to existing systems of diagnosis.

Neuroimaging Advances

Over the past decade, the exponential growth of neuroimaging research, with exciting methodological approaches spanning from connectivity analysis to machine learning, has advanced our understanding of the neural basis of borderline personality disorder. Novel findings suggest that this disorder and other psychiatric disorders may share important neurobiological underpinnings—highlighting biomarkers of psychotherapy outcomes and bringing attention to clinical implications regarding interpersonal sensitivity, negative affectivity, and aggression for people with borderline personality disorder.

The Frontolimbic Hypothesis: Shared or Distinct Neurobiological Disruptions?

The most common brain imaging finding across borderline personality disorder research is that the disorder is associated with an imbalance in frontolimbic circuitry, which is thought to underpin emotion dysregulation (61). The emerging shift favoring dimensional psychopathology is concurrent with a proliferation of neuroimaging studies investigating whether frontolimbic disruptions are distinct to borderline pathology or are shared across disorders. Findings from structural and functional imaging studies have provided convincing data indicating that neurobiological abnormalities in borderline personality disorder may not be diagnosis-specific, pointing to a transdiagnostic model of frontolimbic disruptions. Other findings (62–75) have suggested a distinct and shared neurobiology between borderline personality disorder and other disorders, supporting a hybrid dimensional-categorical formulation (see Table 4 for review).

TABLE 4. Findings supporting dimensional and hybrid models of diagnosis of borderline personality disorder^a

