Skip to main content
Full access

Evidence Tables for Individual Studies Supporting Guideline Statements

Appendix D

Psychoeducation

Psychoeducation vs. Wait-list

Study characteristics and main results of psychoeducation compared with WL control
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Zanarini and Frankenburg (2008)
Design: RCT
Setting: NR
Country: United States
Funding: Eli Lilly
N = 50
G1 (20): Delayed psychoeducation
G2 (30): Psychoeducation
Duration: 12 weeks
Inclusion: Females; ages 18–30 years; met DIB-R and DSM-IV criteria for BPD
Exclusion: Currently in any type of psychiatric treatment; schizophrenia, schizoaffective disorder, bipolar I disorder, or SUD
Mean age, years (SD): 19 (1.4)
Female: 100%
Race/ethnicity:
 White: 86%
Primary outcome: NR
No significant difference between G1 and G2 on ZAN-BPD
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 0
Differential attrition: < 10 percentage points
Moderate
Zanarini et al. (2018)
Design: RCT
Setting: NR
Country: United States
Funding: NIMH, government funding
N = 80
G1 (40): No Psychoeducation
G2 (40): Internet-based psychoeducation
Duration: 12 weeks
Follow-up: 12 months
Inclusion: Met DIB-R and DSM-IV criteria for BPD
Exclusion: Schizophrenia, schizoaffective disorder, or intellectual disability; acutely suicidal or fully manic at time of assessment; current physical condition that can cause serious psychiatric symptoms (e.g., lupus, MS); serious substance abuse
Mean age, years (SD):
 G1: 21 (3.1)
 G2: 22 (3.7)
Female: 100%
Race/ethnicity:
White: 69%
Black: 11%
Hispanic: 10%
Asian: 8%
Other: 3%
Primary outcome: NR
G2 significantly more effective than G1 on SAS (0.5 vs. 0.09, P = 0.049) after 12 weeks
G2 significantly more effective than G1 on ZAN-BPD scale after 12 months (4.46 vs. 0.0, P = 0.035); no significant differences on any other outcome measures after 12 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 4%
Differential attrition: < 10 percentage points
Moderate
AE = adverse event; BPD = borderline personality disorder; DIB-R = Diagnostic Interview for Borderlines–Revised; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; MS = multiple sclerosis; N = sample size; NIMH = National Institute of Mental Health; NR = not reported; RCT = randomized controlled trial; SAS = Social Adjustment Scale; SD = standard deviation; SUD = substance use disorder; WL = wait-list; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Psychosocial Interventions

Interpersonal Psychotherapy vs. Wait-List Plus Clinical Management

Study characteristics and main results of IPT compared with WL plus clinical management
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bozzatello and Bellino (2020)
Design: RCT
Setting: Outpatient, single center
Country: Italy
Funding: Government
N = 43
G1 (21): WL plus clinical management
G2 (22): IPT adapted for treating BPD: 50-minute sessions over 40 weeks; 22 sessions in first 20 weeks and 20 sessions in last 20 weeks
Duration: 10 months
Inclusion: Ages 18–60 years; attended Center for Personality Disorders and met DSM-5 criteria for BPD
Exclusion: Dementia or other cognitive disorders, schizophrenia or other psychotic disorders, or bipolar disorders; co-occurring MDE and/or substance abuse; taken psychotropic medications and/or psychotherapy 3 months previously; females of childbearing age not using birth control
Median age: 35
Female: 67%
Race/ethnicity: NR
Primary outcome: NR
G2 significantly lower severity of BPD on BPDSI (36.1 vs. 44.6, P = 0.01), symptom scores on BIS-11 (64.8 vs. 77.4, P = 0.03), and functioning scores on CGI-S (3.1 vs. 4.1, P = 0.009) and SOFAS (68.2 vs. 57.1, P = 0.02) than G1 after 10 months, but not on SHI
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 14%
Differential attrition: ≤1 0 percentage points
Moderate
AE = adverse event; BIS-11 = Barratt Impulsiveness Scale–version 11; BPD = borderline personality disorder; BPDSI = Borderline Personality Disorder Severity Index; CGI-S = Clinical Global Impression-Severity; DSM-5 = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; G1 = Group 1; G2 = Group 2; IPT = interpersonal psychotherapy; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized controlled trial; SHI = Self-Harm Inventory; SOFAS = Social Occupational Functioning Assessment Scale; WL = wait-list.

Interpersonal Psychotherapy Plus Fluoxetine vs. Clinical Management Plus Fluoxetine

Study characteristics and main results of IPT plus fluoxetine compared with clinical management plus fluoxetine
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bellino et al. (2006)
Design: RCT
Setting: Outpatient, single center
Country: Italy
Funding: None
N = 39
G1 (19): Clinical management plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with opportunity to increase to 40 mg/day beginning week 2
G2 (20): IPT in weekly 1-hour sessions plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with opportunity to increase to 40 mg/day beginning week 2
Duration: 24 weeks
Inclusion: DSM-IV BPD diagnosis; met criteria for MDE
Exclusion: Lifetime diagnosis of delirium, dementia, amnestic or other cognitive disorders, or schizophrenia or other psychotic disorders; MDE as expression of bipolar disorder; current SUD; treatment with psychotropic medication or psychotherapy during 2 months prior to study; females not using adequate birth control
Mean age, years (SD): 26 (3.7)
Female: 60% (reported as ratio of men:women = 3:5)
Race/ethnicity: NR
Primary outcome: NR
G2 significantly more effective than G1 for improving symptoms of depression (measured by Ham-D [9.1 vs. 12, P = 0.005])
No significant differences between G2 and G1 for anxiety for clinical global impressions (measured by CGI-S) or anxiety (measured by Ham-A)
Attrition: 17.9% (7/39)
 G1: 20.0% (4/20)
 G2: 15.8% (3/19)
Moderate
       
Bellino et al. (2010)
Design: RCT
Setting: Outpatient, single center
Country: Italy
Funding: None
N = 55
G1 (28): Clinical management plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with ability to increase to maximum of 40 mg/day beginning in week 2, plus 15–20 minutes of clinical management every 2 weeks dealing with clinical issues
G2 (27): IPT plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with ability to increase to maximum of 40 mg/day beginning in week 2, plus IPT adapted to BPD according to Markowitz’s model (IPT-BPD)
Duration: 32 weeks
Inclusion: DSM-IV-TR BPD diagnosis
Exclusion: Lifetime diagnosis of delirium, amnestic disorder, other cognitive disorders, schizophrenia or other psychotic disorders, bipolar disorder, or Axis I or II disorders; those receiving psychotropic medication in past 2 months and/or psychotherapy in past 6 months; females of childbearing age not using adequate birth control
Mean age, years (SD):
 G1: 26 (7.2)
 G2: 26 (6.4)
Female: 67%
Race/ethnicity: NR
Primary outcome: NR
No significant differences between G2 and G1 on BPDSI, Ham-A, Ham-D, CGI-S, and SOFAS
Attrition: 20% (11/55)
 G1: 21.4% (6/28)
 G2: 18.5% (5/27)
Moderate
BPD = borderline personality disorder; BPDSI = Borderline Personality Disorder Severity Index; CGI-S = Clinical Global Impression-Severity; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; DSM-IV-TR = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision; G1 = Group 1; G2 = Group 2; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; IPT = interpersonal psychotherapy; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SOFAS = Social Occupational Functioning Assessment Scale; SUD = substance use disorder.

Acceptance and Commitment Therapy vs. Treatment as Usual

Study characteristics and main results of ACT compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Morton et al. (2012)
Design: RCT
Setting: Outpatient, multicenter
Country: Australia
Funding: NR
N = 41
G1 (20): TAU; case management provided mostly by public mental health services
G2 (21): ACT; 12 group sessions in psychoeducational format (2 hours/week)
Duration: 13 weeks
Inclusion: Met four or more criteria for BPD (DSM-IV Axis I and Axis II diagnoses using SCID-I and SCID-II, respectively); registered client of public sector adult MHS
Exclusion: Current positive or negative psychotic symptoms other than reactive psychotic symptoms associated with BPD; significant risk of violent and/or threatening behavior toward other participants; intellectual disability, cognitive impairment, or difficulty speaking English, severe enough to interfere with participation
Mean age, years (SD):
 G1: 34 (9.0)
 G2: 36 (9.3)
Female: 93%
Race/ethnicity: NR
Primary outcome: BEST at 13 weeks
G2 significantly more effective than G1 on BEST (- 11.8 vs. - 2.4, P = 0.028), BHS (- 4.7 vs. + 0.7a, P = 0.006), and DERS (- 18.7 vs. + 5.6a, P = 0.008)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 22%
Differential attrition:
 G1: 30% (6/20)
 G2: 14% (3/21)
Moderate
ACT = acceptance and commitment therapy; AE = adverse event; BEST = Borderline Evaluation of Severity Over Time; BHS = Beck Hopelessness Scale; BPD = borderline personality disorder; DERS = Difficulty in Emotion Regulation Scale; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; MHS = mental health service; N = sample size; NR = not reported; RCT = randomized controlled trial; SCID-I = Structured Clinical Interview for DSM-IV Axis I Disorders; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SD = standard deviation; TAU = treatment as usual.
aControl group worsened over time, hence the positive change on the scale.

Manual-Assisted Cognitive Therapy vs. Treatment as Usual

Study characteristics and main results of MACT compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Weinberg et al. (2006)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Other, foundation
N = 30
G1 (15): TAU
G2 (15): MACT: Six sessions adjunctive to ongoing TAU, modified to focus on deliberate self-harm in BPD patients
Six sessions (duration NR)
Inclusion: Females; ages 18–40 years; met DSM-IV and DIB-R criteria for BPD; history of repetitive deliberate self-harm with at least one episode during month before enrollment
Exclusions: Comorbid psychotic disorders, bipolar I disorder, or SUD; elevated suicide risk (scoring  9 on BHS); describing concrete immediate suicide plan
Mean age, years (SD):
 G1: 26 (7.7)
 G2: 30 (8.6)
Female: 100%
Race/ethnicity:
 White: 93%
 Nonwhite: 7%
Primary outcome: NR
G2 significantly more effective than G1 in reducing frequency (1.98 vs. 6.69, P < 0.001) and severity (0.51 vs. 1.01, P < 0.001) of deliberate self-harm 6 months posttreatment
Attrition: 0% (0/30)
 G1: 0% (0/15)
 G2: 0% (0/15)
Moderate
BHS = Beck Hopelessness Scale; BPD = borderline personality disorder; DIB-R = Diagnostic Interview for Borderlines-Revised; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; MACT = manual-assisted cognitive therapy; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SUD = substance use disorder; TAU = treatment as usual.

Cognitive-Behavioral Therapy vs. Treatment as Usual

Study characteristics and main results of CBT compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Davidson et al. (2006); BOSCOT
Design: RCT
Setting: Outpatient, multicenter
Country: United Kingdom
Funding: Other, foundation
N = 106
G1 (52): TAU; inpatient and outpatient hospital services, community-based services, and primary and community care services
G2 (54): CBT; up to 30 sessions over 1 year (1 hour/session) with weekly supervision from CBT experts at each site
Duration: 24 months
Inclusion: Age 18–65 years; met criteria for at least five BPD items using DSM-IV Axis II Personality Disorders; received either inpatient psychiatric services or assessment at accident and emergency services
Exclusion: Currently receiving in-patient treatment for a mental state disorder or systematic psychological therapy or specialist service; evidence of organic illness, mental impairment, substance dependence, schizophrenia, or bipolar disorder
Mean age, years (SD): 32 (9.1)
 G1: 31 (9.4)
 G2: 32 (9.0)
Female: 84%
Race/ethnicity:
 White: 100%
Primary outcome: Suicidal acts, psychiatric hospitalization, accident, and emergency attendance at 24 months
G2 significantly lower number of suicidal acts per person (0.87 vs. 1.73, P = 0.02) and greater improvements on STAI (5.4 vs. 0.5, P = 0.01) than G1 after 24 months
No significant differences in suicidal acts; on STAI, BDI-II, EQ-5D, or SFQ; or in number of hospitalizations after 12 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 15%
Differential attrition:  10 percentage points
Moderate
AE = adverse event; BDI-II = Beck Depression Inventory; BOSCOT = Borderline Personality Disorder Study of Cognitive Therapy; BPD = borderline personality disorder; CBT = cognitive-behavioral therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; EQ-5D = European Quality of Life–5 Dimension; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SFQ = Social Functioning Questionnaire; STAI = State-Trait Anxiety Inventory; TAU = treatment as usual.

