Docstoc

Binks2006-01-01

Document Sample
Binks2006-01-01 Powered By Docstoc
					Pharmacological interventions for people with borderline personality disorder (Review)
Binks CA, Fenton M, McCarthy L, Lee T, Adams CE, Duggan C

This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2007, Issue 3 http://www.thecochranelibrary.com

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

1

TABLE OF CONTENTS ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CRITERIA FOR CONSIDERING STUDIES FOR THIS REVIEW . . . . . . . . . . . . . . . . . . SEARCH METHODS FOR IDENTIFICATION OF STUDIES . . . . . . . . . . . . . . . . . . . METHODS OF THE REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DESCRIPTION OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . METHODOLOGICAL QUALITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . POTENTIAL CONFLICT OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characteristics of included studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Characteristics of excluded studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Comparison 01. ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO . . . . . . . . . . Comparison 02. ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC . . . . . . . . . . . . . Comparison 03. ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO . . . . . . Comparison 04. ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC . . . Comparison 05. ANTIPSYCHOTICS vs PLACEBO . . . . . . . . . . . . . . . . . . . . . . Comparison 06. ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE . . . . . . Comparison 07. ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL . . . . . . Comparison 08. MOOD STABILISERS vs PLACEBO . . . . . . . . . . . . . . . . . . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . COVER SHEET . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . GRAPHS AND OTHER TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 01.01. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 01 Global state: 1. Mean score - by 6 months (GAS, higher scores = better) . . . . . . . . . . . . . . Analysis 01.02. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 02 Global state: 2. Mean symptom score - by 6 months (skewed data) . . . . . . . . . . . . . . . Analysis 01.03. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data) . . . . Analysis 01.04. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 04 Mental state: 1. Anger - not improved by 6 months (POMS not 20% over baseline score. high score=poor) . Analysis 01.05. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 05 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) . . . . . . Analysis 01.06. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 06 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . Analysis 01.07. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 07 Mental state: 3a. Depression - not improved - by 6 months . . . . . . . . . . . . . . . . . . Analysis 01.08. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . Analysis 01.09. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 09 Mental state: 4. Mood disturbance - not improved by 6 months (POMS, not 20% over baseline score) . . . Analysis 01.10. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) . . . . . . . .
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

1 2 2 3 3 4 8 10 13 14 18 19 20 20 20 20 23 23 29 30 30 31 32 33 34 35 36 36 36 36 38 38 38 38 39 40 40 41 41 42 42
i

Analysis 01.11. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 11 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed) . . . . . . . . . . Analysis 01.12. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 12 Behaviour: 1. Attempting suicide within 6 months . . . . . . . . . . . . . . . . . . . . . Analysis 01.13. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 13 Behaviour: 2a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) . . . Analysis 01.14. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 14 Behaviour: 2b. Hostility scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . Analysis 01.15. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 15 Behaviour: 3a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) . . . . Analysis 01.16. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 16 Behaviour: 3b. Impulsiveness - traits by 6 months . . . . . . . . . . . . . . . . . . . . . Analysis 01.17. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 17 Leaving the study early: Completed <2 weeks of treatment . . . . . . . . . . . . . . . . . . Analysis 02.01. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 01 Global state: 1. Global functioning score - by 6 months (GAS, higher scores = better) . . . . . . . . . . . . Analysis 02.02. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 02 Global state: 2. Mean symptom score - by 6 months (skewed data) . . . . . . . . . . . . . . . . . . Analysis 02.03. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data) . . . . . . . Analysis 02.04. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 04 Mental state: 1a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) . . . . . . . . . Analysis 02.05. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 05 Mental state: 1b. Anxiety scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . Analysis 02.06. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 06 Mental state: 2. Depression scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . Analysis 02.07. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 07 Mental state: 3. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) . . . . . . . . . . Analysis 02.08. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 08 Mental state: 4. Other symptoms - by 6 months (high scores = poor, data skewed) . . . . . . . . . . . . . Analysis 02.09. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 09 Behaviour: 1a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) . . . . . . . Analysis 02.10. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 10 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . . Analysis 02.11. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 11 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) . . . . . . . . Analysis 02.12. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 12 Behaviour: 2b. Impulsiveness - traits by 6 months . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 02.13. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 13 Leaving the study early: Completed <2 weeks of treatment . . . . . . . . . . . . . . . . . . . . . . . Analysis 03.01. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) Analysis 03.02. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) . . . . Analysis 03.03. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) . . . . . Analysis 03.04. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data) Analysis 03.05. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) Analysis 03.06. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) . .
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

43 43 44 44 44 45 46 47 47 47 48 48 48 49 50 50 50 55 56 56 56 57 57 57 58 58

ii

Analysis 03.07. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90 subscale higher scores = poor, skewed data) Analysis 03.08. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) . . . . Analysis 03.09. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 09 Mental state: 4. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) . . . Analysis 03.10. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 10 Mental state: 5. Other symptoms - by 6 months (high scores = poor, data skewed) . . . . . Analysis 03.11. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 11 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor) . . . . . . . . Analysis 03.12. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 12 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) . . . . . . Analysis 03.13. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 13 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) Analysis 03.14. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 14 Behaviour: 2b. Impulsiveness - traits by 6 months . . . . . . . . . . . . . . . . . Analysis 03.15. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 15 Appitite and weight (higher scores = poor, skewed data) . . . . . . . . . . . . . . . Analysis 04.01. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.02. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.03. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.04. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.05. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.06. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.07. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90, higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.08. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 08 Mental state: 3a. Depression scores - by 6 months (HAM-D-17, high score= poor)) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.09. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 09 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.10. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 10 Mental state: 4. Dysphoria - hysteroid dysphoria - by 6 months (HDQ, higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.11. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 11 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