Model supported	Citation	Sample	Method	Analysis or aims	Summary	Implications
Dimensional neurobiological model	Baranger et al., 2020 (62)	N=1,398 scans from the Duke Neurogenetics Study and Human Connectome Project	Structural MRI	152 structural metrics examined (e.g., cortical thickness, cortical surface area, subcortical volumes)	Borderline traits not associated with any of 152 structural metrics investigated	Traits that are specific to borderline pathology are not associated with brain structure
	Degasperi et al., 2021 (63)	52 studies N=1,104 borderline personality disorder and 1,100 control participants	Task-based fMRI (any task)	Activation likelihood estimation meta-analysis	No clusters of significant convergence or differences identified during the performance of any task; when analyses were restricted to emotion processing tasks; two significant clusters of greater activation, in borderline personality disorder vs. controls, in amygdalate and ACC	Borderline-specific pathology is not associated broadly with brain function during fMRI task performance
	Traynor et al., 2021 (64)	N=45 borderline personality disorder and 29 control participants	Resting-state fMRI	Connectivity analysis	Limbic system connectivity strength in borderline personality disorder significantly associated with severity of impairments in dimensional self- and interpersonal functioning (DSM-5 AMPD criterion A); typology-specific measure of borderline personality disorder symptom severity not significantly associated with any resting-state connectivity patterns	Limbic system disruptions are related to dimensional psychopathology impairments and are not borderline personality disorder–specific
	van Zutphen et al., 2017 (65); van Zutphen et al., 2020 (66)	N=51 borderline personality disorder, 26 cluster C, and 44 nonpatient control participants	Task-based fMRI (emotion processing task and affective inhibitory control task)	Activation analysis	Insula and TPJ hyperactivation during emotion processing in borderline personality disorder vs. other groups; hyperactivation in FEF and IPL during inhibition of negative vs. neutral stimuli, in borderline personality disorder vs. other groups; activation differences showed significant linear trend related to diagnostic severity vs. category	Differences in brain activation may be related more to the severity of psychopathology vs. diagnostic category
	Kebets et al., 2021 (67)	166 females with borderline personality disorder, bipolar disorder, or ADHD	Resting-state fMRI	Partial least squares analysis of resting-state signal variability	Frontolimbic network signal variability associated with emotion dysregulation across disorders; severity of emotion dysregulation accounted for 74% of shared variance in frontolimbic signal; depression and mania severity not associated with frontolimbic signal	Emotion dysregulation-related frontolimbic function is shared across disorders and not specific to borderline personality disorder
	Bilek et al., 2019 (68)	N=120 borderline personality disorder vs. 115 control participants	Task-based fMRI (threat processing task)	Activation analysis	Significant reduction in amygdala habituation to threat processing in borderline personality disorder vs. controls; effect unrelated to borderline personality disorder symptom severity or category but significantly associated with severity of childhood adversity	Abnormalities in brain activation during threat processing may be related more to severity of early adversity vs. categorical borderline personality diagnosis
	Metz et al., 2019 (69)	N=18 borderline personality disorder, 20 PTSD, 40 control participants	Task-based fMRI	RCT of autobiographical memory (AM) retrieval after administration of 10 mg of hydrocortisone or placebo	No between-group differences in neural activation during AM retrieval after hydrocortisone or placebo; hydrocortisone effects on amPFC, vlPFC, PCC, angular gyrus, and cerebellum positively correlated with childhood trauma questionnaire scores	Hydrocortisone-induced neural activation during AM retrieval is related to childhood trauma (vs. borderline personality disorder diagnosis)
	Schaub et al., 2021 (70)	N=313 with borderline personality disorder, cocaine use disorder, opioid use disorder, major depressive disorder, schizophrenia, or first-episode psychosis	Structural MRI	Examined associations between anhedonia and gray matter volume	Negative associations found between volumes of left putamen and bilateral cerebellum with anhedonia scores across disorders	Clinical problems, such as anhedonia, have shared neurobiological substrates across borderline personality disorder and other disorders
	Neukel et al., 2021 (71)	N=45 borderline personality disorder and 25 control participants	Task-based fMRI (emotion processing task)	Activation analysis	Abnormal precuneus activation during emotion processing of neutral and fearful stimuli in borderline personality disorder related to severity of impairments in self- and interpersonal functioning (DSM-5 AMPD criterion A)	Abnormalities in emotion-related brain activation may be related more to severity of psychopathology vs. diagnostic category
	Cremers et al., 2021 (72)	N=51 borderline personality disorder, 26 cluster C, and 44 nonpatient control participants	Task-based fMRI (emotion regulation task)	Support vector machine classification of borderline personality disorder	Unable to classify borderline personality disorder with high accuracy by using brain activation during emotion regulation; 55% classification accuracy was not statistically significant under permutation model	Brain activation during emotion regulation cannot be used to distinguish individuals with borderline personality disorder from those without the disorder
Hybrid dimensional-categorical neurobiological model	Yu et al., 2019 (73)	13 borderline personality disorder studies (N=395 borderline personality disorder and 415 control participants) and 47 bipolar disorder studies (N=2,111 bipolar disorder and 3,261 control participants)	Structural MRI	Meta-analysis of gray matter volume and density in borderline personality disorder vs. bipolar disorder	Shared: decreased gray matter volume and density in right medial OFC; specific to borderline personality disorder: decreased gray matter volume and density in amygdala and right parahippocampal gyrus; specific to bipolar disorder: decreased gray matter volume and density in right insula and thalamus and increased volume and density in right putamen (and putamen volume correlated with young mania rating scale scores)	Borderline personality disorder and bipolar disorder have both shared and distinct neural substrates
	Lou et al., 2021 (74)	15 borderline personality disorder datasets (N=442 borderline personality disorder and 441 control participants) and 11 PTSD datasets (N=214 PTSD and 258 control participants)	Structural MRI	Meta-analysis of gray matter volume in borderline personality disorder vs. PTSD	Shared: gray matter volume reductions in orbitofrontal gyrus and anterior cingulate cortex; specific to borderline personality disorder: gray matter volume increases in PCC/precuneus; specific to PTSD: gray matter volume decreases in amygdala and hippocampus	Borderline personality disorder and PTSD have both shared and distinct neural substrates
	De la Peña-Arteaga et al., 2021 (75)	N=19 borderline personality disorder, 20 major depressive disorder, and 19 control participants	Task-based fMRI (cognitive reappraisal task)	Activation analysis	Shared: decreased vlPFC activation during cognitive reappraisal; specific to major depressive disorder: overall prefrontal hypoactivation; specific to borderline personality disorder: abnormalities attributed to connectivity disruptions between vlPFC and temporal regions	Borderline personality disorder and major depressive disorder both have shared and distinct neural substrates