Dialectical Behavior Therapy vs. Wait-List/Treatment as Usual

Study characteristics and main results of DBT compared with WL or TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bohus et al. (2004)
Design: Nonrandomized clinical trial
Setting: Inpatient, single center
Country: Germany
Funding: Government, DFG other, BPDRF
N = 60
G1 (20): WL (TAU)
G2 (40): DBT; individual therapy (2 hours/week), group skills training (2 hours/week), group psychoeducation (1 hour/week), peer group meetings (2 hours/week), mindfulness group (1 hour/week), individual body-oriented therapy (1.5 hours/week), and therapist team consultations meetings (2 hours/week)
Duration: 3 months
Inclusion: Met DSM-IV criteria for BPD using SCID-II and DIB-R; one suicide attempt or minimum of two NSSI acts within the past 2 years
Exclusion: Comorbid schizophrenia, bipolar I disorder, substance abuse, or intellectual disability; living > 250 miles from inpatient center; ongoing outpatient DBT or DBT post-discharge
Mean age, years (SD):
 G1: 30 (5.4)
 G2: 29 (7.2)
Female: 100%
Race/ethnicity: NR
Primary outcome: NR
G2 significantly more effective than G1 on GSI (0.56 vs. 0.07, P = 0.005), GAF (11.4 vs. 1.3, P = 0.003), BDI (NR vs. 10.4, P = 0.002), STAI (- 8.2 vs. + 1.2, P < 0.001), and Ham-A (0.6 vs. NR, P = 0.01) and on self-mutilation (62% vs. 31%, P = 0.039)
No significant differences on DES and STAXI
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 17%
Differential attrition:
 G1: 5% (1/20)
 G2: 22% (9/40)
High
       
Carter et al. (2010)
Design: RCT
Setting: Outpatient, single center
Country: Australia
Funding: NR
N = 76
G1 (35): 6 months WL while receiving TAU
G2 (38): DBT; team-based approach including individual therapy, weekly group-based skills training, and telephone access to an individual therapist and therapist supervision groups
Duration: 12 months
Inclusion: Females; ages 18–65 years; met DSM-IV criteria for BPD; history of multiple episodes of deliberate self-harm with at least three self-reported episodes in preceding 12 months
Exclusion: Presence of disabling organic condition, schizophrenia, bipolar disorder, psychotic depression, florid antisocial behavior, or developmental disability
Mean age, years (SD): 25 (6.1)
Female: 100%
Race/ethnicity: NR
Sample demographics and outcome results reported among subset analyzed
Primary outcome: Deliberate self-harm and hospitalizations because of self-harm at 6 months
No significant differences on number of self-harm episodes, proportion of participants with self-harm, or hospitalizations
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 30%
Differential attrition:
 G1: 11% (4/35)
 G2: 47% (18/38)
Low
Feigenbaum et al. (2012)
Design: RCT
Setting: Outpatient, single center
Country: United Kingdom
Funding: Government, C&IHA, NTRHA
N = 42
G1 (16): TAU; standard care of a range of individualized service provisions according to patients’ needs through local crisis services
G2 (26): DBT; goal setting and commitment building (3–6 weeks), individual therapy (1 hour/week), group skills training (2.5 hours/week), and out-of-hours telephone consultation
Duration: 12 months
Inclusion: Males and females; ages 18–65 years; DSM-IV criteria for Cluster B personality disorder
Exclusion: Forensic history with evidence of current high and immediate risk to others; in long-term psychotherapeutic treatment for schizophrenia or bipolar disorder; substance abuse; severe cognitive impairment
Mean age, years (SD):
 G1: 35 (7.4)
 G2: 35 (7.8)
Female: 73%
Race/ethnicity: NR
Primary outcome: CORE-OM at 12 months
No significant differences on CORE-OM, DSH, DES, BDI, STAXI, and OAS-M
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 29%
Differential attrition:
 G1: 13% (2/16)
 G2: 39% (10/26)
High
       
Gregory and Sachdeva (2016)a
Design: Retrospective cohort
Setting: Outpatient, single center
Country: United States
Funding: APsaA
N = 41
G1 (16): TAU; unstructured psychotherapy
G2 (25): DBT; skill group sessions including learning mindfulness, emotion regulation, and distress tolerance, followed by individual sessions
Duration: 12 months
Inclusion: Age > 18 years; SCID-II; Individual Assessment Profile
Exclusion: Schizophrenia, intellectual disabilities, or dementia
Mean age, years (SD):
 G1: 29 (11.5)
 G2: 37 (10.2)
Female: 81%
Race/ethnicity:
 Caucasian: 88%
 Other: 12%
Primary outcome: BEST at 12 months
G2 significantly more effective than G1 on BDI (- 5.5 vs. - 0.6, P < 0.001)
No significant differences on BEST and SDS, or in number of suicide attempts and self-injuries
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/16)
 G2: 0% (0/25)
Attrition: 53%
Differential attrition: < 10 percentage points
High
       
McMain et al. (2017)
Design: RCT
Setting: Outpatient, single center
Country: Canada
Funding: Ontario Mental Health Foundation
N = 84
G1 (42): WL
G2 (42): Brief DBT; skills training only
20 weeks
Follow up: 32 weeks
Inclusion: Ages 18–60 years; met DSM-IV criteria for BPD; two suicidal and/or NSSI episodes in past 5 years, with one occurring within 10 weeks prior to enrollment
Exclusion: Met DSM-IV criteria for psychotic disorder, bipolar I disorder, or dementia; evidence of intellectual disability; participation in DBT program within past year
Mean age, years (SD): 30 (8.6)
Female: 79%
Race/ethnicity: NR
Primary outcome: Frequency of suicidal or NSSI episodes at 32 weeks
G2 significantly more effective than G1 to reduce suicidal and self-injurious episodes on LSASI (7.65 vs. 5.77, P = 0.04) and to improve symptoms on STAXI (8.44 vs. 4.79, P< 0.001) and DERS (20.80 vs. 4.74, P < 0.01)
G2 significantly more clinically relevant improvements on SCL-90-R than G1 (43.8% vs. 18.4%, P = 0.024)
No significant differences on DSHI, BSL-23, BDI, and BIS-11
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 16%
Differential attrition: < 10 percentage points
Moderate
       
Verheul et al. (2003); van den Bosch et al. (2005)
Design: RCT
Setting: Outpatient, with various settings
Country: The Netherlands
Funding: Dutch health insurance company
N = 64
G1 (33): TAU; clinical management from original referral source; generally no more than two sessions per month with a psychologist, psychiatrist, or social worker
G2 (31): Weekly DBT; individual CBT sessions with the primary therapist, skills-training groups (2–2.5 hours/session), and supervision and consultation meetings for therapists
Duration: 52 weeks
Inclusions: Female; ages 18–70 years; BPD; residing near Amsterdam; referred by psychologist or psychiatrist willing to sign agreement committing to deliver 12 months of TAU
Exclusions: DSM-IV diagnosis of bipolar disorder or (chronic) psychotic disorder; insufficient command of Dutch language; severe cognitive impairments
Mean age, years (SD): 35 (7.7)
Female: 100%
Race/ethnicity: NR
Primary outcome: NR
G2 more effective than G1 to reduce self-mutilating behavior (35% vs. 57%, P = 0.003); numerically lower frequency of suicidal attempts (7% vs. 26%, P = 0.06) for G2 than G1
Incidence of AEs: NR
Withdrawal due AEs: NR
Attrition: 19%
Moderate
AE = adverse event; APsaA = American Psychoanalytic Association; BDI: Beck Depression Inventory; BEST = Borderline Evaluation of Severity Over Time; BIS-11 = Barrett Impulsiveness Scale–11; BPD = borderline personality disorder; BPDRF = Borderline Personality Disorder Research Foundation; BSL-23 = Borderline Symptom List–23; C&IHA = Camden and Islington Health Authority; CBT = cognitive-behavioral therapy; CORE-OM = Clinical Outcomes in Routine Evaluation–Outcome Measure; DBT = dialectical behavior therapy; DERS = Difficulties in Emotion Regulation Scale; DES = Dissociations Experiences Scale; DFG = German Research Foundation; DIB-R = Diagnostic Interview for Borderlines-Revised; DSH = deliberate self-harm; DSHI = Deliberate Self-Harm Inventory; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; GSI = Global Severity Index; Ham-A = Hamilton Rating Scale for Anxiety; LSASI = Lifetime Suicide Attempt Self-Injury Interview; N = sample size; NR = not reported; NSSI = nonsuicidal self-injury; NTRHA = North Thames Regional Health Authority; OAS-M = Overt Aggression Scale-Modified; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SCL-90-R = Symptom Checklist–90–Revised; SD = standard deviation; SDS = Sheehan Disability Scale; STAI = State-Trait Anxiety Inventory; STAXI = State-Trait Anger Expression Inventor; TAU = treatment as usual; WL = wait-list.
aGregory and Sachdeva (2016) is a three-arm trial. The two relevant arms to DBT vs. TAU are reported in this table; other eligible comparisons are reported in Tables D–11 and D–20.

Dialectical Behavior Therapy vs. Mentalization-Based Treatment

Study characteristics and main results of DBT compared with MBT
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Barnicot and Crawford (2019)
Design: Nonrandomized clinical trial
Setting: Outpatient, multicenter
Country: United Kingdom
Funding: Government, NIH
N = 90
G1 (58): DBT
G2 (32): MBT
Duration: 12 months
Inclusion: Met DSM-IV criteria for BPD; was about to begin either outpatient DBT or MBT
Exclusion: Intellectual disability; difficulty communicating in English; insufficient capacity to provide informed consent
Mean age, years (SD): 31 (13.0)
Female: 72%
Race/ethnicity:
 White: 64%
 Black and minority: 36%
Primary outcome: NR
No significant differences between G1 and G2 on BEST, DERS, and DES and in number of self-harm incidents at 12 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 13%
Differential attrition: < 10 percentage points
High
AE = adverse event; BEST = Borderline Evaluation of Severity Over Time; BPD = borderline personality disorder; DBT = dialectical behavior therapy;  DERS = Difficulties in Emotion Regulation Scale; DES = Dissociative Experiences Scale; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; MBT = mentalization-based therapy; N = sample size; NIH = National Institutes of Health; NR = not reported; SD = standard deviation.

Dialectical Behavior Therapy vs. General Psychiatric Management

Study characteristics and main results of DBT compared with GPM for BPD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
McMain et al. (2009)
Design: RCT
Setting: Both inpatient and outpatient
Country: Canada
Funding: Government
N = 180
G1(90): Weekly individual therapy and medication management
G2 (90): DBT; weekly individual therapy, skills group sessions, and phone coaching with explicit focus on self-harm and suicidal behavior
1 year
Follow up: 36 months
Inclusion: Met DSM-IV criteria for BPD; ages 18–60 years; Two or more episodes of suicidal or NSSI episodes in past 5 years, at least one of which occurred in 3 months preceding enrollment
Exclusion: DSM-IV psychotic disorder, bipolar I disorder, delirium, dementia, or SUD in preceding 30 days; medical condition that precluded psychiatric medications; any serious medical condition likely to require hospitalization within next year (e.g., cancer)
Mean age, years (SD): 30 (9.9)
Female: 86%
Race/ethnicity: NR
Primary outcome: Suicidal episodes, NSSI at 12 months
No significant differences between G1 and G2 for number of suicidal events and NSSI, or on SCL-90-R, ZAN-BPD, BDI, and IIP after 12 months
G1 significantly greater improvements on BDI (17.4 vs. 12.7, P = 0.004) than G2 at 36-month follow-up; no significant differences at 36 months for any other outcomes
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 38%
Differential attrition: < 10 percentage points
High
AE = adverse event; BDI = Beck Depression Inventory; BPD = borderline personality disorder; DBT = dialectical behavior therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GPM = general psychiatric management; IIP = Inventory of Interpersonal Problems; N = sample size; NR = not reported; NSSI = nonsuicidal self-injury; RCT = randomized controlled trial; SCL-90-R = Symptom Checklist–90–Revised; SD = standard deviation; SUD = substance use disorder; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Dialectical Behavior Therapy vs. Systems Training for Emotional Predictability and Problem-Solving Behavior Therapy

Study characteristics and main results of DBT compared with STEPPS behavior therapy
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Guillén Botella et al. (2021)
Design: Nonrandomized clinical trial
Setting: Outpatient, multicenter
Country: Spain
Funding: NR
N = 72
G1 (45): Weekly individual and group DBT
G2 (27): Weekly STEPPS group therapy plus weekly individual therapy
Duration: 6 months
Inclusion: Met DSM-5 criteria for BPD
Exclusion: Moderate or severe intellectual disability, schizophrenia, or bipolar disorder
Mean age, years (SD): 32 (8.8)
Female: 94%
Race/ethnicity:
 Caucasian: 100%
Primary outcome: BSL-23 at 6 months
G1 significantly more effective than G2 on sum of BSL-23 (23.56 vs. 29.29, P = 0.03) after 6 months
No significant differences for any other measure
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 32%
Differential attrition: < 10 percentage points
High
AE = adverse event; BPD = borderline personality disorder; BSL-23 = Borderline Symptom List–23; DBT = dialectical behavior therapy; DSM-5 = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; SD = standard deviation; STEPPS = systems training for emotional predictability and problem-solving.

Dialectical Behavior Therapy vs. Dynamic Deconstructive Psychotherapy

Study characteristics and main results of DBT compared with DDP
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Gregory and Sachdeva (2016)a; Sachdeva et al. (2013)
Design: Retrospective cohort
Setting: Outpatient, single center
Country: United States
Funding: APsaA
N = 52
G1 (25): DBT; weekly 1-hour individual and 2-hour group sessions and telephone skills coaching
G2 (27): DDP; combined elements of translational neuroscience, object relations theory, and deconstructionist philosophy; weekly 1-hour individual sessions
Duration: 12 months
Inclusion: Age > 18 years; BPD by SCID-II and Individual Assessment Profile
Exclusion: Schizophrenia, intellectual disabilities, or dementia
Mean age, years (SD):
 G1: 29 (11.5)
 G2: 37 (10.2)
Female: 81%
Race/ethnicity:
 Caucasian: 88%
 Other: 12%
Primary outcome: BEST scores at 12 months
G2 significantly greater improvement than G1 on severity (BEST: 33.0 vs. 41.8, P = 0.04), self-injuries (SBQ: 1.3 vs. 2.4, P = 0.02), depression (BDI: 17.1 vs. 27.6, P = 0.009), and disability (SDS: 3.8 vs. 6.1, P = 0.049)
No differences between G1 and G2 in suicide attempts
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 53%
Differential attrition:
 G1: 64% (16/25)
 G2: 33% (9/27)
High
AE = adverse event; APsaA = American Psychoanalytic Association; BDI = Beck Depression Inventory; BEST = Borderline Evaluation of Severity Over Time; BPD = borderline personality disorder; DBT = dialectical behavior therapy; DDP = dynamic deconstructive psychotherapy; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; RCT = randomized controlled trial; SBQ = Suicidal Behaviors Questionnaire; SCID-II = Structured Clinical Interview for DSM-IV Axis II Personality Disorders; SD = standard deviation; SDS = Sheehan Disability Scale.
aGregory and Sachdeva (2016) is a three-arm trial. The two relevant arms to DBT vs. DDT are reported in this table; other eligible comparisons are reported in Tables D–7 and D–20.

Dialectical Behavior Therapy vs. Transference-Focused Psychotherapy vs. Supportive Therapy

Study characteristics and main results of DBT compared with TFP and supportive therapy
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Clarkin et al. (2007)
Design: RCT
Setting: Outpatient, multicenter
Country: United States
Funding: Other, foundation
N = 90
G1 (22): Weekly supportive treatment sessions
G2 (23): TFP; twice weekly individual sessions
G3 (17): DBT; weekly individual and group sessions and available telephone consultation
Duration: 12 months
Inclusion: Ages 18–50 years; met DSM-IV criteria for BPD
Exclusion: Comorbid psychotic disorders, bipolar I disorder, delusional disorder, delirium, dementia, and/or amnestic, other cognitive disorders, or SUD
Mean age, years (SD): 31 (7.9)
Female: 92%
Race/ethnicity:
 White: 68%
 Black: 10%
 Hispanic: 9%
 Asian: 6%
 Other: 8%
Primary outcome: Suicidal behavior at 12 months
No significant differences among G1, G2, and G3 in suicidal behavior or on BDI, BSI, or GAF at 12 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 31%
Differential attrition:
 G1: 27% (8/30)
 G2: 23% (7/30)
 G3: 43% (13/30)
High
AE = adverse event; BDI = Beck Depression Inventory; BPD = borderline personality disorder; BSI = Brief Symptom Inventory; DBT = dialectical behavior therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; G3 = Group 3; GAF = Global Assessment of Functioning; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SUD = substance use disorder; TFP = transference-focused psychotherapy.