58 59 59 60 61 61 61 62 62 62

63

64

64

64

65

65

65

66

66

66

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

iii

Analysis 04.12. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 12 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.13. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 13 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor) Analysis 04.14. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 14 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) Analysis 04.15. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 15 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 04.16. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 16 Behaviour: 2b. Impulsiveness - traits by 6 months . . . . . . . . . Analysis 04.17. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 17 Appetite and weight (higher scores = poor, skewed data) . . . . . . . Analysis 05.01. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) . . . . . . . . . . . . . Analysis 05.02. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) . . . . . . . . . . . . . . . . . . . . . Analysis 05.03. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) . . . . . . . . . . . . . . . . . . . . . . Analysis 05.04. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data) . . . . . . . . . . . . . . . . Analysis 05.05. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) . . . . . . . . . . . . . . . . Analysis 05.06. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 06 Mental state: 2a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) . . . . . . . . . . . . . . . . . . Analysis 05.07. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 07 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . Analysis 05.08. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 08 Mental state: 3a. Depression scores by 6 months (HAM-D-17 higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . Analysis 05.09. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 09 Mental state: 3b. Depression scores by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . Analysis 05.10. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) . . . . . . . . . . . . . . . . . Analysis 05.11. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 11 Mental state: 6. Other symptoms by 6 months (high scores = poor, data skewed) . . . . . . . . . . . . . . . . . . . . . . Analysis 05.12. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 12 Behaviour: 1a. Hostility - by 6 months (higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 05.13. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 13 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . Analysis 05.14. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 14 Behaviour: 2a. Impulsiveness behaviour by 6 months (WSIAP, higher scores = poor, data skewed) . . . . . . . . . . . . . . . Analysis 05.15. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 15 Behaviour: 2b. Impulsiveness - traits by 6 months . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 05.16. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 16 Leaving the study early . . . . Analysis 05.17. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 17 Weight gain score (ADI subscale, higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 06.01. Comparison 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE, Outcome 01 Leaving the study early - before 3 weeks of treatment . . . . . . . . . . . . . . . . Analysis 06.02. Comparison 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE, Outcome 02 Adverse effects: One or more adverse effects reported . . . . . . . . . . . . . . . . Analysis 07.01. Comparison 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL, Outcome 01 Leaving the study early - by 6 weeks . . . . . . . . . . . . . . . . . . . . .
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

67

68 68 69

69 69 70 70 71 71 71 72 72 73 73 74 75 76 76 77 78 78 78 79 79 80
iv

Analysis 07.02. Comparison 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL, Outcome 02 Adverse effects: Adverse reactions to medication by 6 months . . . . . . . . . . . . . Analysis 08.01. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 01 Leaving the study early . . . Analysis 08.02. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 02 Mental state: 1. Not responsive by 6 months (> 2 on CGI-I) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 08.03. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 03 Mental state: 2. Depression scores - by 6 months (BDI, higher scores = poor, skewed data) . . . . . . . . . . . . . . . . . . . Analysis 08.04. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 04 Behaviour: 1. Acting out by 6 months (AOS behavioural dyscontrol) . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 08.05. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 05 Behaviour: 2a. Aggression score by 6 months (AQ, higher scores = poor) . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 08.06. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 06 Behaviour: 2b. Aggression scores by 6 months (higher scores = poor, skewed data). . . . . . . . . . . . . . . . . . . . . . . Analysis 08.07. Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 07 Behaviour: 3. Not improved behavioural dyscontrol (AOS) . . . . . . . . . . . . . . . . . . . . . . . . . . . .

80 81 81 81 82 82 82 83

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

v

Pharmacological interventions for people with borderline personality disorder (Review)
Binks CA, Fenton M, McCarthy L, Lee T, Adams CE, Duggan C

This record should be cited as: Binks CA, Fenton M, McCarthy L, Lee T, Adams CE, Duggan C. Pharmacological interventions for people with borderline personality disorder. Cochrane Database of Systematic Reviews 2006, Issue 1. Art. No.: CD005653. DOI: 10.1002/14651858.CD005653. This version first published online: 25 January 2006 in Issue 1, 2006. Date of most recent substantive amendment: 02 September 2005

ABSTRACT Background Borderline Personality Disorder (BPD) is prevalent (2% in the general population, 20% among psychiatry in-patients) and has a major impact on health facilities as those affected often present in crisis but then make poor use of further attempts to help them. Objectives To evaluate the effects of pharmacological interventions for people with borderline personality disorder. Search strategy We conducted a systematic search of 26 specialist and general bibliographic databases (October 2002) and searched relevant reference lists for further trials. Selection criteria We included all randomised clinical trials comparing any psychoactive drugs with any other treatment for people with borderline personality disorder. Data collection and analysis We independently selected, quality assessed and data extracted studies. For binary outcomes we calculated a standard estimation of the risk ratio (RR), its 95% confidence interval (CI), and where possible the number need to help/harm (NNT/H). For continuous outcomes, endpoint data were preferred to change data. Non-skewed data from valid scales were synthesised using a weighted mean difference (WMD). Main results We found ten small (total n=554), short, randomised studies involving eight comparisons from which we could extract usable data. Studies comparing antidepressants with placebo were small (total n=79, 2 RCTs) but for ratings of anger fluoxetine may offer some improvement for those on antidepressant therapy over placebo (n=22, 1 RCT, RR anger not improved 0.30 CI 0.10 to 0.85, NNT 2 CI 2 to 9). The one small study investigating the important outcome of attempted suicide found no difference between mianserin and placebo (n=38, 1 RCT, RR 0.82 CI 0.44 to 1.54). Haloperidol may be better than antidepressants for symptoms of hostility and psychotism. There were few differences between MAOIs and placebo except that people given MAOIs were less hostile (n=62, 1 RCT, MD -9.19 CI -16.12 to -2.26). Although some ratings were statistically significant the comparison of MAOIs with antipsychotics did not show convincing differences. Antipsychotics may effect some mental state symptoms more effectively than placebo but results are difficult to interpret clinically and there is little evidence of advantage of one antipsychotic over another. Finally mood stabilisers such as divalporex may help mental state (n=16, 1 RCT, RR no improvement in mental state 0.58 CI 0.36 to 0.94, NNT 3 CI 2 to 17) but data are far from conclusive. Authors’ conclusions Pharmacological treatment of people with BPD is not based on good evidence from trials and it is arguable that future use of medication should be from within randomised trials. Current trials suggest that the positive effect of antidepressants, in particular, could be
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 1

considerable. Well designed, conducted and reported clinically meaningful trials are possible and needed with, perhaps, the question of antidepressant versus placebo being addressed first.

PLAIN LANGUAGE SUMMARY People with Borderline Personality Disorder (BPD) tend to have an unstable sense of self with difficulty in interpersonal relationships, impulsiveness, and unstable mood. It is relatively common with rates of about 2% in the general population but it is present in 20% among psychiatry in-patients. Self-harming behaviour is common in this group of people and BPD has a major impact on health facilities as those affected often present in crisis, but then make poor use of further attempts to help them. Evidence of the effects of commonly prescribed drugs is poor but not without some areas of hope. Antidepressants are often employed in routine care and trial evidence does suggest there may be some positive effects. This is an important, urgent area for research.