ACC, anterior cingulate cortex; amPFC, anterior medial prefrontal cortex; FEF, frontal eye fields; FFI-BPD, Five-Factor Inventory borderline personality disorder composite; fMRI, functional magnetic resonance imaging; IPL, inferior parietal lobe; NEO-FFI, NEO Five-Factor Inventory; OFC, orbitofrontal cortex; PCC, posterior cingulate cortex; RCT, randomized control trial; TPJ, temporal parietal junction; vlPFC, ventrolateral prefrontal cortex.

Structural and functional imaging studies powered to detect effects (∼N>1,000) provide strong evidence that neurobiological abnormalities in borderline personality disorder are not disorder specific. The largest structural MRI study (62) to date found no associations between borderline personality disorder traits and any of the 152 structural metrics examined. Similarly, a large meta-analysis of task-based functional MRI (fMRI) studies, which used activation likelihood estimation to compare participants with borderline personality disorder versus healthy control participants (63), found no clusters of significant convergence or difference. Instead, neurobiological alterations in borderline personality disorder appeared to be associated more with severity rather than with diagnostic category, as well as with phenomena that cut across disorders, such as emotion dysregulation, severity of childhood adversity, and anhedonia (64–72). On the other hand, some findings (73–75) have pointed to both shared and distinct neurobiological alterations in borderline personality disorder versus major depressive disorder, bipolar disorder, and PTSD.

Interpersonal Hypersensitivity, Negative Affectivity, and the Default Mode Network (DMN)

Disruptions in the functioning of the DMN may play a role in the interpersonal and affective features of borderline personality disorder. The DMN is composed of a set of brain structures along the cortical midline and in temporal, parietal, and subcortical areas. The DMN is most active during introspection or rumination (i.e., self-referential processing) as well as during mentalizing (i.e., social cognition) (76–78). During more “task on” executive performance (e.g., working memory or inhibition tasks), the DMN deactivates or “ramps down” to support the activation and coordination of other networks involved in executive functioning (e.g., central executive and attentional networks). Findings from a study of working memory (79) suggest that relative to control participants, the DMN in participants with borderline personality disorder did not show typical deactivation during executive task performance. Participants with borderline personality disorder also showed less activation in more typical executive areas of the brain during working memory performance (79). These findings suggest that individuals with the disorder may have difficulty “turning off the DMN” (i.e., disengaging from social cognition) while completing executive tasks (79) (e.g., focusing on how others are perceiving them at the expense of task efficiency). In turn, such difficulties may underpin hypermentalization, which is defined as making complex inferences of what another person is thinking without ostensible evidence (80, 81). A recent systematic review (82) reported findings that the DMN is often overactive in people with borderline personality disorder, and this overactivation is related to impaired social cognition. In another study (83), relative to healthy control participants and individuals with low self-esteem, the DMN of participants with borderline personality disorder was overactivated in response to insults, compared with compliments. Similar results have been observed during the Cyberball task, which simulates conditions of social inclusion and exclusion. Compared with healthy control participants, self-harming participants without borderline personality disorder, and individuals with major depressive disorder, participants with borderline personality disorder show overactivation in regions of the DMN during both social inclusion and exclusion, reflecting heightened relational sensitivity (84–86). One of these studies (84) found that high negative affect among participants with borderline personality disorder moderated the extent of DMN reactivity to social exclusion, suggesting that interpersonal reactivity among people with borderline personality disorder may oscillate with transient emotion states. Social exclusion also appears to impair effortful control and mentalization abilities of individuals with high, but not low, borderline personality disorder traits (87). Taken together, these findings underscore the magnitude of interpersonal hypersensitivity among those with borderline personality disorder and the strong influence that social exclusion can have on executive and social functioning.