Components of Dialectical Behavior Therapy vs. Other Components of Dialectical Behavior Therapy

Study characteristics and main results of components of DBT compared with other components of DBT for BPD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Andión et al. (2012)
Design: Nonrandomized clinical trial
Setting: Outpatient, single center
Country: Spain
Funding: Government, health department, La Caixa
N = 51
G1 (37): Weekly individual DBT therapy
G2 (14): Combined weekly individual and group DBT sessions
Duration: 12 months intervention, followed through 18 months
Inclusion: Ages 18–50 years; one or more suicide attempts and/or self-harm behaviors during previous month; met criteria for DSM-IV Axis II and Axis I Disorders
Exclusion: Intellectual disability, schizophrenia, or bipolar I disorder; previous DBT treatment
Mean age, years (SD): 26 (6.5)
Female: 100%
Race/ethnicity: NR
Primary outcome: Suicide attempts and self-harm at 12 and 18 months
No significant differences between groups on any outcome
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 10%
Differential attrition: < 10 percentage points
High
       
Linehan et al. (2015)
Design: RCT
Setting: Outpatient, multicenter
Country: United States
Funding: Government, NIMH
N = 99
G1 (33): Weekly standard DBT: skills training, individual therapy, telephone coaching, and a therapist consultation team
G2 (33): DBT individual therapy with no group skills training
G3 (33): Weekly DBT group skills training with no individual therapy
Duration: 1-year treatment followed through 2 years
Inclusion: Females; ages 18–60 years; met DSM-IV criteria for BPD; two or more suicide attempts and/or NSSI episodes in past 5 years, one or more suicide attempts or NSSI acts in 8 weeks prior to entering study, and one or more suicide attempts in past year
Exclusion: IQ < 70; DSM-IV criteria for current psychotic or bipolar disorders; seizure disorder requiring medication; required primary treatment for another life-threatening condition
Mean age, years (SD): 30 (8.9)
Female: 100%
Race/ethnicity:
 White: 71%
 Asian American: 5%
 Biracial: 22%
 Other: 2%
Primary outcome: Frequency and severity of suicide attempts and NSSI episodes at 12 and 24 months
No significant difference between groups in suicide attempts, NSSI acts, or suicide ideation
During the treatment year, G1 and G3 significantly greater improvement in depression than G2 (12.3 and 10.4 vs. 18.2, P = 0.02 on Ham-D) with no differences between groups in anxiety (Ham-A)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 26%
Differential attrition:
 G1: 18% (6/33)
 G2: 33% (11/33)
 G3: 27% (9/33)
High
       
Lyng et al. (2020)
Design: Prospective cohort
Setting: Outpatient, multicenter
Country: Ireland
Funding: NR
N = 88
G1 (54): Weekly standard DBT; individual therapy, group skills training, phone consultation (as needed), and therapist consultation team meeting
G2 (34): Weekly DBT group skills training
Duration: 6 months
Inclusion: DSM-IV-TR BPD or equivalent diagnosis of emotionally unstable personality disorder by community psychiatrist
Exclusion: Enduring psychotic disorder or primary (i.e., main reason for seeking treatment) alcohol or substance abuse disorder; suicide attempt in previous 6 months and/or ongoing medically serious self-harm; other weekly counseling
Mean age, years (SD): 33 (range 18–59)
Female: 83%
Race/ethnicity: NR
Primary outcome: Borderline symptoms, general psychopathology, suicidal ideation at 6 months
No significant differences between groups in BPD symptomatology, suicide ideation, and symptom severity index
G2 significantly greater improvement on BHS (8.0 vs. 11.91, P = 0.02) and DERS (96.24 vs. 115.12, P = 0.02)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 25%
Differential attrition:
 G1: 17% (9/54)
 G2: 38% (13/34)
High
AE = adverse event; BHS = Beck Hopelessness Scale; BPD = borderline personality disorder; DBT = dialectical behavior therapy; DERS = Difficulties in Emotion Regulation Scale; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; DSM-IV-TR = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision; G1 = Group 1; G2 = Group 2; G3 = Group 3; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; IQ = intelligence quotient; N = sample size; NIMH = National Institute of Mental Health; NR = not reported; NSSI = nonsuicidal self-injury; RCT = randomized controlled trial; SD = standard deviation.

Component of Dialectical Behavior Therapy Skills Training vs. Another Component of Dialectical Behavior Therapy Skills Training

Study characteristics and main results of component of DBT skills training compared with another component of DBT skills training for BPD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Carmona i Farrés et al. (2019a)
Design: RCT
Setting: Outpatient, single center
Country: Spain
Funding: Other, mixed
N = 70
G1 (35): Weekly group DBT interpersonal effectiveness skills training for 10 sessions
G2 (35): Weekly group DBT mindfulness skills training for 10 sessions
Duration: 10 weeks
Inclusion: Ages 18–50 years; DSM-IV BPD diagnosis; no comorbidities with schizophrenia, drug-induced psychosis, organic brain syndrome, SUD, bipolar disorder, intellectual disability, or MDE in course; no concurrent psychotherapy at study enrollment; no previous training in mindfulness, other meditation-contemplative practices, or any other mind-body practices
Exclusion: NR
Mean age, years (SD):
 G1: 33.29 (8.54)
 G2: 30.51 (6.9)
Female: 90%
Race/ethnicity: NR
Primary outcome: Emotional dysregulation (DERS) and impulsivity (BIS-11) at 10 weeks
No significant differences between G1 and G2 on DERS at 10 weeks
G2 significantly greater improvement on the BIS-11 (75.3 vs. 79.3, P = 0.03) at 10 weeks
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 28%
 G1: 20%
 G2: 37%
High
       
Carmona I Farrés et al. (2019b)
Design: RCT
Setting: Outpatient, single center
Country: Spain
Funding: Other, mixed
N = 65
G1 (32): Weekly group DBT interpersonal effectiveness skills training for 10 sessions
G2 (33): Weekly group DBT mindfulness skills training for 10 sessions
Duration: 10 weeks
Inclusion: Ages 18–50 years; DSM-IV BPD diagnosis; no comorbidities with schizophrenia, drug-induced psychosis, organic brain syndrome, SUD, bipolar disorder, intellectual disability, or major depressive episode in course; no concurrent psychotherapy at study enrollment; right-handed; IQ within normal range
Exclusion: NR
Mean age, years (SD):
 G1: 33.75 (8.78)
 G2: 31.03 (6.76)
Female: 89.2%
Race/ethnicity: NR
Primary outcome: DMN activation and deactivation during an executive task
No significant differences between G1 and G2 on BSL-23, BDI, STAI-T, or STAI-S at 10 weeks (decreases on outcome measures in both groups)
No between‐group differences in DMN activation or deactivation
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 23%
Differential attrition:  10 percentage points
High
Elices et al. (2016)
Design: RCT
Setting: Outpatient, single center
Country: Spain
Funding: Government
N = 64
G1 (32): Weekly group DBT interpersonal effectiveness skills training
G2 (32): Weekly group DBT mindfulness training
Duration: 10 weeks
Inclusion: Ages 18–45 years; met BPD criteria according to SCID-II and DIB-R
Exclusion: Lifetime diagnosis of schizophrenia, drug-induced psychosis, organic brain syndrome, or bipolar disorder; participation in any psychotherapy during study or having previously received DBT; having meditation/yoga experience
Mean age, years (SD): 32 (6.9)
 G1: 32 (6.82)
 G2: 32 (7.25)
Female: 86%
Race/ethnicity: NR
Primary outcome: Borderline severity at 10 weeks
G2 significantly reduced BPD symptoms on BSL-23 than G1 at 10 weeks (33.5 vs. 52.5, P = 0.001)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 31%
Differential attrition:
 G1: 22% (7/32)
 G2: 41% (13/32)
High
       
Schmidt et al. (2021)
Design: RCT
Setting: Outpatient, single center
Country: Spain
Funding: Other, mixed
N = 102
G1 (52): Weekly group DBT interpersonal effectiveness skills training for 10 sessions
G2 (50): Weekly group DBT mindfulness skills training for 10 sessions
Duration: 10 weeks
Inclusion: Ages 18–50 years; DSM-IV BPD diagnosis; no comorbidities with schizophrenia, drug-induced psychosis, organic brain syndrome, SUD, bipolar disorder, intellectual disability, or MDE in course; no concurrent psychotherapy at study enrollment; no previous experience in mindfulness meditation and DBT skills training
Exclusion: NR
Mean age, years (SD):
 G1: 33 (8.0)
 G2: 32 (8.0)
Female: 93%
Race/ethnicity: NR
Primary outcome: Borderline severity (BSL-23) at 10 weeks
G2 significantly greater improvements on BSL-23 at 10 weeks (37.38 vs. 48.90, P = 0.000) and on EQ at 10 weeks (31.28 vs. 27.48, P = 0.001)
No differences between groups on DERS
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: NR
Differential attrition: NR
High
AE = adverse event; BDI = Beck Depression Inventory; BIS-11 = Barrett Impulsiveness Scale-11; BPD = borderline personality disorder; BSL-23 = Borderline Symptom List–23; DBT = dialectical behavior therapy; DERS = Difficulties in Emotion Regulation Scale; DIB-R = Diagnostic Interview for Borderlines-Revised; DMN = default mode network; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; EQ = European Quality of Life; G1 = Group 1; G2 = Group 2; IQ = intelligence quotient; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Personality Disorders; SD = standard deviation; STAI‐S = State‐Trait Anxiety Inventory–State; STAI‐T = State‐Trait Anxiety Inventory–Trait; SUD = substance use disorder.

Dialectical Behavior Therapy vs. Cognitive Therapy

Study characteristics and main results of DBT compared with CT
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Lin et al. (2019)
Design: RCT
Setting: Other: university counseling centers
Country: Taiwan
Funding: Government
N = 82
G1 (40): Weekly CT group sessions, phone consultations as needed, and closed social media community for group members
G2 (42): Weekly DBT group skills training, phone consultation as needed, and closed social media community for group members
Duration: 8 weeks
Inclusion: College students; met criteria for BPD per BPDFS; score  21 on Ko’s Depression Inventory; one or more suicide attempts in past 6 months
Exclusions: Lifetime diagnosis of schizophrenia, schizoaffective disorder, bipolar disorder, or psychotic disorder; current severe depression and suicide risk indicating need for inpatient care and crisis intervention; current neurological signs and substance abuse during past 6 months
Mean age, years (SD):
 G1: 20.47 (0.71)
 G2: 20.40 (0.76)
Female: 87.8%
Race/ethnicity: NR
Primary outcome: Suicide attempt at 32 weeks
No significant difference between G1 and G2 on suicide reattempt (CMSADS-L Short form) and Ko’s Depression Inventory at 32 weeks
Compared with G1, G2 significant improvements on BPDFS (5.87 vs. 4.91, P <0.01) and in suicide ideation (ASIQ-S; 42.96 vs. 40.27, P <0.01) at 32 weeks
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 1%
Differential attrition: < 10 percentage points
Moderate
AE = adverse event; ASIQ-S = Adult Suicidal Ideation Questionnaire-Shortened Version; BPD = borderline personality disorder; BPDFS = Borderline Personality Disorder Features Scale; CMSADS-L = Chinese Version of the Modified Schedule of Affective Disorders and Schizophrenia–Lifetime; CT =cognitive therapy; DBT = dialectical behavior therapy; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation.

Dialectical Behavior Therapy vs. Community Therapy by Experts

Study characteristics and main results of DBT compared with community therapy by experts
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Linehan et al. (2006)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Government, NIMH
N = 111
G1 (49): Community treatment by experts; developed specifically for this study to control for factors previously uncontrolled for in DBT studies
G2 (52): DBT
Duration: 1 year
Inclusion: Females; ages 18–45 years; BPD; two suicide attempts or self-injuries in past 5 years, with at least one occurring in past 8 weeks
Exclusion: Comorbid schizophrenia, schizoaffective disorder, bipolar, psychotic disorder, or intellectual disability; seizure disorder requiring medication; mandate to treatment; need for primary treatment for another debilitating condition
Mean age, years (SD): 29 (7.5)
Female: 100%
Race/ethnicity:
 White: 87.0%
 Black: 4.0%
 Asian American: 2.0%
 Native American or Alaskan Native: 1.0%
 Other: 5.0%
Primary outcome: NR
G2 more effective than G1 in preventing suicide attempts (23% vs. 46%, P = 0.01), ED visits for suicide ideation (10.6% vs. 18.4%, P = 0.02), and hospital admissions for suicide ideation (14.9% vs. 18.4%, P = 0.004)
No significant differences between groups on NSSI, Ham-D, and RLI
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition:
 Overall: 18%
 G1: 27.5% (14/51)
 G2: 10.0% (6/60)
High
AE = adverse event; BPD = borderline personality disorder; DBT = dialectical behavior therapy; ED = emergency department; G1 = Group 1; G2 = Group 2; Ham-D = Hamilton Rating Scale for Depression; N = sample size; NIMH = National Institute of Mental Health; NR = not reported; NSSI = nonsuicidal self-injury; RCT = randomized controlled trial; RLI = Reasons for Living Inventory; SD = standard deviation.

Dialectical Behavior Therapy Plus REMS Treatments vs. REMS Treatments

Study characteristics and main results of DBT plus REMS treatments compared with REMS treatments alone
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bianchini et al. (2019)
Design: RCT
Setting: Inpatient, single center
Country: Italy
Funding: NR
N = 21
G1 (11): REMS (Residenze per l’Esecuzione delle Misure di Sicurezza, a small‐scale intensive therapeutic unit)
G2 (10): DBT plus REMS treatments
Duration: 12 months
Inclusion: Met criteria for BPD as measured by PAI; history of violence toward others
Exclusion: Cognitive deficit (IQ < 70) and/or comorbid neurological diseases
Mean age, years (SD): 42 (8.14)
Female: 0%
Race/ethnicity: NR
Primary outcome: NR
No between-group comparisons at end of treatment
Significant change on only two outcomes, DERS and BIS-11, within intervention group
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: NR
Moderate
AE = adverse event; BIS-11 = Barrett Impulsiveness Scale-11; BPD = borderline personality disorder; DBT = dialectical behavior therapy; DERS = Difficulties in Emotion Regulation Scale; G1 = Group 1; G2 = Group 2; IQ = intelligence quotient; N = sample size; NR = not reported; PAI = Personality Assessment Inventory; RCT = randomized controlled trial; REMS = Residenze per l’Esecuzione delle Misure di Sicurezza; SD = standard deviation.