BACKGROUND Borderline personality disorder (BPD) was a condition recognised in the 19th century as existing in the “borderland” between psychosis and neurosis (Stone 1993). Subsequent psychoanalytic contributions (especially that of Kernberg 1975) have reaffirmed this distinction emphasising that sense of identity is weak but nevertheless the capacity to test reality remains. There are three main clinical components to this disorder (a) an unstable sense of self with difficulty in interpersonal relationships; (b) impulsivness; and (c) affective instability. Some believe that it is a variant of affective disorders (Coid 1993). Despite its controversial nature, borderline personality disorder is the focus of great interest with more reports and books appearing on this disorder than any other personality problems (Stone 1993) and its importance stems from the significant impact it has on mental health services. The Diagnostic and Statistical Manual of Mental Disorders-IV (DSM-IV) definition (APA 1994) comprises nine criteria that cover the above features with a definite diagnosis requiring that five are met, and probable diagnosis requiring four. The competing International Classification of Diseases-10 (ICD-10) classification system refers to the condition of Emotionally Unstable Personality Disorder (F60.3 of which there is an impulsive type (F60.30) and a borderline type (F60.31) (ICD-10 1992). The latter essentially overlaps with the DSM-IV definition. A significant problem is that with this type of polythetic definition it is possible for two people to satisfy the criteria and yet have very different personalities. This heterogeneity is a major problem in assessing the impact of an intervention. The prevalence of borderline personality disorder is about 2% in the general population but it is present in 20% among psychiatry in-patients (DSM-IV, APA 1994) and is predominantly diagnosed in women (75%). There are particular problems in its diagnosis in adolescents and young adults where existential dilemmas may be mistakenly be classified as having BPD (DSM-IV). BPD commonly occurs with mood disorders, substance misuse, eating dis-

orders, post traumatic stress disorder (PTSD) and, of course, is also associated with other personality disorders. BPD has a high rate of suicide when associated with mood disorders or alcoholism (Stone 1990).

There is compelling evidence that, while the short to mediumterm outcome of BPD is poor, being similar to that of schizophrenia, longer term follow-ups show a more favourable course (Stone 1990, Paris 1987, McGlashan 1986, Plakum 1985). These longterm follow-up studies found that almost identical results despite differences in the intensity of treatment and socio-demographic status. Despite the suggestion of a favourable general course, the majority of people with BPD still had significant levels of symptoms and disability irrespective of the length of follow-up (Perry 1993). Of the nine studies that examined the rate suicide in BPD, there was an average of about 6% (range of 3-9%) by approximately 7 years.

The direct costs of BPD are considerable in that many people so affected make major demands on health professionals. These are often so intense that the professional eventually burns out and the same cycle is repeated with another. Hence, it has been suggested that a team, rather than an individual manages, people with BPD (NIMHE 2003). The problem of deliberate self-harm is a particular problem in this group (Linehan 1997).

In summary, BPD is a relatively common personality disorder that has been extensively studied. It has a major impact on health facilities as those affected often present in crisis but then make poor use of further attempts to help them. Its long-term course leads to improvement but people continue to have considerable problems. The polythetic nature of the diagnosis is likely to lead to heterogeneity and therefore make it difficult to assess treatment efficacy.
2

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

OBJECTIVES To evaluate the effects of pharmacological interventions for people with borderline personality disorder.

locked general hospital ward), non-secure hospital, prison (category A-D) and community. Types of intervention People with personality disorders may respond to pharmacological interventions that target both their state and trait symptoms. Drug treatments that target the cognitive-perceptual, affective, impulsive-behavioural and anxious-fearful domains of personality disorder need to be evaluated (Soloff 1998). To this end the following classes of pharmacological interventions as defined by the British National Formulary 2003 (BNF 2003) were included in the review: • Hypnotics, anxiolytics and barbiturates • Antipsychotic drugs (including depot injections), antimanic or mood stabilising drugs • Antidepressant drugs; tricyclic and related, monoamine-oxidase inhibitors (MOAIs), SSRIs and related and other antidepressant drugs • Central nervous system stimulants • Antiepileptics • Drugs used in essential tremor, chorea, tics and related disorders • Drugs used in substance dependence We would have included combination drug interventions studies. Types of outcome measures 1. Death - Sudden and unexpected death and natural causes 2. Global state 2.1 Relapse* 2.2 Time to relapse 2.3 No clinically important change in global state 2.4 Not any change in global state 2.5 Average endpoint global state score 2.6 Average change in global state scores 3. Behaviour 3.1 General behaviour 3.1.1 No clinically important change in general behaviour 3.1.2 Not any change in general behaviour 3.1.3 Average endpoint general behaviour score 3.1.4 Average change in general behaviour scores 3.1.5 Compulsory administration of treatment** 3.1.6 Use of further doses of medication 3.2 Specific behaviours 3.2.1 Self-harm, including suicide 3.2.2 Injury to others* 3.2.3 Aggression 3.2.3.1 No clinically important change in aggression* 3.2.3.2 Not any change in aggression
3

CRITERIA FOR CONSIDERING STUDIES FOR THIS REVIEW Types of studies All relevant randomised controlled trials with or without blinding. We excluded quasi-randomised trials, such as those where allocation was undertaken on surname. We included trials described as double blind, where randomisation was implied, and included these in a sensitivity analysis. If there were no substantive differences within primary outcomes when these ’implied randomisation’ studies were added, then we included these in the final analysis. However, if there was a substantive difference then only clearly randomised trials were used. Types of participants Adults (over the age of 18 years) with either: 1. A formal diagnosis of BPD as described by DSM. 2. A formal diagnosis of BPD diagnosed by criteria other than DSM but which nonetheless satisfy three or more of the DSM criteria set out below. 3. Have no formal diagnosis of BPD but nonetheless satisfy three or more DSM criteria. Diagnostic Statistical Manual, version IV, criteria for borderline personality disorder (APA 1994) • Fails to conform to social norms (e.g. repeatedly performing acts that are grounds for arrest) • Is deceitful, lies or cons other people for personal profit or pleasure • s impulsive and fails to plan ahead • Is irritable or aggressive and is often involved in physical fights or assaults • Has a reckless disregard for the safety of others • Is consistently irresponsible and fails to hold down work or honour financial obligations • Shows no remorse after hurting, mistreating or stealing from others People with co-morbid mental health problems other than the major functional mental illnesses (i.e. schizophrenia, schizoaffective disorder, schizotypal or bipolar disorder) are included in the review. We recorded the type of treatment setting according to the following locations: secure hospital (high, medium, low security,