Two other studies have further underscored the impact of negative affectivity on hypersensitivity among people with borderline personality disorder. One study (88) found increased neurobiological reactivity to repeated presentations of both negative and neutral emotional stimuli among those with borderline personality disorder, relative to those with avoidant personality disorder and to a healthy comparison group, and neurobiological reactivity was again moderated by ratings of negative affect among the group with borderline personality disorder. The other study (89) found a relationship between the magnitude of brain activation during a negative affective-inhibitory control task and the lethality of past suicidal behavior of individuals with borderline personality disorder, suggesting that negative affective states may also have an impact on cognitive processing and the medical seriousness of suicide attempts.

These findings highlight the relevance of strategies to enhance self-management to reduce negative affect and associated cognitive impairment and disinhibition (e.g., self-harm or suicidal behavior), which are likely to follow interpersonal conflict (e.g., rejection) among people with the disorder (90). The disorder appears associated with reactivity to both acceptance and rejection and to negative and neutral emotional stimuli, thereby underscoring the importance of interventions that target emotional and interpersonal hypersensitivity.

Biomarkers of Aggression and Hostility in Borderline Personality Disorder

Sex-specific aggression biomarkers in borderline personality disorder represent another emerging research direction. In a study with a sex-balanced sample (91), a script-driven aggression imagery fMRI study found that anger in males with the disorder (relative to females with the disorder and to control males) may be related to abnormal coupling of the prefrontal cortex with the amygdala. This finding converges with that of another study (92) which found a sex-specific association between externalizing anger and prefrontal-amygdala function in males with the disorder. Another analysis of females with the disorder (93) showed increased anger-related brain connectivity across a more extensive functional brain network, suggesting more comprehensive neural coordination during anger processing among females with the disorder.

Poor top-down control of aggression among males with borderline personality disorder has been inferred and the importance of sex-specific, anger regulation treatment targets emphasized (91). Interestingly, findings from a pilot treatment study of aggression-specific psychotherapy versus treatment as usual (71) found significant effects of aggression-specific treatment, but not of treatment as usual, on amygdala-prefrontal coupling, suggesting that interventions specifically targeting aggression may indeed help to correct or normalize anger-related frontolimbic disturbances among those with borderline personality disorder. A recent finding also suggests that fatty acid amide hydrolase (FAAH) inhibitors may be explored as a future pharmacotherapeutic option for aggression in borderline personality disorder; in the first positron emission tomography study of its kind (94), FAAH in the prefrontal cortex and amygdala of participants with borderline personality disorder was positively associated with measures of hostility.

Self-Harm in Borderline Personality Disorder and Biomarkers of Psychotherapy Outcomes

Progress in identifying biomarkers of self-harm and psychotherapy outcomes has developed in the last decade. A recent systematic review (95) on the neural substrates of self-harm in the disorder shows that self-harm is related to disruptions in impulse-control and affective brain regions. In one study (96), compared with control participants and participants with the disorder without self-harming behavior, patients with the disorder and self-harming behavior exhibited greater amygdala activation during repeated presentations of aversive stimuli, reflecting increased neurobiological sensitivity. Other studies have shown that the application of painful stimuli after social exclusion versus inclusion is associated with hyperactivation of the nucleus accumbens, a reward-related region, among participants with the disorder. This effect was mediated by anxious attachment style, implicating a reward-related function to self-harm following social rejection among anxiously attached individuals with borderline personality disorder (97). Along the same lines, another study (96) observed a dampening of amygdala activation in response to painful stimuli, providing further evidence for a regulating function of self-harm among people with the disorder. In Niedtfeld et al.’s study (98), DBT, but not treatment as usual, restored amygdala reactivity during pain processing among people with borderline personality disorder to that seen among healthy control participants.