Dialectical Behavior Therapy vs. Conversational Model

Study characteristics and main results of DBT compared with conversational model
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Walton et al. (2020)
Design: RCT
Setting: Outpatient, single center
Country: Australia
Funding: None
N = 166
G1 (83): Conversational model involving twice-weekly individual therapy
G2 (83): Weekly DBT; individual therapy, group training, and access to after-hours coaching
Duration: 14 months
Inclusion: Ages 18–65 years; DSM-IV BPD; three or more suicidal and/or NSSI episodes in past 12 months
Exclusion: Disabling organic conditions, current acute psychotic illness, antisocial behavior that posed significant threat to staff and fellow patients, or developmental disability; substance dependence; living > 1-hour’s drive from treatment facility; inability to speak or read English; prior treatment with DBT or conversational model
Mean age, years (SD): 27 (7.8)
Female: 77%
Race/ethnicity:
 White: 86%
 Aboriginal: 6%
 Other: 8%
Primary outcome: Suicide attempts and NSSI at 14 months and depression severity (BDI-II) at 14 months
No differences between groups in suicide attempts, NSSI, BPD severity (BPDSI-IV), interpersonal problems (IIP), dissociation (DES), and mindfulness at 14 months
G2 significantly greater reductions in BDI-II scores at 14 months (15.94 vs. 22.13, P = 0.005) and greater improvements in emotion regulation (DERS; 87.08 vs. 105.16, P = 0.008)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 29%
 G1: 24% (20/83)
 G2: 34% (28/83)
Moderate
AE = adverse event; BDI-II = Beck Depression Inventory-II; BPD = borderline personality disorder; BPDSI-IV = Borderline Personality Disorder Severity Index-IV; DBT = dialectical behavior therapy; DERS = Difficulties in Emotion Regulation Scale; DES = Dissociative Experiences Scale; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; IIP = Inventory of Interpersonal Problems; N = sample size; NR = not reported; NSSI = nonsuicidal self-injury; RCT = randomized controlled trial; SD = standard deviation.

Dialectical Behavior Therapy Skills Training vs. Standard Group Therapy

Study characteristics and main results of DBT compared with standard group therapy
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Soler et al. (2009)
Design: RCT
Setting: Outpatient, single center
Country: Spain
Funding: Government
N = 60
G1 (30): Weekly standard group therapy
G2 (29): Weekly group DBT skills training
Duration: 12 weeks
Inclusion: Age s 18–45 years; met DSM-IV criteria for BPD as assessed by SCID-II and DIB-R; CGI-S score of  4
Exclusion: Comorbid schizophrenia, drug-induced psychosis, organic brain syndrome, alcohol or other psychoactive SUD, bipolar disorder, intellectual disability, or MDE in course; current psychotherapy
Mean age, years (SD):
 G1: 29.97 (5.63)
 G2: 28.45 (6.55)
Female: 83.0%
Race/ethnicity: NR
Primary outcome: NR
No significant differences between G1 and G2 on CGI-BPD and SCL-90-R at 12 weeks
G2 significantly greater improvement on the Ham-D (11.1 vs. 16.0, P = 0.001) and Ham-A (16.6 vs. 13.0, P = 0.03) at 12 weeks
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 49%
Differential attrition:
 G1: 35% (10/29)
 G2: 63% (19/30)
High
AE = adverse event; BPD = borderline personality disorder; CGI-BPD = Clinical Global Impression Scale for Borderline Personality Disorder; CGI-S = Clinical Global Impression-Severity; DBT = dialectical behavior therapy; DIB-R = Diagnostic Interview for Borderlines–Revised; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SCL-90-R = Symptom Checklist–90–Revised; SD = standard deviation; SUD = substance use disorder.

Dynamic Deconstructive Psychotherapy vs. Treatment as Usual

Study characteristics and main results of DDP compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Gregory and Sachdeva (2016)a
Design: Retrospective cohort
Setting: Outpatient, single center
Country: United States
Funding: APsaA
N = 44
G1 (16): TAU: Unstructured psychotherapy
G2 (28): DDP: combined elements of translational neuroscience, object relations theory, and deconstructionist philosophy in weekly 1-hour individual sessions
Duration: 12 months
Inclusion: Age >18 years; BPD by SCID-II and Individual Assessment Profile
Exclusion: Schizophrenia, intellectual disabilities, or dementia
Mean age, years (SD):
 G1: 29 (11.5)
 G2: 28 (11.7)
Female: 81%
Race/ethnicity:
 Caucasian: 88%
 Other: 12%
Primary outcome: BEST at 12 months
G2 significantly more effective than G1 on change from baseline in BEST (14.1 vs. - 2.6, P = 0.006), BDI (- 12.6 vs. - 0.6, P < 0.001), and SDS (- 2.5 vs. 0.6, P <0.001)
No significant differences in the number of suicide attempts and self-injuries
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 53%
Differential attrition:
 G1: 69% (11/16)
 G2: 33% (9/27)
High
AE = adverse event; APsaA = American Psychoanalytic Association; BDI = Beck Depression Inventory; BEST = Borderline Evaluation of Severity Over Time; BPD = borderline personality disorder; DDP = dynamic deconstructive psychotherapy; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SD = standard deviation; SDS = Sheehan Disability Scale; TAU = treatment as usual.
aGregory and Sachdeva (2016) is a three-arm trial. The two relevant arms to DDP vs. TAU are reported in this table; other eligible comparisons are reported in Tables D–7 and D–11.

Mentalization-Based Treatment vs. Other Active Comparators

Study characteristics and main results of MBT compared with other active comparators
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Beck et al. (2020); M-GAB
Design: double-blinded RCT
Setting: Outpatient, multicenter
Country: Denmark
Funding: Government, region Zealand other, TrygFonden
N = 112
G1 (56): Standardized to at least 12 individual monthly sessions, and additional contact varied across clinics and therapists and according to patient need; therapists were nurses, psychologists, social workers, or psychiatrists not trained in or practicing MBT; treatment was not manualized
G2 (56): MBT delivered as a 1-year program with three components: 1) MBT introduction, 2) MBT group, and 3) MBT parents (90-minute sessions)
Duration: 12 months
Inclusion: Ages 14–17 years; met four or more DSM-5 BPD criteria and total score above clinical cutoff (> 67) on BPFS-C
Exclusion: Comorbid PDD, learning disability (IQ < 75), anorexia, psychosis, schizophrenia or schizotypal personality disorder, ASPD, or any mental disorder other than BPD considered primary diagnosis; current SUD (past 2 months; not substance abuse); current psychiatric inpatient treatment
Mean age, years (SD):
 G1: 16 (1.0)
 G2: 16 (1.1)
Female: 99% (111/112)
Race/ethnicity: NR
Primary outcome: BPFS-C
No significant differences between G2 and G1 on BPFS-C, BPFS for Parent, ZAN-BPD, Risk-Taking and Self-Harm Inventory for Adolescents, BDI for Youth, internalizing or externalizing symptoms on Youth Self-Report, Child Behavior Checklist, or Children’s GAS
Report of any AEs: 0
Attrition: 25.0% (28/112)
 G1: 19.6% (11/56)
 G2: 30.3% (17/56)
High
AE = adverse event; ASPD = antisocial personality disorder; BDI = Beck Depression Inventory; BPD = borderline personality disorder; BPFS = Borderline Personality Features Scale; BPFS-C = Borderline Personality Features Scale for Children; DSM-5 = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; G1 = Group 1; G2 = Group 2; GAS = Global Assessment Scale; IQ = intelligence quotient; MBT = mentalization-based treatment; M-GAB = Mentalization-Based Treatment in Groups for Adolescents with Borderline Personality Disorder; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SUD = substance use disorder; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Mentalization-Based Treatment vs. Supportive Therapy

Study characteristics and main results of MBT compared with supportive therapy
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bateman and Fonagy (2009); Bateman et al. (2021)
Design: RCT
Setting: Outpatient, single center
Country: United Kingdom
Funding: Other, foundation
N = 134
G1 (63): SCM individual and group sessions plus medication; therapy based on supportive approach with case management, advocacy support, and problem-oriented psychotherapeutic interventions
G2 (71): MBT plus medication; 18-month manualized weekly combined individual and group psychotherapy
Duration: 18 months
Follow-up: 6 years
Inclusion: Ages 18–65 years; DSM-IV BPD diagnosis; suicide attempt or episode of life-threatening self-harm within past 6 months
Exclusion: Current long-term psychotherapeutic treatment; met DSM-IV criteria for psychotic disorder or bipolar I disorder; opiate dependence requiring specialist treatment; mental impairment or evidence of organic brain disorder
Mean age, years (SD):
 G1: 31 (7.9)
 G2: 31 (7.6)
Female: 80%
Race/ethnicity:
 White: 72%
 Black: 18%
 Other: 10%
Primary outcome: Suicide, self-injury, and hospitalizations at 18 months
At 18 months, G2 significantly more effective than G1 in reducing life-threatening suicide attempts in prior 6-month period on SCL-90-R (0.03 vs. 0.32, P < 0.001), reducing severe self-harm incidents on SCL-90-R (0.38 vs. 1.66, P < 0.001), and reducing hospitalizations (0.03 vs. 0.19, P < 0.001) (composite of all three measures, 0.5 vs. 2.2, P < 0.001)
At 18 months, G2 greater improvement in 6-month periods free of suicidal behavior, severe self-harm, and hospitalizations (OR 0.28, 95% CI 0.13–0.61, P <0.002)
G2 significantly greater than G1 in number of participants who achieved primary recovery criteria (free of self-harm, suicide attempts, or inpatient hospital stays) and who remained well over 6-year follow-up period (75% vs. 51%, P = 0.02)
At 18 months, G2 significantly greater improvements in BDI (14.80 vs. 18.68, P <0.01), IIP (1.28 vs. 1.65, P <0.001), SCL-90-GSI (1.12 vs. 1.55, P <0.001), and GAF (60.9 vs. 53.2, P <0.001)
Incidence of AEs: NR at 18 months
Withdrawal due to AEs: NR
Attrition: 26% at 18 months; 39% at 6 years post-treatment
Differential attrition: < 10 percentage points
Moderate
Carlyle et al. (2020)
Design: RCT
Setting: Outpatient, single center
Country: New Zealand
Funding: NR
N = 72
G1 (34): Enhanced therapeutic case management with case managers using published manual of SCM
G2 (38): MBT: manualized weekly 1-hour individual sessions and weekly 1.5-hour group sessions
Duration: 18 months
Inclusion: BPD diagnosis using SCID-II
Exclusion: Patients diagnosed with psychoses or primary substance dependence; insufficient proficiency in English; concurrent engagement in structured psychological treatment for personality disorder
Mean age, years (SD):
 G1: 32 (11.7)
 G2: 32 (9.8)
Female: 99%
Race/ethnicity:
 NZ European: 79%
 Maori: 6%
 European other: 12.5%
 Other: 3%
Primary outcome: NSSI and suicide attempts at 18 months
At 18 months, no significant differences between groups on incidents of NSSI, suicide attempts, or hospitalizations
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 14%
Differential attrition: < 10 percentage points
Moderate
Jørgensen et al. (2013)
Design: RCT
Setting: Outpatient, single center
Country: Denmark
Funding: NR
N = 111 randomized; n = 85 treated
G1 (27): Biweekly group therapy and monthly group psychoeducational program for 6 months
G2 (58): Weekly individual and group MBT therapy and monthly group MBT psychoeducational program for 6 months
Duration: 24 months
Inclusion: Age  21 years; met DSM-IV BPD criteria as assessed by SCID-II; GAF score > 34
Exclusion: Met diagnostic criteria for antisocial or paranoid personality disorder at time of assessment; severe substance abuse (daily) requiring specialist treatment
Mean age, years (SD):
 G1: 30 (6.8)
 G2: 30 (6.5)
Female: 96%
Race/ethnicity: NR
Primary outcome: GAF at 24 months
At 24 months, G2 significantly greater improvement than G1 on therapist-rated GAF-F (56.7 vs. 51.3, P = 0.007) and GAF-S (58.5 vs. 54.0, P <0.001)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 32%
Differential attrition: < 10 percentage points
High
AE = adverse event; BDI = Beck Depression Inventory; BPD = borderline personality disorder; CI = confidence interval; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; GAF-F = Global Assessment of Functioning-Functioning; GAF-S = Global Assessment of Functioning-Symptoms; IIP = Inventory of Interpersonal Problems; MBT = mentalization-based treatment; N = sample size; NR = not reported; OR = odds ratio; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Personality Disorders; SCL-90-R = Symptom Checklist–90–Revised; SCL-90-GSI = Symptom Checklist-90-Global Severity Index; SCM = structured clinical management; SD = standard deviation.

Mentalization-Based Treatment vs. Psychodynamic Treatment Program

Study characteristics and main results of MBT compared with psychodynamic treatment program
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Kvarstein et al. (2015)
Design: Prospective cohort
Setting: Outpatient, single center
Country: Norway
Funding: NR
N = 345
G1 (281): 18 weeks of weekly inpatient group therapies followed by weekly outpatient group therapy
G2 (64): Weekly individual and group MBT therapy and monthly group MBT psychoeducational program
Duration: 36 months
Inclusion: NR; assessed baseline diagnostic status with the M.I.N.I. version 4.4 for DSM Axis-I diagnosis and SCID-II at baseline
Exclusion: Treated in transition period between G1 and G2; included in RCT during 2004–2006
Mean age, years (SD):
 G1: 30 (7.0)
 G2: 26 (6.0)
Female: 83.2%
Race/ethnicity: NR
Primary outcome: GAF, CIP, BSI-18 at 36 months
G2 significantly greater improvements on CIP (0.9 vs. 1.4, P <0.001), BSI-18 (0.8 vs. 0.9, P < 0.001), and GAF (63.0 vs. 56.0, P < 0.001)
No difference between G1 and G2 in self-harm and suicide attempts at 36 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 16%
Differential attrition:
 G1: 8% (22/281)
 G2: 50% (32/64)
High
AE = adverse event; BPD = borderline personality disorder; BSI-18 = Brief Symptom Inventory-18; CIP = Circumplex of Interpersonal Problems; DSM = Diagnostic and Statistical Manual of Mental Disorders; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; MBT = mentalization-based treatment; M.I.N.I. = Mini International Neuropsychiatric Interview; N = sample size; NR = not reported; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SD = standard deviation.