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

3.2.3.3 Average endpoint aggression score 3.2.3.4 Average change in aggression scores 3.2.4 Self-care 3.2.4.1 No clinically important change in self-care 3.2.4.2 Not any change in self-care 3.2.4.3 Average endpoint self-care score 3.2.4.4 Average change in self-care scores 3.2.5 Compliance 3.2.5.1 No clinically important change in compliance 3.2.5.2 Not any change in compliance 3.2.5.3 Average endpoint compliance score 3.2.5.4 Average change in compliance scores 4. Mental state 4.1 General mental state 4.1.1 No clinically important change in general mental state* 4.1.2 Not any change in general mental state 4.1.3 Average endpoint general mental state score 4.1.4 Average change in general mental state scores 5. Engagement with services 5.1 No clinically important engagement 5.2 Not any engagement 5.3 Average endpoint engagement score 5.4 Average change in engagement scores 6. Adverse effects 6.1 No clinically important general adverse effects 6.2 Not any general adverse effects 6.3 Average endpoint general adverse effect score 6.4 Average change in general adverse effect scores 6.5 No clinically important change in specific adverse effects 6.6 Not any change in specific adverse effects 6.7 Average endpoint specific adverse effects 6.8 Average change in specific adverse effects 7. Prison and service outcomes 7.1 Treatment of people in the community 7.2 Duration of treatment programme 7.3 Changes in services provided by through care/probation teams 7.4 Changes of level of supervision by staff/police (re sex offender registration etc.) 8. Satisfaction with treatment 8.1 Recipient of treatment not satisfied with treatment** 8.2 Recipient of treatment average satisfaction score 8.3 Recipient of treatment average change in satisfaction scores 8.4 Informal treatment provider not satisfied with treatment* 8.5 Informal treatment providers’ average satisfaction score 8.6 Informal treatment provider’ average change in satisfaction scores 8.7 Professional providers not satisfied with treatment 8.8 Professional providers’ average satisfaction score 8.9 Professional providers’ average change in satisfaction scores 9. Acceptance of treatment

9.1 Not accepting treatment** 9.2 Average endpoint acceptance score 9.3 Average change in acceptance score 10. Leaving the study early 10.1 For specific reasons (release, parole, move establishment, changes in security status (for example, changes from HMP Category B to Category C levels) 10.2 For general reasons 11. Quality of life 11.1 No clinically important change in quality of life 11.2 Not any change in quality of life 11.3 Average endpoint quality of life score 11.4 Average change in quality of life scores 11.5 No clinically important change in specific aspects of quality of life 11.6 Not any change in specific aspects of quality of life 11.7 Average endpoint specific aspects of quality of life 11.8 Average change in specific aspects of quality of life 12. Recidivism 12.1 Time before re-offence 12.2 Non-sexual violent offence 12.3 Any offence 13. Substance Use 13.1 Change in illicit drug use or in abuse of prescribed drugs 13.2 Change in alcohol use 14. Changes in employment status 14.1 Gaining employment 14.2 Retaining employment 14.3 Losing employment 15 Economic outcomes 15.1 Direct costs 15.2 Indirect costs We divided outcomes into immediate (within 6 months), short term (>6 months to 24 months) and medium term (>24 months to 5 years) and long term (beyond 5 years). N.B. Trials that provided data for this review reported outcomes for intervals. * Primary outcomes for immediate and short term. ** Primary outcomes for medium term.

SEARCH METHODS FOR IDENTIFICATION OF STUDIES See: Cochrane Developmental, Psychosocial and Learning Problems Group methods used in reviews. 1. Electronic searches
4

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

1.1 ASSIA: we searched Applied Social Sciences Index and Abstracts (1987-October 2002) using the phrase: (de=((antisocial personality disorder) OR (avoidant personality disorders) OR (borderline personality disorder) OR (dependent personality) OR (depressive personality disorders) OR (gender identity disorder) OR (histrionic personality disorder) OR (multi-impulsive personality disorder) OR (multiple personality disorder) OR (narcissistic personality disorder) OR (passiveaggressive personality disorder) OR (sadistic personality disorder) OR (schizotypal personality disorders) OR (self defeating personality disorder) OR (antisocial behaviour))) OR (((parano* NEAR person*) OR ((asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath*) NEAR person*)or (psychopath OR sociopath OR (moral NEAR insanity) OR dissocial)) OR (diagnostic within 2 statistical manual iii) OR (diagnostic within 2 statistical manual iv) OR (diagnostic within 2 statistical manual ii)) and (ab=(random*) OR ti=(random*) OR de=(randomi?ed controlled trials) OR ab=(double* blind*)or ti=(double* blind*)or de=(double blind studies) OR (single* NEAR blind*)) 1.2 BHI: we searched British Humanities Index (1962 to October 2002) using the phrase: (de=((antisocial behaviour) OR (psychopaths)) OR ((personality NEAR disorder*) OR (gender NEAR identity)) OR ((sadistic OR schizotypal OR selfdefeating OR borderline OR avoidant OR dependent OR depressive OR histrionic OR multi-impulsive OR multiple OR narcissistic OR passive-aggressive) NEAR (person*)) OR ((parano* NEAR person*) OR (asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath*) NEAR (person*)) OR (psychopath OR sociopath OR (moral NEAR insanity) OR dissocial) OR (diagnostic NEAR statistical NEAR manual*)) and (ab=(random*) OR ti= (random*) OR de=(random) OR ab=(double* blind*) OR ti= (double* blind*)or de=(double* blind*) OR (single* NEAR blind*)) 1.3 BIOSIS: we searched BIOSIS (1985 to October2002) using the phrase: ((al: ((randomi* OR crossover OR random-assignment) OR ((singl* OR doubl* OR tripl* OR trebl*) and (mask* OR blind*)))) and (((((((histrionic OR multi-impulsive OR multiple OR narcissistic OR passive-aggressive)) OR ((psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR selfdefeating OR borderline OR avoidant OR dependent OR depressive))))) and (person*))) OR (((((((((moral and insanity) OR (moral and insanity)) OR ((asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath*) and person*)) OR (self and defeating)) OR (parano* and person*)) OR (gender and identity)) OR (personality and disorder)) OR (antisocial and behaviour))))))