Psychotherapy improvements among people with the disorder appear to be mediated by changes in prefrontal-limbic circuitry, with treatment-related effects commonly yoked to the structure and function of the amygdala and the anterior cingulate cortex (ACC) (99, 100). Neurobiological changes after treatment for borderline personality disorder are thought to reflect increased executive control and improved emotion regulation (100). After DBT and psychodynamic therapy, less amygdala activation is observed, with more heterogeneity in the directionality of prefrontal effects (99, 101, 102). In particular, the dorsal ACC is a candidate biomarker of psychotherapy change, which may be related to its role in appraisal and error-monitoring (99). Multimodal imaging findings (103) have shown an association between ACC connectivity and the use of negative cognitive emotion regulation strategies among unmedicated individuals with borderline personality disorder, relative to control participants. Changes in dorsal ACC connectivity during cognitive reappraisal following psychotherapy have also been shown (104).

Treatment

Numerous assessments are now validated for screening and diagnosing and for measuring changes in borderline personality disorder symptoms (33). However, clinical diagnosis can be made easily by reviewing the DSM-5 section II or III criteria collaboratively with patients to evaluate which features relate to their history of care. Stigma perpetuates reluctance to deliver “bad news” to patients despite the positive prognosis and relevant treatment options. Studies have reported that the diagnostic disclosure of borderline personality disorder is tolerated similarly to other diagnoses among adults (104) and was not a negative experience for adolescents but rather one that accurately captured their problems (105). Screening all patients by using the criterion of affective instability has been shown sensitive to detecting cases of borderline personality disorder for further evaluation (106).

Psychotherapy

Several psychotherapies effectively treat borderline personality disorder (1, 2, 107). Focus’s prior review described four empirically supported approaches—DBT, MBT, transference-focused psychotherapy, and good psychiatric management, which all have some empirical support and dissemination systems. Schema-focused therapy has also proven effective for the disorder, with a recent report (108) showing that a combination of group and individual schema-focused therapy led to greater reductions in borderline personality disorder severity than group schema-focused therapy alone or treatment as usual. Although these treatment approaches differ in their underlying theories of the development of the disorder and mechanism of change, they share common characteristics requiring structured roles and goals that organize the working relationship between clinician and patient. Across meta-analyses, no one therapy approach has been shown to be superior to others nor has duration or intensity determined outcome (59). Psychotherapies for borderline personality disorder remain among the few validated interventions for self-harm, among both adolescents and adults (109, 110). These therapies also improve anger, impulsivity, depression, and interpersonal problems (107). DBT and MBT stand out as the approaches with the best evidence (107). Treatments that are specifically tailored for the disorder work better compared with treatments that are not, whether they are specialist treatments requiring intensive training or generalist approaches, such as good psychiatric management and structured clinical management (2, 111–114).

In recognition that the treatment supply remains severely inadequate to meet demand (11), research has focused on less intensive, briefer treatments, which are proving effective and adequate for patients at acute and outpatient levels of care (115–117). Studies comparing 6-month to usual 12-month variants of DBT and MBT are nearing completion, showing promising results that justify shortening the length of treatment to expand supply (118). A recent report from McMain and colleagues’ (119) Feasibility of a Shorter Treatment and Evaluating Responses (FASTER) randomized controlled trial with 240 participants showed that 6 months of DBT was not inferior to 12 months, even at 24-month follow-up. Adjunctive treatments have also received substantial empirical support. Trials of systems training for emotional predictability and problem solving, DBT-skills training, and a brief psychoeducational bibliotherapy called manual-assisted cognitive therapy, demonstrate the utility of augmenting ongoing “good-enough” clinical care with these add-on modalities (120). Briefer treatments, ranging in length from 1 to 6 months, prior to potential transition to longer-term outpatient care, also show promise in reaching symptomatic stability and limiting suicidality and self-harm (115–119, 121). Stepped-care models that aim to optimally allocate resources according to clinical staging have been developed but need testing (122, 123).