Day-Hospital Mentalization-Based Treatment vs. Specialized Psychotherapy

Study characteristics and main results of day-hospital MBT compared with specialized psychotherapy
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bales et al. (2015)
Design: Nonconcurrent cohort
Setting: Other day-hospital
Country: The Netherlands
Funding: NR
N = 204
G1 (175): Variety of psychotherapeutic treatments in inpatient, outpatient, and day-hospital settings
G2 (29): MBT in a day-hospital setting; daily group therapy, weekly individual therapy, individual crisis planning and art therapy twice a week, mentalizing cognitive group therapy, and writing therapy; medication consultation when indicated
Duration: 18-month treatment phase; actual treatment was mean of 15.5 months (3.8)
Inclusion: Age  18 years; met DSM-IV criteria for BPD
Exclusion: Schizophrenia, ADHD, bipolar disorder, psychotic disorders, or SUDs; intellectual impairment; organic brain disorder
Mean age, years (SD):
 G1: 30 (7.9)
 G2: 30 (6.2)
Female:
 G1: 86%
 G2: 69%
Race/ethnicity: NR
Primary outcome: Psychiatric symptoms, personality functioning at 18 months
G2 significantly greater improvements than G1 on GSI at 18 months (1.04 vs. 1.21, P = 0.01) and 36 months (0.73 vs. 1.04, P = 0.02)
G2 favored on SIPP-118 changes in (mal)adaptive personality functioning (results NR)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: NR (could assume none)
High
       
Laurenssen et al. (2018)
Design: RCT
Setting: Other inpatient and outpatient therapies, multicenter
Country: The Netherlands
Funding: Other, organization
N = 95
G1 (41): Manualized psychiatric treatment and system-oriented tailored care
G2 (54): Day-hospital MBT consisting of daily group psychotherapy, weekly individual psychotherapy, individual crisis planning, art therapy twice a week, mentalizing cognitive group therapy, and writing therapy
Duration: 18 months
Inclusion: Met DSM-IV criteria for BPD; score of  20 on BPDSI
Exclusion: Schizophrenia or bipolar disorder; substance abuse requiring specialist treatment; organic brain disorder
Mean age, years (SD):
 G1: 34 (10.6)
 G2: 34 (9.4)
Female: 79%
Race/ethnicity: NR
Primary outcome: BPDSI total score at 18 months
At 18 months, no significant differences between groups on any outcome
Incidence of AEs (among completers):
 G1: 0% (0/15)
 G2: 0% (0/33)
Withdrawal due to AEs: NR
Attrition: 50%
Differential attrition: < 10 percentage points
Moderate
ADHD = attention-deficit/hyperactivity disorder; AE = adverse event; BPD = borderline personality disorder; BPDSI = Borderline Personality Disorder Severity Index; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GSI = Global Severity Index; MBT = mentalization-based treatment; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SIPP-118 = Severity Indices of Personality Problems; SUD = substance use disorder.

Systems Training for Emotional Predictability and Problem Solving vs. Treatment as Usual

Study characteristics and main results of STEPPS compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Blum et al. (2008); STEPPS
Design: RCT
Setting: Other outpatient, inpatient, and community
Country: United States
Funding: Government, NIMH
N = 165
G1 (72 [data analysis based on 65]): TAU; 20 weeks of continued usual care, including individual psychotherapy, medication, and case management
G2 (93 [data analysis based on 59]): STEPPS plus individual therapy; 20 sessions (2 hours/week) of manual-based STEPPS group treatment that combines cognitive-behavioral elements with skills training; components included psychoeducation about BPD, emotion management skills training, and behavior management skills training
Duration: 20 weeks
Inclusion: Subjects with DSM-IV BPD who could designate a mental health professional and friend or relative to serve as system members
Exclusion: Non-English speaker; psychotic or primary neurological disorder; prior participation in STEPPS
Among 124 who received allocated intervention:
Mean age, years (SD): 32 (9.5)
Female: 83%
Race/ethnicity:
 White: 95%
 Black: 2%
 Other: 3%
Primary outcome: BPD-specific psychiatric symptoms (ZAN-BPD) measured at 20 weeks
Significantly improved symptoms of BPD with G2 than G1 at 20 weeks on ZAN-BPD (9.8 vs. 13.4, P = 0.001)
Significantly improved impulsivity of BPD with G2 than G1 on BIS (72.7 vs. 76.8, P = 0.004)
Significantly improved depression on BDI (22.0 vs. 25.8, P = 0.03)
Significantly improved global impressions and functioning with G2 than G1 on SCL-90 (12.5 vs. 14.1, P = 0.03), CGI-S (4.4 vs. 4.7, P < 0.001), CGI-I (2.7 vs. 3.8, P < 0.001), and GAS (50.5 vs. 43.5, P < 0.001)
No significant differences between groups on suicide attempts and self-harm acts (BEST); on SCL-90, BDI, CGI, and GAS between 20 weeks and 1 year; or SAS at 20 weeks or 20 weeks to 1 year
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 42%
Differential attrition:
 G1: 29% (21/72)
 G2: 52% (48/93)
High
       
Bos et al. (2010)
Design: RCT
Setting: Outpatient, multicenter
Country: The Netherlands
Funding: Other
N = 79
G1 (37): TAU
G2 (42): STEPPS plus individual therapy: 18 weekly sessions of STEPPS and follow-up session 3–6 months after intervention; components included psychoeducation about BPD, emotion management skills training, and behavior management skills training
Duration: 24 weeks
Inclusion: Met DSM-IV criteria for BPD by administering BPD modules and SCID-II; BDSI-IV with scores exceeding established cutoff on one or both subscales
Exclusion: Did not speak Dutch; cognitively impaired (IQ < 70); age < 18 years; treated involuntarily; presented imminent danger to self or others
Mean age, years (SD):
 G1: 32 (9.2)
 G2: 32 (5.6)
Female: 86%
Race/ethnicity: NR
Primary outcome: BPD-specific (BPD-40) and general psychiatric symptoms (SCL-90) at 1 year
Significantly improved BPD-specific symptoms (BPD-40: 78.2 vs. 88.6, P = 0.001), general psychiatric symptoms (SCL-90: 199.2 vs. 222.7, P = 0.001), and quality of life (WHOQOL-Bref: 12.6 vs. 11.3, P = 0.006) for G2 than G1
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 17%
Differential attrition:
 G1: 21% (9/42)
 G2: 11% (4/37)
Moderate
González-González et al. (2021)
Design: Prospective cohort
Setting: Outpatient, single center
Country: Spain
Funding: None
N = 118
G1 (98 [data analysis based on 28]): TAU
G2 (20 [data analysis based on 9]): STEPPS: 20 weekly sessions of group STEPPS psychotherapy, 5 sessions of group psychotherapy for companions, monthly sessions of individual and family psychotherapy, and possibility of therapy in case of emergency; combined with usual medication and/or psychiatric consultations
Duration: 18 months
Inclusion: DSM-5 BPD diagnosis, including self-harm or aggressive impulsive behaviors for past 2 years
Exclusion: Acute patients or those with comorbid pathology; cognitive, intellectual, or psychopathological impairment for daily life activities requiring care in rehabilitation center; receiving another psychotherapy treatment
Mean age, years (range): 34 (18–58)
Female: 85%
Race/ethnicity: NR
Primary outcome: NR
Significantly improved BPD-specific symptoms (BEST: 47.3 [14.1] vs. 28.8 [10.9], P < 0.01)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 59%
Differential attrition:
 G1: 55% (11/20)
 G2: 71% (70/98)
High
AE = adverse event; BDI = Beck Depression Inventory; BDSI-IV = Borderline Syndrome Index IV; BEST = Borderline Evaluation of Severity Over Time; BIS = Barratt Impulsivity Scale; BPD = borderline personality disorder; BPD-40 = Borderline Personality Disorder checklist-40; CGI = Clinical Global Impressions; CGI-I = Clinical Global Impression-Improvement; CGI-S = Clinical Global Impression-Severity; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; DSM-5 = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; G1 = Group 1; G2 = Group 2; GAS = Global Assessment Scale; IQ = intelligence quotient; N = sample size; NIMH = National Institute of Mental Health; NR = not reported; RCT = randomized controlled trial; SAS = Social Adjustment Scale; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SCL-90 = Symptom Checklist-90; SD = standard deviation; STEPPS = systems training for emotional predictability and problem-solving; TAU = treatment as usual; WHOQOL-Bref = World Health Organization Quality of Life Bref; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Transference-Focused Psychotherapy vs. Treatment as Usual

Study characteristics and main results of TFP compared with TAU
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Doering et al. (2010)
Design: RCT
Setting: Outpatient, multicenter
Countries: Austria and Germany
Funding: Other, Austrian bank
N = 104
G1 (52): TAU: treatment by community psychotherapists and medication treatments as needed
G2 (52): TFP: Two 50-minute sessions every week from experienced clinical psychologists or medical doctors; medications as needed
Duration: 12 months
Inclusion: Female; ages 18–45 years; DSM-IV BPD diagnosis; sufficient knowledge of German language
Exclusions: ASPD, schizophrenia, or bipolar I and II disorder with a major depressive, manic, or hypomanic episode during the previous 6 months; SUD during the previous 6 months; organic pathology or intellectual disability
Mean age, years (SD):
 G1: 27 (7.5)
 G2: 28 (6.8)
Female: 100%
Race/ethnicity: NR
Primary outcome: Suicide attempts, dropout from therapy at 12 months
Significantly fewer suicide attempts with G2 than G1 (13.7% vs. 21.2%, P = 0.009) for LOCF analysis but not for completers analysis (P 0.025)
G2 significantly more effective than G1 for achieving fewer than five DSM-IV criteria for BPD (42.3% vs. 15.4%, P = 0.002) and on GAF (58.62 vs. 56.06, P = 0.002)
No significant differences for self-harm acts, severity of symptoms, depression (BDI), and anxiety (STAI)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 34%
Differential attrition:
 G1: 42.3% (22/52)
 G2: 25% (13/52)
High
AE = adverse event; ASPD = antisocial personality disorder; BDI = Beck Depression Inventory; BPD = borderline personality disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; LOCF = last observation carried forward; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; STAI = State-Trait Anxiety Inventory; SUD = substance use disorder; TAU = treatment as usual; TFP = transference-focused psychotherapy.

Transference-Focused Psychotherapy vs. Schema-Focused Therapy

Study characteristics and main results of TFP compared with SFT
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Giesen-Bloo et al. (2006); Spinhoven et al. (2007)
Design: RCT
Setting: Outpatient, multicenter
Country: The Netherlands
Funding: Government, Dutch Health Care Insurance Board
N = 88
G1 (43): TFP: 50-minute sessions twice a week from therapists
G2 (45): SFT: 50-minute sessions twice a week from trained therapists
Duration: 3 years
Inclusion: Ages 18–60 years; DSM-IV BPD diagnosis; BPDSI-IV score > 20
Exclusion: Psychotic disorders, bipolar disorder, DID, ASPD, or ADHD; addiction requiring clinical detoxification; psychiatric disorders secondary to medical conditions
Mean age, years (SD):
 G1: 29.5 (6.5)
 G2: 31.7 (8.9)
Female: 93%
Race/ethnicity: NR
Primary outcome: BPDSI-IV at 36 months
G2 significantly more effective than G1 to improve BPDSI-IV at 36 months (16.24 vs. 21.87, P = 0.005, RR = 2.33, 95% CI 1.24–4.37)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 39%
Differential attrition:
 G1: 51% (22/43)
 G2: 27% (12/45)
High
ADHD = attention-deficit/hyperactivity disorder; AE = adverse event; ASPD = antisocial personality disorder; BPD = borderline personality disorder; BPDSI-IV = Borderline Personality Disorder Severity Index-IV; CI = confidence interval; DID = dissociative identity disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; N = sample size; NR = not reported; RCT = randomized controlled trial; RR = risk ratio; SD = standard deviation; SFT = schema-focused therapy; TFP = transference-focused psychotherapy.

Special Populations

Borderline Personality Disorder and Substance Use Disorder: Comprehensive Validation Therapy Plus 12-Step vs. Dialectical Behavior Therapy

Study characteristics and main results of CVT+12S compared with DBT in patients with BPD and SUD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Linehan et al. (2002)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Government, NIDA, NIH
N = 24
G1 (12): Weekly individual CVT+12S; weekly “12-and-12” NA group, case management, and phone consultation as needed and opiate agonist therapy
G2 (12): Weekly DBT individual and group skills training, case management, and phone consultation as needed and opiate agonist therapy
Duration: 12 months
Inclusion: Female; ages 18–45 years; BPD diagnosis according to PDE and SCID-II; current opiate dependence according to SCID-I; no indication of treatment coercion (e.g., court-ordered/agency-ordered to retain housing)
Exclusion: Did not meet criteria for BPD; met criteria for bipolar mood disorder; currently pregnant; did not complete pretreatment and/or medical evaluation
Mean age, years (SD): 36 (7.3)
Female: 100%
Race/ethnicity:
 Caucasian: 66%
 African American: 26%
 Other (Asian and Hispanic American): 4%
Primary outcome: Percentage of opiate-positive urine specimens
At end of 12-month treatment, G2 significantly lower percentage of opiate-positive urine specimens than G1 (t = 2.32, P < 0.02); no significant differences at 12 months for any other outcomes
No significant differences between G1 and G2 for percentage of opiate-positive urine specimens or parasuicidal behavior and on BSI or GAS at 16 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 21%
Differential attrition:
 G1: 0% (0/12)
 G2: 42% (5/12)
Moderate
AE = adverse event; BPD = borderline personality disorder; BSI = Brief Symptom Inventory; CVT+12S = comprehensive validation therapy plus 12-step; DBT = dialectical behavior therapy; G1 = Group 1; G2 = Group 2; GAS = Global Assessment Scale; N = sample size; NA = Narcotics Anonymous; NIDA = National Institute on Drug Abuse; NIH = National Institute of Health; NR = not reported; PDE = Personality Disorders Exam; RCT = randomized controlled trial; SCID-I = Structured Clinical Interview for DSM-IV Axis I Disorders; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SD = standard deviation; SUD = substance use disorder.