1.4 C2-SPECTR (http://128.91.198.137/RIS/RISWEB. ISA#TOPOFREFLIST) (September 2002) SPECTR is a registry of over 10,000 randomised and possibly randomised trials in education, social work and welfare, and criminal justice and we searched this using the phrase: ((antisocial OR avoidant OR borderline OR dependent OR depressive OR histrionic OR impulsive OR multiple OR narcissistic OR paran OR psychopa OR sadistic OR schizotypal OR self-defeating OR sociopath) AND person) OR (gender AND identity) OR (passive AND aggressive) OR (antisocial AND behav) OR (moral AND insanity) OR (asocial OR antisocial OR dissocial OR psychopath OR sadist OR sociopath) 1.5 CareData: we searched this database of social work and social care literature (1985 to November 2002) using the phrase: (randomi* OR (double* AND blind*) OR (control* AND clinical*)) We downloaded the results into a Procite5 database and searched again using the terms: (antisocial* OR asocial* OR avoidant OR borderline OR dependent OR depressive OR dissocial OR dissocial* OR histrionic OR moral OR multi-impulsive OR multiple* OR narcissistic OR parano* OR passive-aggressive OR psychopath* OR sadis* OR schizotypal OR self-defeating OR sociopath*) 1.6 CINAHL: we searched (1982 to January 2003) using the phrase: exp personality disorders/ OR exp antisocial personality disorder/ OR exp borderline personality disorder/ OR exp compulsive personality disorder/ OR exp dependent personality disorder/ OR exp impulse control disorders/ OR exp passive-aggressive personality disorder/ OR (histrionic$ adj2 person$) OR (parano$ adj2 person$).mp. OR (schizo$ adj2 person$) OR ((asocial$ OR antisocial$ OR dissocial$ OR psychopath$ OR sadist$ OR sociopath$) adj2 person$) OR (psychopath OR sociopath OR (moral adj2 insanity) OR dyssocial OR (DSM and (Axis and II))) 1.7 Cochrane CENTRAL Register of Controlled Trials: we searched (October 2002) using the phrase: [(antisocial-personality-disorder*:me OR personalitydisorders*:me OR sexual-and-gender-disorders*:me OR multiple-personality-disorder*:me OR paraphilias*:me) OR (multi-impulsive and personality) OR (parano* NEAR person*) OR (asocial* NEAR person) OR (dissocial* NEAR person) OR (psychopath* NEAR person) OR (sadist* NEAR person) OR (sociopath* NEAR person*) OR (moral NEAR insanity) OR ((personality and disorder*) and ((((avoidant OR multiimpulsive) OR narcissistic) OR self-defeating) OR personality)] 1.8 Cochrane Schizophrenia Groups Register of trials related to Forensic Mental Health Services (2000): we searched this using the phrase:
5

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

(psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR self-defeating OR borderline OR avoidant OR dependent OR depressive OR histrionic OR multi-impulsive OR multiple OR narcissistic OR passive-aggressive) AND (person*) OR (antisocial AND behaviour) OR (personality AND disorder*) OR (gender AND identity) OR (parano* AND person*) OR (self AND defeating) OR ((asocial* or antisocial* or dissocial* or psychopath* or sadist* or sociopath*) AND person*) OR (moral AND insanity) 1.9 COPAC: we searched the Consortium of University Research Libraries joint catalogue (October 2002) using the following terms: randomi* OR ((double OR single OR triple OR treble) and blind) OR prospective OR (clinical and trial) We then downloaded results into a Procite5 database and searched again using the terms: (antisocial* OR asocial* OR avoidant OR borderline OR dependent OR depressive OR dissocial OR dissocial* OR histrionic OR moral OR multi-impulsive OR multiple* OR narcissistic OR parano* OR passive-aggressive OR psychopath* OR sadis* OR schizotypal OR self-defeating OR sociopath*) 1.10 CORDIS: we searched the Community Research and Development Information Service (1986 to November 2002) with the phrase: (randomi* OR (double and blind) OR (controlled and trial). We read entries on the web page and downloaded relevant records. 1.11 Criminal Justice Periodical Index: we searched (1968 to November 2002) using the phrase: (((personality-disorder) OR (personality-disordered)) OR ((antisocial-person) OR (antisocial-manipulator) OR (antisociality) OR (antisocially) OR (antisocials)) OR ((psychopath) OR (psychopathic) OR (psychopathic-deviant) OR (psychopathic-deviate) OR (psychopathically) OR (psychopaths-) OR (psychopathy) OR (psychopathy-)) OR (personality-disorder) OR (borderlines) OR (borderlines) OR (histrionicity) OR ((narcissistic) OR (narcissistically)) OR (passive-aggressive) OR ((sadistically) OR (sadists) OR (sadomasochism) OR (sado-masochistic) OR (sadomascochism) OR (sadomasochism) OR (sadomasochistic)) OR (schizotypy) OR (((asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath*) near person*)or (psychopath OR sociopath OR (moral near insanity) OR dissocial))) and (((single blind OR double blind OR triple blind OR treble blind OR randomi* OR randomised OR randomized OR controlled trial) OR ((random-assignment) OR (random-design) OR (randomization) OR (randomize) OR (randomized) OR (randomized-groups) OR (randomly-selected))) OR (double-

blind) OR ((blind-treatment) OR (blinded) OR (blinding)) OR ((controls) OR (controlled))) 1.12 Dissertation Abstracts: we searched (Digital Dissertations 1861 to December 2002) using the phrase: ((randomi? w/2 control?) OR (triple w/1 blind?) OR (double w/1 blind?) OR (treble w/1 blind?) OR (single w/1 blind?) OR ab(double blind?) OR ab(randomi?) OR (controlled clinical trial)) and ((asocial? OR antisocial? OR dissocial? OR psychopath? OR sadist? OR sociopath? OR histrionic OR multiimpulsive OR multiple? OR narcissistic) and person?) OR ((passive-aggressive OR psychopath? OR sociopath? OR dissocial OR sadis? OR schizotypal OR self-defeating OR borderline OR avoidant OR dependent OR depressive OR parano?) and person?) OR ((moral and insanity) OR (self and defeating) OR (gender and identity) OR (personality and disorder) OR (antisocial and behaviour)) 1.13 EMBASE: we searched (1980 to January 2003) using the phrase: exp personality disorder/ OR exp borderline state/ OR exp character disorder/ OR exp compulsive personality disorder/ OR exp delusion/ OR exp dependency/ OR exp depersonalization/ OR exp jealousy/ OR exp kleptomania/ OR exp multiple personality/ OR exp narcissism/ OR exp psychopathy/ OR exp schizoidism/ OR exp sociopathy/ OR (antisoci$ adj2 person$) OR (aggres$ adj2 person$) OR (border$ adj2 person$) OR histrion$ person$ OR paranoid person$ OR (passive adj2 aggressive) OR ((asocial$ OR antisocial$ OR dissocial$ OR psychopath$ OR sadist$ OR sociopath$) adj person$) OR (moral adj2 insan$) OR dyssocial OR (DSM and (Axis and II)) 1.14 Federal Research in Progress/CRSP: This is one of the databases within the GRC database. See 1.15 below. 1.15 GOV.Research_Center: we searched (1964 to December 2002) using the phrase: ((randomi* OR (double and blind*) OR (control* and clinical and trial*)) and ((antisocial OR avoidant OR borderline OR dependent OR depressive OR histrionic OR impulsive OR multiple OR narcissistic OR paran* OR psychopa* OR sadistic OR schizotypal OR self-defeating OR sociopath) and person) OR (gender and identity) OR (passive and aggressive) OR (antisocial and behav) OR (moral and insanity) OR (asocial OR antisocial OR dissocial OR psychopath OR sadist OR sociopath)) 1.16 IBSS: we searched the International bibliography of the Social Sciences (1951 to October 2002) using the phrase: (randomi* OR double blind) We then downloaded results into a Procite5 Database and searched this using the terms:
6