Complex comorbidity is the rule rather than exception for people with borderline personality disorder (Table 2). Nearly half of those with borderline personality disorder also meet criteria for substance or alcohol use disorder (36), and one-third have co-occurring PTSD (28, 30). Progress in combining treatment for common co-occurring disorders has been made. Dynamic deconstructive psychotherapy, a psychodynamic-oriented approach designed to promote emotional processing to reduce black and white thinking, has two trials supporting its use (124, 125). DBT-PTSD, found to be effective in reducing PTSD and borderline personality disorder symptoms among patients with severe chronic childhood abuse, integrates DBT, acceptance and commitment therapy, and exposure, specifically targeting traumatic memory processing, self-concept, and social interactions (126). These treatments enable clinicians to deliver more parsimonious care.

Treatments for the disorder routinely encourage involvement of family members (i.e., caregivers) and significant others, which reduces relapse, facilitates treatment, and improves quality of family life (127). Several family support programs have been developed on the basis of DBT, MBT, and psychoeducation (127). All combine psychoeducation with skills training tailored for caregivers and family members, which enhances support for the patient to foster and sustain changes. Examples of guidelines for families are widely available at no cost (e.g., www.borderlinepersonalitydisorder.org/family-guidelines).

Pharmacology and Novel Brain-Based Therapeutics

Limited research on medication management for borderline personality disorder has been published in the last decade (127–130), and no medication has been approved by the Food and Drug Administration for the treatment of the disorder. Paradoxically, over 90% of patients with the disorder receive at least one psychotropic medication. Polypharmacy is common, with almost one in five patients receiving more than four psychotropic agents simultaneously (130). Treatment guidelines (131–133) diverge, with some advising against the use of medications and others prescribing symptom-based algorithms. Meta-analytic studies have indicated that mood stabilizers diminish symptoms of affective lability (especially anger), and second-generation antipsychotics improve general psychiatric severity of patients with borderline personality disorder (128). However, the quality of evidence for pharmacology studies for the disorder remains low and inconsistent. Placebo response is significant. In one of the largest trials to date, lamotrigine and placebo yielded significant treatment responses in 4 months, but without significant differences between them, suggesting the risks and costs inherent in lamotrigine were not justified (134). These outcomes suggest high quality psychiatric management focused on the disorder may exert an important treatment effect apart from specific psychotropic mechanisms (135). According to expert consensus, the main role of medications is to treat the usual co-occurring disorders that patients with borderline personality disorder have (128, 129). Recent studies (129) also have shown that combining medications with psychotherapy yields greater improvement in borderline personality disorder symptoms than either treatment alone.

Guidelines recommend conservative prescribing with attention to collaborative monitoring of targeted symptoms, discontinuing when not producing clear benefits, and short-term usage to minimize side effects. For most co-occurring diagnoses, the treatment of borderline personality disorder as a priority will improve symptoms of the other disorders (111). In the case of ADHD, anorexia, and substance use disorder, the priority is to treat these entities to promote learning. With OCD, panic disorder, and PTSD, stabilization of borderline personality disorder symptoms may make exposure treatments more tenable (136).