Borderline Personality Disorder and Substance Use Disorder: Substance Use Disorder Treatment vs. Mentalization-Based Treatment Plus Substance Use Disorder Treatment

Study characteristics and main results of SUD treatment compared with MBT plus SUD treatment in patients with BPD and SUD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Philips et al. (2018)
Design: RCT
Setting: Outpatient, multicenter
Country: Sweden
Funding: Multiple
N = 46
G1 (22): Standard SUD treatment
G2 (24): Standard SUD treatment plus combined individual and group MBT
Duration: 18 months
Inclusion: Males and females; ages 18–65 years; DSM-IV BPD and SUD diagnoses; currently undergoing treatment at SUD treatment clinic
Exclusion: Schizophrenia, schizoaffective disorder, bipolar disorder type I, cognitive impairment, ASD, or psychopathy; participation in psychotherapy outside of study; inability to communicate in Swedish
Mean age, years (SD): 36.7 (9.6)
Female: 80.4%
Race/ethnicity: NR
Primary outcome: BPDSI-IV, deliberate self-harm, suicide attempts, IIP, reflective functioning scale, GSI, at 18 months
No significant difference between groups on any outcome measure at 18 months
Attrition: 48%
Differential attrition: < 10 percentage points
High
ASD = autism spectrum disorder; BPD = borderline personality disorder; BPDSI-IV = Borderline Personality Disorder Severity Index-IV; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GSI = Global Severity Index; IIP = Inventory of Interpersonal Problems; MBT = mentalization-based treatment; N = sample size; NR = not reported; RCT = randomized clinical trial; SD = standard deviation; SUD = substance use disorder.

Borderline Personality Disorder and Alcohol Use Disorder: Dynamic Deconstructive Psychotherapy vs. Treatment as Usual in the Community

Study characteristics and main results of DDP compared with TAU in the community in patients with BPD and AUD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Gregory et al. (2008, 2009, 2010)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Other, university
N = 30
G1 (15): TAU: combination of individual psychotherapy, medication management, alcohol and drug counseling, professional and self-help groups, and/or case management
G2 (15): DDP: weekly, 1-hour sessions administered by the PI or by one of five psychiatry residents
Duration: 12 months
Inclusion: Ages 18–45 years; DSM-IV BPD diagnosis; active alcohol abuse or dependence
Exclusion: Schizophrenia or schizoaffective disorder, intellectual disability, or neurological condition that may produce secondary psychiatric symptoms (e.g., stroke, MS, partial complex seizures, TBI)
Mean age, years (SD): 29 (7.7)
Female: 80%
Race/ethnicity:
 White: 90%
 Black: 3.3%
 Hispanic or Latino: 3.3%
 American Indian or Alaska Native: 3.3%
Primary outcome: Parasuicide behavior, alcohol misuse, and institutional care at 12 months
No significant difference between G2 and G1 for parasuicide behavior, alcohol misuse, and dissociation at 12 months
G2 significant improvements in depression (21.0 vs. 25.9, P < 0.05) and in core symptoms of BPD (BEST) (33.6 vs. 38.4, P < 0.05) at 12 months
Attrition at 12 months: 37%
Differential attrition: < 10 percentage points
High
AUD = alcohol use disorder; BEST = Borderline Evaluation of Severity Over Time; BPD = borderline personality disorder; DDP = dynamic deconstructive psychotherapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; MS = multiple sclerosis; N = sample size; PI = principal investigator; RCT = randomized clinical trial; SD = standard deviation; TAU = treatment as usual; TBI = traumatic brain injury.

Borderline Personality Disorder and Eating Disorder: Cognitive-Behavioral Therapy vs. Dialectical Behavior Therapy

Study characteristics and main results of CBT compared with DBT in patients with BPD and eating disorder
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Navarro-Haro et al. (2018)
Design: Nonrandomized controlled trial
Setting: Outpatient, multicenter
Country: Spain
Funding: Government, national agency
N = 118
G1 (47): Weekly individual CBT, weekly group session, and pharmacological treatment
G2 (71): Weekly individual DBT, weekly DBT group skills training, and pharmacological treatment
Duration: 6 months
Inclusion: Age  18 years; DSM-IV BPD and eating disorder diagnoses
Exclusion: Psychotic disorder and/or bipolar I disorder; alcohol or other SUD; organic disease that could interfere with psychological treatment
Mean age, years (SD): 27 (8.8)
Female: 100%
Race/ethnicity: NR
Primary outcome: Suicide attempt frequency, NSSI at 6 months
G2 significantly more improved on BDI than G1 (23.9 vs. 29.8, P = 0.02) at 6 months
No significant differences between groups for suicide attempts, NSSI, or on GAF after 6 months; no significant differences between groups for depression, emotional regulation, or resilience at 6 years
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 8% at 6 months and 41.5% at 6 years
Differential attrition:  10 percentage points
Moderate
AE = adverse event; BDI = Beck Depression Inventory; BPD = borderline personality disorder; CBT = cognitive-behavioral therapy; DBT =  dialectical behavioral therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; N = sample size; NR = not reported; NSSI = nonsuicidal self-injury; SD = standard deviation; SUD = substance use disorder.

Borderline Personality Disorder and Eating Disorder: Specialist Supportive Clinical Management vs. Modified Mentalization-Based Treatment

Study characteristics and main results of SSCM compared with modified MBT in patients with BPD and eating disorder
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Robinson et al. (2016); NOURISHED
Design: RCT
Setting: Outpatient, multicenter
Country: United Kingdom
Funding: Government, NIHR
N = 68
G1 (34): SSCM: one session every 1–4 weeks for 20–26 sessions over 1 year
G2 (34): MBT: one individual and one group session per week for 1 year
Duration: 12 months
Inclusion: Age  18 years; DSM-IV eating disorder and BPD diagnoses or “BPD symptoms” from DSM-IV (impulsivity in two or more potentially self-damaging areas, recurrent suicidal or self-mutilating behavior)
Exclusion: Current psychosis; current inpatient or day-patient (3+ days/week); currently in individual or group psychological therapy; received MBT < 6 months prior to randomization; organic brain disease leading to significant cognitive impairment; BMI < 15
Mean age, years (SD): 31 (9.9)
Female: 93%
Race/ethnicity:
 White: 84%
Primary outcome: EDE global score at 18 months
No significant differences between G1 and G2 on ZAN-BPD at 18 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 78%
Differential attrition:
 G1: 85% (29/34)
 G2: 71% (24/34)
High
AE = adverse event; BMI = body mass index; BPD = borderline personality disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; EDE = Eating Disorder Examination; G1 = Group 1; G2 = Group 2; MBT = mentalization-based treatment; N = sample size; NIHR = National Institute for Health Research; NOURISHED = Nice OUtcomes for Referrals with Impulsivity, Self Harm and Eating Disorders; NR = not reported; RCT = randomized clinical trial; SD = standard deviation; SSCM = specialist supportive clinical management; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Borderline Personality Disorder and Major Depressive Disorder: Cognitive Therapy Plus Fluoxetine vs. Interpersonal Therapy Plus Fluoxetine

Study characteristics and main results of CT plus fluoxetine compared with IPT plus fluoxetine in patients with BPD and MDD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bellino et al. (2007)
Design: RCT
Setting: Outpatient, single center
Country: Italy
Funding: None
N = 32
G1 (16): Weekly CT plus fluoxetine
G2 (16): Weekly IPT plus fluoxetine
Duration: 24 weeks
Inclusion: Met DSM-IV-TR criteria for BPD and an MDE
Exclusion: Lifetime diagnosis of delirium, dementia, amnestic or other cognitive disorders, schizophrenia or other psychotic disorders, or bipolar disorder; current SUD; treated with psychotropic medication or psychotherapy during 2 months prior to study
Based on report among completers:
Mean age, years (SD): 31 (5.8)
Female: 73%
Race/ethnicity: NR
Primary outcome: Ham-D at 24 weeks
No significant differences between G1 and G2 on Ham-D, Ham-A, BDI-II, CGI-S, SOFAS, or SAT-P at 24 weeks
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 19%
Differential attrition:
 G1: 25% (4/16)
 G2: 13% (2/16)
Moderate
AE = adverse event; BDI-II = Beck Depression Inventory–II; BPD = borderline personality disorder; CGI-S = Clinical Global Impression-Severity; CT = cognitive therapy; DSM-IV-TR = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition, Text Revision; G1 = Group 1; G2 = Group 2; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; IPT = interpersonal psychotherapy; MDD = major depressive disorder; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized clinical trial; SAT-P = Satisfaction Profile; SD = standard deviation; SOFAS = Social Occupational Functioning Assessment Scale; SUD = substance use disorder.

Borderline Personality Disorder and Major Depressive Disorder: Interpersonal Psychotherapy Plus Fluoxetine vs. Clinical Management Plus Fluoxetine

Study characteristics and main results of IPT plus fluoxetine compared with clinical management plus fluoxetine in patients with BPD and MDD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bellino et al. (2006)
Design: RCT
Setting: Outpatient, single center
Country: Italy
Funding: None
N = 39
G1 (19): Clinical management plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with opportunity to increase to 40 mg/day beginning week 2
G2 (20): IPT in weekly 1-hour sessions plus fluoxetine 20–40 mg/day; initial fixed 20 mg/day with opportunity to increase to 40 mg/day beginning week 2
Duration: 24 weeks
Inclusion: DSM-IV BPD diagnosis; met criteria for MDE
Exclusion: Lifetime diagnosis of delirium, dementia, amnestic or other cognitive disorders, or schizophrenia or other psychotic disorders; MDE as an expression of bipolar disorder; current SUD; treatment with psychotropic medication or psychotherapy during 2 months prior to study; females not using adequate birth control
Mean age, years (SD): 26 (3.7)
Female: 60% (reported as ratio of males:females = 3:5)
Race/ethnicity: NR
Primary outcome: NR
G2 significantly more effective than G1 for improving symptoms of depression (measured by Ham-D [9.1 vs. 12, P = 0.005])
No significant differences between G2 and G1 in anxiety on clinical global impressions (measured by CGI-S) or anxiety (measured by Ham-A)
Attrition: 17.9% (7/39)
 G1: 20.0% (4/20)
 G2: 15.8% (3/19)
Moderate
BPD = borderline personality disorder; CGI-S = Clinical Global Impression-Severity; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; IPT = interpersonal psychotherapy; MDD = major depressive disorder; MDE = major depressive episode; N = sample size; NR = not reported; RCT = randomized clinical trial; SD = standard deviation; SUD = substance use disorder.

Borderline Personality Disorder and Posttraumatic Stress Disorder: Dialectical Behavior Therapy Alone vs. Dialectical Behavior Therapy Plus Dialectical Behavior Therapy-Prolonged Exposure

Study characteristics and main results of DBT alone compared with DBT plus DBT-PE in patients with BPD and PTSD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Harned et al. (2014, 2018)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Government, NIMH
N = 26
G1 (9): DBT: weekly individual therapy, group training, and therapist consultation team meeting and as-needed phone consultation
G2 (17): DBT-PE: weekly PE protocol and DBT as well as group DBT skills training and as-needed phone consultation
Duration: 1 year
Inclusion: Female; ages 18–60 years; DSM-IV BPD and PTSD diagnoses; can remember at least some part of index trauma; recent and recurrent intentional self-injury; lives within commuting distance of clinic
Exclusion: Met criteria for psychotic disorder, bipolar disorder, or intellectual disability; legally mandated to treatment; required primary treatment for another debilitating condition (i.e., life-threatening anorexia nervosa)
Mean age, years (SD): 33 (12)
Female: 100%
Race/ethnicity:
 White: 81%
 Biracial: 15%
 Asian-American: 4%
Primary outcome: PTSD (PSS-I), intentional self-injury (SASII) at 15 months
Numerically greater improvements in suicide attempts and on NSSI, PSS-I, SASII, Ham-A, Ham-D, and GSI across both groups; no statistical tests performed
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 42%
Differential attrition:  10 percentage points
High
AE = adverse event; BPD = borderline personality disorder; DBT = dialectical behavior therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GSI = Global Severity Index; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; N = sample size; NIMH = National Institute of Mental Health; NR = not reported; NSSI = nonsuicidal self-injury; PE = prolonged exposure; PSS-I = PTSD Symptom Scale–Interview; PTSD = posttraumatic stress disorder; RCT = randomized clinical trial; SASII = Suicide Attempt Self-Injury Interview; SD = standard deviation.

Adolescents With Borderline Personality Disorder: Manualized Good Clinical Care vs. Cognitive Analytic Therapy

Study characteristics and main results of manualized good clinical care compared with CAT in adolescents with BPD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Chanen et al. (2008)
Design: RCT
Setting: Outpatient, single center
Country: Australia
Funding: Government, NHMRC other, VicHealth, Colonial
N = 86
G1 (42): Weekly group standardized good clinical care
G2 (44): Weekly CAT
Duration: 24 months
Inclusion: Ages 15–18 years; met two to nine DSM-IV criteria for BPD; any personality disorder or disruptive behavior disorder symptom; low socioeconomic status; depressive symptoms; history of abuse or neglect
Exclusion: Learning disability, psychiatric disorder, pervasive developmental disorder, or severe primary Axis I disorder; more than nine sessions of specialist mental health treatment in previous 12 months; sustained psychosis and met criteria for Early Psychosis Prevention and Intervention Centre
Mean age: NR
Female: NR
Race/ethnicity: NR
Primary outcome: Psychopathology, parasuicidal behavior, global functioning at 24 months
No significant differences between G1 and G2 for parasuicidal behavior or on BPD Total Score and SOFAS at 24 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 9%
Differential attrition:  10 percentage points
Moderate
AE = adverse event; BPD  = borderline personality disorder; CAT = cognitive analytic therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; N = sample size; NHMRC = National Health and Medical Research Council; NR = not reported; RCT = randomized clinical trial; SOFAS = Social Occupational Functioning Assessment Scale.

Adolescents with Borderline Personality Disorder and Substance Use Disorder: Individual Drug Counseling vs. Integrative Borderline Personality Disorder-Oriented Adolescent Family Therapy

Study characteristics and main results of individual drug counseling compared with I-BAFT in adolescents with BPD and SUD
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Santisteban et al. (2015)
Design: RCT
Setting: Outpatient, single center
Country: United States
Funding: Government
N = 40
G1 (20): Twice-weekly individual drug counseling and monthly family meeting with caregivers
G2 (20): Twice-weekly I-BAFT, including family and individual therapy and skills-building interventions
Duration: 7 months
Inclusion: Ages 14–17 years; DSM-IV BPD and substance use diagnoses
Exclusion: NR
Mean age, years (SD):
 G1: 16 (0.8)
 G2: 16 (0.8)
Female: 38%
Race/ethnicity:
 Hispanic: 85%
Primary outcome: Substance use, BPD behaviors at 12 months
No significant differences between G1 and G2 in substance use or BPD behavior at 12 months
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 33%
Differential attrition:
 G1: 40% (8/20)
 G2: 25% (5/20)
High
AE = adverse event; BPD = borderline personality disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; I-BAFT = integrative borderline personality disorder–oriented adolescent family therapy; N = sample size; NR = not reported; RCT = randomized clinical trial; SD = standard deviation; SUD = substance use disorder.