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

(antisocial* OR asocial* OR avoidant OR borderline OR dependent OR depressive OR dissocial OR dissocial* OR histrionic OR moral OR multi-impulsive OR multiple* OR narcissistic OR parano* OR passive-aggressive OR psychopath* OR sadis* OR schizotypal OR self-defeating OR sociopath*) 1.17 ISI - Proceedings (instead of ISI-ISTP): we searched (1990 to October 2002) using the phrase: (double blind OR randomi*) AND ((passive-aggressive OR psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR self-defeating OR borderline OR avoidant OR dependent OR depressive OR parano* OR asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath* OR histrionic OR multi-impulsive OR multiple* OR narcissistic) AND personality*) OR ((moral AND insanity) OR (self AND defeating) OR (gender AND identity) OR (personality AND disorder) OR (antisocial AND behaviour)) 1.18 ISI-SCI: we searched the Science Citation Index Expanded (1981 to November 2002) using the phrase: (double blind OR randomi*) AND ((passive-aggressive OR psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR self-defeating OR borderline OR avoidant OR dependent OR depressive OR parano* OR asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath* OR histrionic OR multi-impulsive OR multiple* OR narcissistic) AND personality*) OR ((moral AND insanity) OR (self AND defeating) OR (gender AND identity) OR (personality AND disorder) OR (antisocial AND behaviour)) 1.19 ISI-SSCI: we searched Social Sciences Citation Index (1981 to November 2002) using the phrase: (double blind OR randomi*) and ((passive-aggressive OR psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR self-defeating OR borderline OR avoidant OR dependent OR depressive OR parano* OR asocial* OR antisocial* OR dssocial* OR psychopath* OR sadist* OR sociopath* OR histrionic OR multi-impulsive OR multiple* OR narcissistic) and personality*) OR ((moral and insanity) OR (self and defeating) OR (gender and identity) OR (personality and disorder) OR (antisocial and behaviour)) 1.20 MEDLINE: we searched (1966 to January 2003) using the phrase: [exp personality disorders/ OR exp antisocial personality disorder/ OR exp borderline personality disorder/ OR exp compulsive personality disorder/ OR exp dependent personality disorder/ OR exp histrionic personality disorder/ OR exp hysteria/ OR exp paranoid personality disorder/ OR exp passiveaggressive personality disorder/ OR exp schizoid personality disorder/ OR exp schizotypal personality disorder/ OR ((asocial$ OR antisocial$ OR dissocial$ OR psychopath$ OR sadist$ OR

sociopath$) adj2 person$) OR psychopath OR sociopath OR (moral adj2 insanity) OR dyssocial OR (DSM and (axis and II))] 1.21 National Criminal Justice Reference Service Abstracts Database: we searched (1970 to November 2002) using the phrase: (randomi* OR double blind) and (antisocial* OR asocial* OR avoidant OR borderline OR dependent OR depressive OR dissocial OR dissocial* OR histrionic OR moral OR multiimpulsive OR multiple* OR narcissistic OR parano* OR passiveaggressive OR psychopath* OR sadis* OR schizotypal OR selfdefeating OR sociopath*) 1.22 National Research Register: we searched (Issue 3 2002) using the phrase: (antisocial-personality-disorder*:me OR (self AND (near3 AND defeating)) OR (antisocial AND behaviour) OR (self AND defeating) OR (parano* NEAR person*) or((((((asocial* OR antisocial*) OR dissocial*) OR psychopath*) OR sadist*) OR sociopath*) NEAR person*) OR (((((asocial* OR antisocial*) OR dissocial*) OR psychopath*) OR sadist*) OR sociopath*) OR (moral NEAR insanity) OR (((psychopath OR sociopath) OR (moral NEAR insanity)) OR dissocial)) AND (randomizedcontrolled-trials*:me OR double-blind-method*:me OR single-blind-method*:me OR controlled-clinical-trial*:me OR randomi* OR ((double OR single OR triple OR treble) AND blind) 1.23 PSYCINFO: we searched (1872 to December 2002) using the phrase: [((((personality disorders/ OR exp antisocial personality/ OR exp avoidant personality/ OR exp borderline personality/ OR exp dependent personality/ OR exp histrionic personality disorder/ OR exp narcissistic personality/ OR exp obsessive compulsive personality/ OR exp paranoid personality/ OR exp passive aggressive personality/ OR exp sadomasochistic personality/ OR exp schizoid personality/ OR exp schizotypal personality/) OR (((Personality adj disorders) OR (antisocial adj personality) OR (avoidant adj personality) OR (borderline adj personality) OR (dependent adj personality) OR (histrionic adj (personality AND disorder)) OR (narcissistic adj personality) OR (obsessive adj (compulsive AND personality)) OR (paranoid adj personality) OR (passive adj (aggressive AND personality)) OR (sadomasochistic adj personality) OR (schizoid adj personality) OR (schizotypal adj personality) and combined with the Cochrane Schizophrenia Groups search strategy for controlled clinical trials] 1.24 REGARD: we searched (1980’s to November 2002) using the phrase: (randomi* OR (double AND blind) OR (single AND blind) OR (control* AND trial))
7