Noninvasive brain stimulation interventions are effective for reducing depression and suicidal ideation among people with treatment-resistant major depressive disorder (137), inspiring studies of their effectiveness with these symptoms for patients with borderline personality disorder. Although most findings have been preliminary, from small samples and unblinded trials, they have shown initial promise and warrant more rigorous testing. Evidence from a systematic review (138) of 24 studies on brain stimulation (most with repetitive transcranial magnetic stimulation [rTMS] of the dorsolateral or dorsomedial prefrontal cortex) among patients with borderline personality disorder supports the safety, tolerability, and preliminary efficacy of rTMS for borderline personality disorder, with the majority of these investigations reporting reductions in symptom domains of impulsivity, anger, emotion dysregulation, anxiety, and depression severity. However, only two rTMS trials to date (139, 140) have used a sham-controlled design involving rTMS over the dorsomedial prefrontal cortex, with findings favoring active rTMS for improving depression, anxiety, and impulsiveness. Prior reports also have shown reduced effectiveness of electroconvulsive therapy (ECT) for depression among patients with borderline personality disorder. However, two recent studies (141, 142) found that ECT was effective for treating depression of patients who screened positive on the McLean Screening Instrument for borderline personality disorder, although those findings may be limited in their generalizability to individuals meeting formal diagnostic criteria (143).

Conclusions

Over the last decade of progress, our field’s understanding of borderline personality disorder’s specific characteristics as a prevalent, sometimes fatal, disabling, but treatable condition, has solidified as dimensional conceptualizations of psychopathology have moved psychiatry into new directions. Research demonstrating borderline personality disorder’s unique and shared neuroscientific characteristics has provided a basis for understanding both common and distinct features of the disorder. These advances support a move toward a hybrid or integrative model of understanding borderline personality disorder that appears to represent fundamental features of general personality dysfunction more broadly. The treatments proven for the disorder may in fact be transdiagnostically effective across personality disorders as well as the major comorbid conditions of borderline personality disorder. Research elucidates that the disorder’s biological basis relates to both emotional and interpersonal hypersensitivity, showing that negative affectivity interrupts usual regulation of cognitive processes that support switches between DMN-driven reflection and executive and attentional functions that support effective decision making and action. Furthermore, there are sex-specific differences regarding aggression, but few treatments have been tested on adequately balanced samples of males and females, pointing to the need for more representative treatment samples in future studies.

Treatment research is actively refining what we already know. Psychotherapeutic interventions are primary, whereas pharmacologic and other brain-based therapeutics are adjunctive, specifically to treat co-occurring disorders. In the last decade, evidence that expands our clinical armamentarium to generalist, acute, and briefer steps of care confronts the reality that treatment is rarely one-size-fits-all for any disorder. Future studies should focus on answering which, if any, pharmacologic or brain-based therapeutics enhance psychotherapies or clinical management enough to clarify new basic standards of care. Earlier intervention and focus on functional rehabilitation over symptomatic remission are new directions that hold the potential to curb the detriments of the disorder over the course of a patient’s lifetime to enable more stable engagement in the usual life-building activities, such as work and love, that define human personality functioning. With the life-altering impact of COVID-19, the need for digital and telehealth interventions, which are showing preliminary promise, will become more central (144, 145). Continuing research and treatment for borderline personality disorder is crucial and hopeful, as reflected in life-changing discoveries and developments.

REFERENCES

Gunderson JG, Weinberg I, Choi-Kain L: Borderline personality disorder. Focus 2013; 11:129–145

Format	RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks
Direct importt
Citation style
Style

Copy to clipboard
Tips for downloading citations

Abstract

Basic Facts: Definitions, Epidemiology, Longitudinal Course

Controversies: Dimensional Versus Categorical Diagnosis

Neuroimaging Advances

The Frontolimbic Hypothesis: Shared or Distinct Neurobiological Disruptions?

Interpersonal Hypersensitivity, Negative Affectivity, and the Default Mode Network (DMN)

Biomarkers of Aggression and Hostility in Borderline Personality Disorder

Self-Harm in Borderline Personality Disorder and Biomarkers of Psychotherapy Outcomes

Treatment

Psychotherapy

Pharmacology and Novel Brain-Based Therapeutics

Conclusions

REFERENCES

Information

Published In

History

Keywords

Authors

Details

Notes

Competing Interests

Metrics

Citations

Export Citations

View options

PDF/EPUB

Get Access

Login options

Purchase Options

Not a subscriber?

Figures

Other

Share

Share article link

Share