Pharmacotherapy

Second-Generation Antipsychotics vs. Placebo

Study characteristics and main results of SGAs compared with placebo
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Black et al. (2014)
Design: Double-blinded RCT
Setting: Outpatient, multicenter
Country: United States
Funding: AstraZeneca
N = 95
G1 (29): Placebo
G2 (33): Quetiapine ER (150 mg/day)
G3 (33): Quetiapine ER (300 mg/day)
Duration: 8 weeks
Inclusion: Males and females; ages 18–45 years; DSM-IV criteria for personality disorders; score   9 on ZAN-BPD
Exclusion: History of psychotic disorder, neurological condition, or cognitive impairment; current SUD or abuse; medically unstable; history of lack of response to SGA; pregnant or lactating; acutely suicidal
Mean age, years (SD):
 G1: 30 (8.8)
 G2: 28 (8.0)
 G3: 30 (8.1)
Female: 30%
Race/ethnicity:
 European-Caucasian: 78%
 Other: 21%
Primary outcome: ZAN-BPD at 8 weeks
G2 (but not G3) significantly more effective than G1 on ZAN-BPD (P = 0.03)
G3 (but not G2) significantly more effective on SCL-90 than G1 (P = 0.03)
G2 and G3 significantly more effective on MOAS (P = 0.01)
No significant differences on BIS, MADRS, and SDS
Incidence of AEs:
 G1: 86% (25/29)
 G2: 88% (29/33)
 G3: 91% (30/33)
Withdrawal due to AEs: NR
Attrition: 33%
Differential attrition:
 G1: 21% (6/29)
 G2: 33% (11/33)
 G3: 42% (14/33)
Moderate
       
Bogenschutz and Nurnberg (2004)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: Eli Lilly
N = 40
G1 (20): Placebo
G2 (20): Olanzapine (2.5–20 mg/day)
Duration: 12 weeks
Inclusion: Ages 18–60 years; DSM-IV BPD diagnosis; medically stable
Exclusion: Other psychiatric disorders, SUD, or actively suicidal
Mean age, years (SD): 32 (10.3)
Female: 63%
Race/ethnicity:
 White: 58%
 Hispanic: 25%
 Asian/Pacific Islander: 8%
 Other: 10%
Primary outcome: CGI-BPD at 12 weeks
Significantly greater improvement of G2 than G1 on CGI-BPD (P = 0.03)
No significant differences on SCL-90, Ham-A, Ham-D, MOAS, and GAF
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/20)
 G2: 20% (4/20)
Attrition: 43%
Differential attrition:
 G1: 35% (7/20)
 G2: 50% (10/20)
High
Linehan et al. (2008)
Design: Double-blinded RCT
Setting: University hospital
Country: United States
Funding: Eli Lilly
N = 24
G1 (12): Placebo
G2 (12): Olanzapine (5 mg/day)
Duration: 6 months
Inclusion: Females; ages 18–60 years; met SCID-II and Borderline Personality Disorder Examination criteria for BPD; MOAS irritability subscale score  6
Exclusion: Schizophrenia, bipolar I disorder, schizoaffective disorder, MDD with psychotic features or other psychotic disorder, intellectual disability, seizure disorder, or SUD
Mean age, years (SD): 37 (9.0)
Female: 100%
Race/ethnicity:
 White: 79%
 Black: 4%
 Native American: 4%
 Latino: 4%
 Other: 8%
Primary outcome: NR
No significant differences between G1 and G2 on MOAS and Ham-D and for self-inflicted injury
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/12)
 G2: 8% (1/12)
Attrition: 33%
Differential attrition:  10 percentage points
High
Nickel et al. (2006, 2007)
Design: Double-blinded RCT
Setting: University hospitals
Country: Austria, Germany
Funding: None
N = 52
G1 (26): Placebo
G2 (26): Aripiprazole (15 mg/day)
Duration: 8 weeks
Follow up: 18 months
Inclusion: Males and females; age  16 years; DSM-IV BPD diagnosis
Exclusion: Schizophrenia; current use of other psychotropic medication; past termination of aripiprazole; current psychotherapy; pregnancy; suicidal ideation; severe somatic illness; alcohol or drug abuse
Mean age, years (SD):
 G1: 21 (4.6)
 G2: 22 (3.4)
Female: 83%
Race/ethnicity: NR
Primary outcome: SCL-90-R, Ham-D, Ham-A, STAXI at 8 weeks
G2 significantly greater improvements than G1 on SCL-90-R (15.0 vs. 4.9, P < 0.001), Ham-D (6.4 vs. 2.1, P = 0.002), Ham-A (7.0 vs. 3.3, P = 0.007), and STAXI (13.6 vs. 5.7, P < 0.001)
18-month follow-up for SCL-90-R: 17.9 vs. 1.4, P < 0.01
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 25%
Differential attrition:  10 percentage points
Moderate
Pascual et al. (2008)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: Spain
Funding: Pfizer, government funding
N = 60
G1 (30): Placebo
G2 (30): Ziprasidone (40–200 mg/day)
Duration: 12 weeks
Inclusion: Males and females; ages 18–45 years; DSM-IV BPD diagnosis; current use of medically accepted contraception for females
Exclusion: Schizophrenia, drug-induced psychosis, organic brain syndrome, alcohol or other SUD, bipolar disorder, intellectual disability, or MDE in course; CGI-S score  4
Mean age, years (SD):
 G1: 29 (6.3)
 G2: 29 (6.0)
Female: 82%
Race/ethnicity: NR
Primary outcome: CGI-BPD at 12 weeks
No significant differences on CGI-BPD, SCL-90, Ham-A, Ham-D, and in clinical psychotic symptoms
Incidence of AEs:
 G1: 13% (4/30)
 G2: 37% (11/30)
Withdrawal due to AEs:
 G1: 0% (0/30)
 G2: 30% (9/30)
Attrition: 52%
Differential attrition:  10 percentage points
High
Schulz et al. (2008)
Design: Double-blinded RCT
Setting: Outpatient, multicenter
Country: Multicountry
Funding: Eli Lilly
N = 314
G1 (159): Placebo
G2 (155): Olanzapine (2.5–20 mg/day)
Duration: 12 weeks
Inclusion: Males and females; ages 18–65 years; DSM-IV BPD diagnosis; ZAN-BPD total score of 9
Exclusion: Schizophrenia, bipolar I disorder, bipolar II disorder, delusional disorder, MDD, SUD, PTSD, panic disorder, or OCD; BMI < 17; use of antidepressants, mood stabilizer, or antipsychotic medication within 1 week of randomization; new psychotherapy treatment
Mean age, years (SD):
 G1: 32 (9.6)
 G2: 32 (9.5)
Female: 71%
Race/ethnicity:
 White: 87%
Primary outcome: ZAN-BPD at 12 weeks
No significant differences on ZAN-BPD, SCL-90-R, and MADRS
SDS, GAF, MOAS: data NR
Incidence of AEs:
 G1: 57% (90/159)
 G2: 66% (102/155)
Withdrawal due to AEs:
 G1: 11% (18/159)
 G2: 11% (17/155)
Attrition: 43%
Differential attrition:  10 percentage points
High
Soler et al. (2005)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: Spain
Funding: Eli Lilly
N = 60
G1 (30): DBT plus placebo
G2 (30): DBT plus olanzapine (5–20 mg/day)
Duration: 12 weeks
Inclusion: Females; ages 18–45 years; DSM-IV BPD diagnosis without comorbid, unstable Axis I disorder; CGI-S score  4; not receiving psychotherapy
Exclusion: NR
Mean age, years (SD):
 G1: 26 (5.4)
 G2: 28 (6.3)
Female: 87%
Race/ethnicity: NR
Primary outcome: NR
Significantly greater improvements for G2 than G1 on Ham-D (8.79 vs. 4.87, P = 0.004) and frequency of aggressive behavior (data NR, P = 0.03)
No significant differences on Ham-A, CGI-S, and episodes of suicide attempts and self-injury
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 30%
Differential attrition:  10 percentage points
High
Zanarini and Frankenburg (2001)
Design: Double-blinded RCT
Setting: Outpatients, single center
Country: United States
Funding: Eli Lilly
N = 28
G1 (9): Placebo
G2 (19): Olanzapine (2.5 mg/day)
Duration: 6 months
Inclusion: Females; ages 18–40 years; DSM-IV BPD diagnosis
Exclusion: MDD; previous treatment with olanzapine; currently taking psychotropic medications; actively abusing alcohol or drugs
Mean age, years (SD):
 G1: 26 (4.5)
 G2: 28 (7.7)
Female: 100%
Race/ethnicity:
 White: 71%
 Nonwhite: 29%
Primary outcome: SCL-90 at 6 months
G2 significantly greater improvements than G1 on four domains of SCL-90 (interpersonal sensitivity, anxiety, anger/hostility, paranoia); overall SCL-90 score NR
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/9)
 G2: 16% (3/19)
Attrition: 68%
Differential attrition:
 G1: 89% (8/9)
 G2: 58% (11/19)
High
Zanarini et al. (2011b)
Design: Double-blinded RCT
Setting: Outpatient, multicenter
Country: Multicountry
Funding: Eli Lilly
N = 451
G1 (153): Placebo
G2 (150): Olanzapine (2.5 mg/day)
G3 (148): Olanzapine (5–10 mg/day)
Duration: 12 weeks
Inclusion: Males and females; ages 18–65 years; DSM-IV BPD diagnosis; ZAN-BPD total score  9
Exclusion: Schizophrenia, schizoaffective disorder, schizophreniform disorder, bipolar I disorder, bipolar II disorder, delusional disorder, MDD, SUD within past 3 months, PTSD, panic disorder, or OCD; actively suicidal; BMI < 17; Cluster A personality disorder; new psychotherapy within 3 months prior to visit 1; use of anticholinergic medication as prophylaxis for extrapyramidal symptoms
Mean age, years (SD):
 G1: 34 (11.3)
 G2: 33 (11.2)
 G3: 33 (10.0)
Female: 74%
Race/ethnicity:
 White: 65%
 African descent: 7%
 East/Southeast Asian: 2%
 Western Asian: 0.2%
 Hispanic: 24.6%
 Other origin: 11.1%
Primary outcome: ZAN-BPD at 12 weeks
G3 significantly more effective than G1 on ZAN-BPD (- 8.5 vs. - 6.8, P = 0.01; response: 74% vs. 60%, P = 0.018) and SCL-90-R (- 0.7 vs. - 0.6, P < 0.05)
No significant differences between G1 and G3 on MADRS, GAF, and MOAS
No significant differences between G1 and G2 on most outcome measures
Incidence of AEs:
 G1: 61% (93/153)
 G2: 65% (98/150)
 G3: 67% (99/148)
Withdrawal due to AEs:
 G1: 3% (5/153)
 G2: 3% (5/150)
 G3: 6% (9/148)
Attrition: 35%
Differential attrition:  10 percentage points
Moderate
AE = adverse event; BIS = Barratt Impulsiveness Scale; BMI = body mass index; BPD = borderline personality disorder; CGI = Clinical Global Impression Scale; CGI-BPD = Clinical Global Impression Scale for Borderline Personality Disorder; CGI-S = Clinical Global Impression-Severity; DBT = dialectical behavior therapy; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; ER = extended release; G1 = Group 1; G2 = Group 2; G3 = Group 3; GAF = Global Assessment of Functioning; Ham-A = Hamilton Rating Scale for Anxiety; Ham-D = Hamilton Rating Scale for Depression; MADRS = Montgomery-Åsberg Depression Rating Scale; MDD = major depressive disorder; MDE = major depressive episode; MOAS = Modified Overt Aggression Scale; N = sample size; NR = not reported; OCD = obsessive compulsive disorder; PTSD = posttraumatic stress disorder; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SCL-90 = Symptom Checklist-90; SCL-90-R = Symptom Checklist–90–Revised; SD = standard deviation; SDS = Sheehan Disability Scale; SGA = second-generation antipsychotic; STAXI = State-Trait Anger Expression Inventory; SUD = substance use disorder; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Second-Generation Antipsychotics vs. Antidepressants

Study characteristics and main results of SGAs compared with antidepressants
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Zanarini et al. (2004c)
Design: double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: Eli Lilly
N = 45
G1 (14): Fluoxetine (10–30 mg/day)
G2 (16): Olanzapine (2.5–7.5 mg/day)
G3 (15): Fluoxetine (10–30 mg/day) and olanzapine (2.5–7.5 mg/day)
Duration: 8 weeks
Inclusion: Females; ages 18–40 years; DSM-IV BPD diagnosis; does not meet criteria for current MDD
Exclusion: Current MDD, current or lifetime schizophrenia, schizoaffective disorder, or bipolar disorder; current use of psychotropic medications; medical illness; seizure disorder; substance abuse; acutely suicidal
Mean age, years (SD): 23 (5.7)
Female: 100%
Race/ethnicity:
 White: 80%
Primary outcome: NR
G2 and G3 significantly more effective than G1 on MOAS (19.7 vs. 20.2 vs. 15.4, P = 0.003 for G2 vs. G1, P < 0.001 for G3 vs. G1) at 8 weeks
G2 and G3 significantly more effective than G1 on MADRS (13.6 vs. 11.9 vs. 8.2, P < 0.001 for G2 vs. G1, P = 0.02 for G3 vs. G1) at 8 weeks
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 7% (1/14)
 G2: 0% (0/16)
 G3: 7% (1/15)
Attrition: 7%
Differential attrition:  10 percentage points
Moderate
AE = adverse event; BPD = borderline personality disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; G3 = Group 3; MADRS = Montgomery-Åsberg Depression Scale; MDD = major depressive disorder; MOAS = Modified Overt Aggression Scale; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; SGA = second-generation antipsychotic.