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

We then downloaded results into a Procite5 Database and searched using the terms: (personality* OR antisocial* OR avoidant* OR borderline* OR dependent* OR histrionic* OR narcissistic* OR obsessive* OR compulsive* OR paranoid* OR passive* OR aggress* OR sadomasochistic* OR schizo*) 1.25 SIGLE: we searched System for Information on Grey Literature in Europe (1980 - November 2002) using the phrase: ((randomisation) OR (randomised) OR (randomisee) OR (randomises) OR (randomize) OR (randomized) OR (randomly) OR ((double AND blind) OR double-blind OR double* blind* OR randomi?ed controlled trials)) AND ((psychopath* OR sociopath* OR dissocial OR sadis* OR schizotypal OR selfdefeating OR borderline OR avoidant OR dependent OR depressive OR histrionic OR multi-impulsive OR multiple OR narcissistic OR passive-aggressive) AND (person*) OR (antisocial AND behaviour) OR (personality AND disorder*) OR (gender AND identity) OR (parano* AND person*) OR (self AND defeating) OR ((asocial* OR antisocial* OR dissocial* OR psychopath* OR sadist* OR sociopath*) AND person*) OR (moral AND insanity)) 1.26 Sociological Abstracts: we searched (1963 to November 2002) using the phrase: (((personality* OR antisocial* OR avoidant* OR borderline* OR dependent* OR histrionic* OR narcissistic* OR obsessive* OR compulsive* OR paranoid* OR passive* OR aggress* OR sadomasochistic* OR schizo*) near disorder) OR ((personality* OR antisocial* OR avoidant* OR borderline* OR dependent* OR histrionic* OR narcissistic* OR obsessive* OR compulsive* OR paranoid* OR passive* OR aggress* OR sadomasochistic* OR schizo*) AND disorder)) AND (AB=randomi* OR TI= random* OR DE=(randomi?ed controlled trials) OR AB= (double* blind*) OR TI=(double* blind*) OR DE=(double blind studies) OR (single* near blind*)) 1.27 ZETOC: we searched this database (1993 to November 2002) using the phrase: (personality* OR antisocial* OR avoidant* OR borderline* OR dependent* OR histrionic* OR narcissistic* OR obsessive* OR compulsive* OR paranoid* OR passive* OR aggress* OR sadomasochistic* OR schizo*) We then downloaded results into a Procite5 Database and searched using the terms: (randomi* or blind* or control*) 2. Hand searching We searched reference lists of included and excluded studies for additional relevant trials. Specific journals not previously hand searched, which gave a high yield of studies were to have been hand searched.

3. Requests for additional data We contacted authors of relevant studies to enquire about other sources of information and the first author of each included study for information regarding unpublished data.

METHODS OF THE REVIEW 1. Study selection CB and TL, the principal authors, inspected all electronic reports. These were then re-inspected by a third author (LMcC) in order to ensure reliable selection. We resolved any disagreement by discussion or, if doubt remained, acquired the full article for further inspection. Once we had obtained the full articles the three authors, again working independently, decided whether they met the criteria for inclusion. Again, when disputes arose, resolution was attempted by discussion but if this was not possible, we did not enter data and allocated the study to the list of those awaiting assessment whilst authors were contacted. MF re-inspected all included and excluded studies. 2. Assessment of quality Again working independently, we allocated trials to three quality categories, as described in the Cochrane Collaboration Handbook (Clarke 2002). When disputes arose as to which category a trial was to be allocated, again, we attempted resolution by discussion. When this was not possible, we excluded studies of possible category C until further details became available. We hoped to include only trials in category A or B in the review. 3. Data management 3.1 Data extraction CB, LMcC and MF, all working independently, extracted data from selected trials. When disputes arose we resolved this by discussion. When this was not possible and we needed further information to resolve the dilemma, we did not enter data and added this outcome of the trial to the list of those awaiting assessment. 3.2 Intention to treat analysis We excluded data from studies where more than 50% of participants in any group were lost to follow-up. In studies with less than 50% drop out rate, we considered people leaving early to have had the negative outcome, except for adverse effects such as death. We analysed the impact of including studies with high attrition rates (25 to 50%) in sensitivity analyses. If inclusion of data from this group did result in a substantive change in the estimate of effect of the primary outcomes, we did not add data from these studies to trials with less attrition, but presented them separately. 4. Data analysis 4.1.1 Binary data For binary outcomes we calculated a standard estimation of the fixed effects ratio (RR) and its 95% confidence interval (CI).
8

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Where possible, for results that reached conventional levels of statistical significance, we calculated the weighted number needed to treat/harm statistic (NNT/H), and its 95% confidence interval (CI). 4.1.2 Valid scales: we included data from rating scales only if the measuring instrument had been previously described in a peerreviewed journal (Marshall 2000) and the instrument was either a self-report or completed by an independent rater or relative (not the therapist). 4.2. Continuous data 4.2.1 Skewed data: continuous data on clinical and social outcomes are often not normally distributed. To avoid the pitfall of applying parametric tests to non-parametric data, we applied the following standards to all data before inclusion: (a) standard deviations and means were reported in the paper or were obtainable from the authors; (b) when a scale started from the finite number zero, the standard deviation, when multiplied by two, was less than the mean (as otherwise the mean was unlikely to be an appropriate measure of the centre of the distribution, (Altman 1996); (c) if a scale started from a positive value (such as PANSS which can have values from 30 to 210) the calculation described above was modified to take the scale starting point into account. In these cases skew is present if 2SD>(S-Smin), where S is the mean score and Smin is the minimum score. Endpoint scores on scales often have a finite start and end point and these rules can be applied to them. When continuous data are presented on a scale which includes a possibility of negative values (such as change on a scale), there is no way of telling whether data are non-normally distributed (skewed) or not. It is thus preferable to use scale end point data, which typically cannot have negative values. If end point data were not available, the authors would have used change data, but not subjected them to meta-analysis, and reported them in ’Additional data’ tables. 4.2.2 Summary statistic: for continuous outcomes we estimated a fixed effects weighted mean difference (WMD) between groups. Again, if we found heterogeneity (see section 5) we employed a random effects model. 4.2.3 Valid scales: we included continuous data from rating scales only if the measuring instrument had been previously described in a peer-reviewed journal (Marshall 2000) and the instrument was either a self-report or completed by an independent rater or relative (not the therapist). 4.2.4 Endpoint versus change data: where possible we present endpoint data and if both endpoint and change data had been available for the same outcomes then only the former would have been reported in this review. 4.2.5 Cluster trials: studies increasingly employ ’cluster randomisation’ (such as randomisation by clinician or practice) but analysis and pooling of clustered data poses problems. Firstly, authors often fail to account for intra class correlation in clustered