Second-Generation Antipsychotics vs. Second-Generation Antipsychotics

Study characteristics and main results of SGAs compared with SGAs
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Bozzatello et al. (2017)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: Italy
Funding: None
N = 51
G1 (26): Olanzapine (5–10 mg/day)
G2 (25): Asenapine (5–10 mg/day)
Duration: 12 weeks
Inclusion: Ages 18–50 years; DSM-5 BPD diagnosis
Exclusion: Dementia, schizophrenia or other psychotic disorders, bipolar disorders, co-occurring MDE, or substance abuse; past use of psychotropic medications and/or psychotherapy
Mean age, years (SD): 25 (5.3)
Female: 63%
Race/ethnicity: NR
Primary outcome: NR
No significant differences between G1 and G2 on BPDSI, CGI-S, BIS, MOAS, Ham-D, and SHI at 12 weeks
Incidence of AEs (among completers):
 G1: 26% (5/19)
 G2: 19% (4/21)
Withdrawal due to AEs:
 G1: 11% (2/19)
 G2: 10% (2/21)
Attrition: 22%
Differential attrition:  10 percentage points
High
García-Carmona et al. (2021)
Design: Retrospective cohort study
Setting: Outpatient; multicenter
Country: Spain
Funding: None
N = 116
G1 (66): Oral antipsychotics
G2 (50): LAI antipsychotics
Duration: 1–3 months
Inclusion: Age  18 years; DSM-5 BPD diagnosis; treated with an oral or an LAI SGA continuously for > 12 months
Exclusion: institutionalized; intellectual disability or other psychiatric disorder; concomitant use of two LAI antipsychotics, or missing clinical records
Mean age, years (SD):
 G1: 42.4 (1.4)
 G2: 39.4 (1.7)
Female: 46%
Race/ethnicity: NR
Primary outcome: NR
G1 significantly more ED visits than G2 (7.9 vs. 6.2, P = 0.041)
No significant differences in suicidal behavior and hospital admissions
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: NR
Differential attrition: NR
High
AE = adverse event; BIS = Barratt Impulsiveness Scale; BPD = borderline personality disorder; BPDSI = Borderline Personality Disorder Severity Index; CGI-S = Clinical Global Impression-Severity; DSM-5 = Diagnostic and Statistical Manual of Mental Disorders, 5th Edition; ED = emergency department; G1 = Group 1; G2 = Group 2; Ham-D = Hamilton Rating Scale for Depression; LAI = long-acting injectable; MDE = major depressive episode; MOAS = Modified Overt Aggression Scale; N = sample size; NR = not reported; SD = standard deviation; SE = standard error; SGA = second-generation antipsychotic; SHI = Self-Harm Inventory.

Anticonvulsants vs. Placebo

Study characteristics and main results of anticonvulsants compared with placebo
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Crawford et al. (2018); LABILE
Design: Double-blinded RCT
Setting: Outpatient, multicenter
Country: United Kingdom
Funding: NIHR
N = 276
G1 (139): Placebo
G2 (137): Lamotrigine (200 mg/day)
Duration: 52 weeks
Inclusion: DSM-IV BPD diagnosis
Exclusion: Met diagnostic criteria for bipolar disorder (type I or II) or psychotic disorder; history of liver or kidney impairment
Mean age, years (SD):
 G1: 36 (11.0)
 G2: 36 (11.0)
Female: 75%
Race/ethnicity:
 White: 89%
 Black: 4%
 Asian: 1%
 Other: 6%
Primary outcome: ZAN-BPD at 52 weeks
No significant differences on ZAN-BPD, SHI, SFQ, and EQ-5D-3L
Incidence of AEs:
 G1: 67% (93/139)
 G2: 56% (77/137)
Withdrawal due to AEs:
 G1: 1% (1/139)
 G2: 4% (4/137)
Attrition: 29%
Differential attrition: < 10 percentage points
Moderate
Frankenburg and Zanarini (2002)
Design: Double-blinded RCT
Setting: Community recruitment with advertisements
Country: United States
Funding: Abbott Laboratories
N = 30
G1 (10): Placebo
G2 (20): Divalproex sodium (250 mg/day)
Duration: 24 weeks
Inclusion: Females; ages 18–40 years; DIB-R and DSM-IV BPD and bipolar II disorder diagnoses
Exclusion: Formerly treated with divalproex sodium; medically ill; seizure disorder; current substance abuse; current criteria for an MDE or hypomanic episode; current or lifetime criteria for schizophrenia, schizoaffective disorder, psychotic disorder, or bipolar I disorder
Mean age, years (SD):
 G1: 26 (7.3)
 G2: 27 (7.4)
Female: 100%
Race/ethnicity:
 White: 67%
 Black: 10%
 Hispanic: 13%
 Biracial: 7%
Primary outcome: MOAS, SCL-90-R (subscales on anger, interpersonal hostility, depression) at 24 weeks
G2 significantly more effective than G1 on MOAS (3.0 vs. 1.9, P = 0.03) and SCL-90-R subscales on anger/hostility (0.8 vs. 0.6, P = 0.01) and interpersonal sensitivity (0.8 vs. 0.4, P = 0.04)
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 30% (3/10)
 G2: 5% (1/20)
Attrition: 63%
Differential attrition: < 10 percentage points
High
Hollander et al. (2001)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: Abbott Laboratories, NIMH
N = 16
G1 (4): Placebo
G2 (12): Divalproex sodium (250 mg/day)
Duration: 10 weeks
Inclusion: DSM-IV BPD diagnosis
Exclusion: Medical or neurological illness; psychotic disorders, substance abuse, bipolar disorder I or II, or MDD; suicidal ideation
Mean age, years (SD): NR
Female: 52%
Race/ethnicity:
 White: 67%
 Black: 14%
 Hispanic: 19%
Primary outcome: NR
No significant differences on CGI-I, GAS, MOAS, and AQ
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/4)
 G2: 0% (0/12)
Attrition: 63%
Differential attrition:
 G1: 100% (4/4)
 G2: 50% (6/12)
High
Loew et al. (2006)
Design: Double-blinded RCT
Setting: Single center or multicenter
Country: Germany and Austria
Funding: None
N = 56
G1 (28): Placebo
G2 (28): Topiramate (200 mg/day)
Duration: 10 weeks
Inclusion: Females; ages 18–35 years; DSM-IV BPD diagnosis
Exclusion: Schizophrenia; current use of psychotropic medication or psychotherapy; suicidal; substance abuse; severe somatic illness
Mean age, years (SD):
 G1: 26 (5.7)
 G2: 25 (5.3)
Female: 100%
Race/ethnicity: NR
Primary outcome: SCL-90-R, SF-36, and IIP at 10 weeks
G2 significantly more effective than G1 on SCL-90-R (7.4 vs. 1.8, P < 0.001), SF-36 (data NR, P < 0.01), and IIP (data NR)
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 7%
Differential attrition: < 10 percentage points
Low
Moen et al. (2012)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: Abbott
N = 15
G1 (5): Placebo
G2 (10): Divalproex sodium (NR)
Duration: 12 weeks
Inclusion: Ages 21–55 years; DSM-IV BPD diagnosis; score  150 on SCL-90; score  5 on SCID-II
Exclusion: Current or past history of bipolar disorder, schizophrenia, or MDD with psychotic features; current psychotropic medication; acutely suicidal; SUD; seizure disorder and/or anticonvulsant medications
Mean age, years (range)
 G1: 37 (22–51)
 G2: 34 (23–45)
Female: 80%
Race/ethnicity:
 White: 80%
 Black: 7%
 Hispanic: 7%
 Mixed: 7%
Primary outcome: NR
No significant differences on SCL-90, BIS, and BEST
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 40%
Differential attrition: < 10 percentage points
High
Nickel et al. (2004)
Design: Double-blinded RCT
Setting: Community recruitment
Country: Germany
Funding: None
N = 31
G1 (10): Placebo (50 mg/day)
G2 (21): Topiramate (250 mg/day)
Duration: 8 weeks
Inclusion: Females; ages 20–35 years; DSM-IV BPD diagnosis
Exclusion: Current schizophrenia, MDD, or bipolar disorder; current use of psychotropic medication or psychotherapy; somatically ill; actively suicidal; substance abuse
Mean age, years (SD):
 G1: 27 (NR)
 G2: 26 (NR)
Female: 100%
Race/ethnicity: NR
Primary outcome: STAXI at 8 weeks
G2 significantly more effective than G1 on four out of five subscales on STAXI (P values from 0.05 to 0.01); no significant improvement on subscale assessing tendency to repress anger
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/10)
 G2: 0% (0/21)
Attrition: 6%
Differential attrition: < 10 percentage points
Moderate
Nickel et al. (2005)
Design: Double-blinded RCT
Setting: Outpatient recruitment and community advertisement
Country: Germany
Funding: None
N = 44
G1 (22): Placebo
G2 (22): Topiramate (250 mg/day)
Duration: 8 weeks
Inclusion: Males; age > 18 years; DSM-IV BPD diagnosis
Exclusion: Acute psychosis, severe MDD or bipolar disorder; current use of psychotropic medication or psychotherapy; somatically ill; actively suicidal; SUD
Mean age, years (SD):
 G1: 29 (NR)
 G2: 30 (NR)
Female: 0%
Race/ethnicity: NR
Primary outcome: STAXI at 8 weeks
G2 significantly more effective than G1 on four out of five subscales on STAXI (P values from 0.05 to 0.01); no significant improvement on subscale assessing tendency to repress anger
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 0% (0/22)
 G2: 0% (0/22)
Attrition: 5%
Differential attrition: < 10 percentage points
Moderate
Reich et al. (2009)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: GlaxoSmithKline
N = 28
G1 (13): Placebo
G2 (15): Lamotrigine (50–275 mg/day)
Duration: 12 weeks
Inclusion: DSM-IV BPD diagnosis; score  8 on DIB-R; “serious” score on affective instability item of ZAN-BPD; score  14 on ALS
Exclusion: Dementia, psychiatric disorder, bipolar disorder, psychotic disorder, or SUD; currently hospitalized; previous treatment with lamotrigine or psychotherapy; active suicidal or homicidal ideation
Mean age, years (SD):
 G1: 35 (9.7)
 G2: 28 (9.5)
Female: 89%
Race/ethnicity:
 White: 89%
Primary outcome: ALS, affective instability item of ZAN-BPD at 12 weeks
G2 significantly greater improvements than G1 on ALS (0.71 vs. 0.40, P = 0.012) and affective lability of ZAN-BPD (1.5 vs. 1.1, P = 0.043)
No significant difference on ZAN-BPD
Incidence of AEs:
 G1: 31% (4/13)
 G2: 40% (6/15)
Withdrawal due to AEs:
 G1: 0% (0/13)
 G2: 0% (3/15)
Attrition: 39%
Differential attrition: < 10 percentage points
High
Tritt et al. (2005)
Design: Double-blinded RCT
Setting: single center or multicenter
Country: Germany and Austria
Funding: None
N = 27
G1 (9): Placebo
G2 (18): Lamotrigine (200 mg/day)
Duration: 8 weeks
Inclusion: Females; ages 20–40 years; DSM-IV BPD diagnosis
Exclusion: Schizophrenia, MDD, or bipolar disorder; current use of psychotropic medication or psychotherapy; somatically ill; actively suicidal; substance abuse
Mean age, years (SD):
 G1: 29 (NR)
 G2: 29 (NR)
Female: 100%
Race/ethnicity: NR
Primary outcome: STAXI at 8 weeks
G2 significantly more effective than G1 on all five subscales of STAXI (P values from < 0.05 to < 0.01; overall STAXI score NR)
G2 improved more than G1 with respect to all STAXI scales on assessments after 8 weeks of treatment
Incidence of AEs: NR
Withdrawal due to AEs:
 G1: 11% (1/9)
 G2: 6% (1/18)
Attrition: 11%
Differential attrition:  10 percentage points
Low
AE = adverse event; ALS = Affective Liability Scale; AQ = Aggression Questionnaire; BEST = Borderline Evaluation of Severity Over Time; BIS = Barratt Impulsiveness Scale; BPD = borderline personality disorder; CGI-I = Clinical Global Impression-Improvement; DIB-R = Diagnostic Interview for Borderlines-Revised; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; EQ-5D-3L = European Quality of Life–5 Dimension-3 level version; G1 = Group 1; G2 = Group 2; GAS = Global Assessment Scale; IIP = Inventory of Interpersonal Problems; LABILE = Lamotrigine and Borderline Personality Disorder: Investigating Long-Term Effects; MDD =  major depressive disorder; MDE = major depressive disorder; MOAS = Modified Overt Aggression Scale; NIHR = National Institute for Health Research; NIMH = National Institute of Mental Health; N = sample size; NR = not reported; RCT = randomized controlled trial; SCID-II = Structured Clinical Interview for DSM-IV Axis II Disorders; SCL-90 = Symptom Checklist-90; SCL-90-R = Symptom Checklist–90–Revised; SD = standard deviation; SF-36 = Short Form Survey; SFQ = Social Functioning Questionnaire; SHI = Self-Harm Inventory; STAXI = State-Trait Anger Expression Inventory; SUD = substance use disorder; ZAN-BPD = Zanarini Rating Scale for Borderline Personality Disorder.

Antidepressants vs. Placebo

Study characteristics and main results of antidepressants compared with placebo
Author (year) and/or trial nameStudy characteristicsParticipants, N; interventions; durationStudy population, including main inclusion and exclusion criteriaSample demographicsPrimary outcome; main results; attritionRisk of bias
Simpson et al. (2004)
Design: Double-blinded RCT
Setting: Outpatient, single center
Country: United States
Funding: Eli Lilly
N = 25
G1 (13): Placebo
G2 (12):
Fluoxetine (40 mg/day)
Duration: 12 weeks
Inclusion: Admission to Women’s Partial Program; DSM-IV BPD diagnosis
Exclusion: SUD; seizure disorder; unstable medical conditions; history of schizophrenia or bipolar disorder; previous adequate trial of fluoxetine
Mean age, years (SD): 35 (10.1)
Female: 100%
Race/ethnicity:
 White: 72%
 Black: 20%
 Native American: 8%
Primary outcome: NR
When corrected for multiple testing, no significant differences between G1 and G2 on STAXI, MOAS, or GAF at mean of 10 weeks
Incidence of AEs: NR
Withdrawal due to AEs: NR
Attrition: 20%
Differential attrition:  10 percentage points
High
AE = adverse event; BPD = borderline personality disorder; DSM-IV = Diagnostic and Statistical Manual of Mental Disorders, 4th Edition; G1 = Group 1; G2 = Group 2; GAF = Global Assessment of Functioning; MOAS = Modified Overt Aggression Scale; N = sample size; NR = not reported; RCT = randomized controlled trial; SD = standard deviation; STAXI = State-Trait Anger Expression Inventory; SUD = substance use disorder.

Information & Authors

Information

Published In

Go to The American Psychiatric Association Practice Guideline for the Treatment of                 Patients With Borderline Personality Disorder
The American Psychiatric Association Practice Guideline for the Treatment of Patients With Borderline Personality Disorder
November 2024
©American Psychiatric Association Publishing

Authors

Metrics & Citations

Metrics

Citations

If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.

For more information or tips please see 'Downloading to a citation manager' in the Help menu.

Format
Citation style
Style
Copy to clipboard

View Options

View options

Login options

Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.

Personal login Institutional Login Open Athens login

Not a subscriber?

Subscribe Now / Learn More

PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.

Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).

Media

Figures

Other

Tables

Share

Share

Share article link

Share