studies, leading to a ’unit of analysis’ error (Divine 1992) - whereby p values are spuriously low, confidence intervals unduly narrow and statistical significance overestimated - causing type I errors (Bland 1997, Gulliford 1999). Secondly, RevMan does not currently support meta-analytic pooling of clustered dichotomous data, even when these are correctly analysed by the authors of primary studies, since the ’design effect’ (a statistical correction for clustering) cannot be incorporated. Where clustering was not accounted for in primary studies, we presented the data in a table, with a (*) symbol - to indicate the presence of a probable unit of analysis error. Subsequent versions of this review will seek to contact first authors of studies to seek intra-class correlation co-efficients of their clustered data and to adjust for this using accepted methods (Gulliford 1999). Where clustering has been incorporated into the analysis of primary studies, then we would also have presented these data as if from a non-cluster randomised study, but adjusted for the clustering effect. We have sought statistical advice and have been advised that the binary data as presented in a report should be divided by a ’design effect’. This is calculated using the mean number of participants per cluster (m) and the intraclass correlation coefficient (ICC) [Design effect = 1+(m-1)*ICC] (Donner 2002). If the ICC was not reported it was assumed to be 0.1 (Ukoumunne 1999). If cluster studies had been appropriately analysed taking into account intra-class correlation coefficients and relevant data documented in the report, synthesis with other studies would have been possible using the generic inverse variance technique. 5. Test for heterogeneity Firstly, we undertook consideration of all the included studies within any comparison to estimate clinical heterogeneity. Then we used visual inspection of graphs to investigate the possibility of statistical heterogeneity. We supplemented this by employing, primarily, the I-squared statistic. This provides an estimate of the percentage of inconsistency thought to be due to chance. Where the I-squared estimate included 75% this was interpreted as evidence of high levels of heterogeneity (Higgins 2003). We then reanalysed data using a random effects model to see if this made a substantial difference. If it did, and results became more consistent, falling below 75% in the estimate, the studies were added to the main body trials. We did not summate data if using the random effects model did not make a difference and inconsistency remained high, but presented the data separately and investigated and discussed reasons for heterogeneity. 6. Addressing publication bias With more included studies we would have entered data from all included studies into a funnel graph (trial effect against trial size) in an attempt to investigate the likelihood of overt publication bias (Egger 1997). 7. Sensitivity analyses
9

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

We were to have analysed the effect of including studies with high attrition rates in a sensitivity analysis. Where a trial was described as ’double-blind’ but implied that the study was randomised, these trials were also to have been included in a sensitivity analysis. We were to have compared results of studies using Grade A diagnoses with those using Grade B labels (primary outcomes only). 8. General We entered data in such a way that the area to the left of the line of no effect indicated a favourable outcome for the intervention.

3. Awaiting assessment No studies await assessment. 4. Ongoing We have not identified any ongoing studies, but would welcome contact from anyone who is aware of any. 5. Included studies We found ten relevant studies that we could include (De la Fuente 1994; Goldberg 1986; Hollander 2001; Leone 1982; Montgomery 1983; Salzman 1995; Serban 1984; Soloff 1989; Soloff 1993). 5.1. Length of trials Study duration was from just over one month (De la Fuente 1994, Leone 1982, Soloff 1989, Soloff 1993), through about three months (Goldberg 1986, Hollander 2001, Salzman 1995, Serban 1984) to six months (Montgomery 1983, Zanarini 2001). 5.2 Participants Five hundred and fifty four participants are included in these studies. People who participated in De la Fuente 1994 were inpatients fulfilling the criteria of DSM-IIIR. They also had to have a score of seven on the Gunderson and Kolb Diagnostic Interview for borderline patients (Gunderson 1981), and had to be able to abstain from alcohol or psychoactive substances use, have a history of compliance and not have any abnormal physical results following examination. Goldberg 1986 randomised people with a borderline and/or schizotypal personality disorder who volunteered following an advertisement containing a shortened version of the Schedule of Interviewing Schizotypal personalities (SIB) in the local newspaper. Once people attended the full SIB was administered. Goldberg 1986 also used similar exclusion criteria to De la Fuente 1994. Hollander 2001 included people who attended North American outpatient clinics, self help groups, answered private practice advertisements and articles in local media and who could be diagnosed BPD using DSM-IV, Axis II criteria. Participants in Leone 1982 were ’patients with acute, disruptive symptoms’ who were required to exhibit four or more of the Gunderson and Kolb Diagnostic Interview criteria. Montgomery 1983 randomised people following admission to a medical ward following a suicidal act, who were then screened for the presence of a psychiatric illness. People who had a history of at least two documented acts prior to the index act and who were not suffering from a physical illness, overt schizophrenia or depression were included. Salzman 1995 included people who responded to a newspaper advertisement and gave written informed consent to treatment. Participants had to meet screening requirements of not quite meeting full diagnosis criteria for borderline personality disorder following a screening interview using the Diagnostic Interview for Borderlines Revised (DIB-R0 Gunderson 1983), and the Structured Clinical Interview for DSM-III-R personality disorders (SCID-II , First 1997). All participants ’acknowledged an unstable sense of identity, irritability, and “trigger temper”, an inner sense of emptiness, fluctuating and environmentally sensitive
10

DESCRIPTION OF STUDIES For substantive descriptions of studies please see Included and Excluded studies tables. 1. The search A comprehensive electronic search was undertaken. It is perfectly feasible, however, that despite the extensive efforts, we could have failed to identify some studies. With only a few exceptions, each database yielded a few more unique reports to inspect. We think it unlikely, however, that large studies have not been identified by this search. 2. Excluded In a review such as this there is always an issue of what to report in the excluded studies section. We identified thousands of electronic reports in our search but reporting on the great majority of those would, we feel, do no service to the reader. Most could be easily dismissed as not relevant to the review. However, we selected 48 studies for closer inspection as the electronic record suggested that they could be relevant. We had to reject 33 articles of the 48 because they were not randomised and are not reported upon in the review, with the majority having used a case series design, and one a cohort study. Five studies are of particular note (Benedetti 1998, Bohus 1999, Chengappa 1999, Cowdry 1988, Links 1990). Benedetti 1998 used a case series design, but included a pre and post treatment comparison on 12 people diagnosed with BPD using DSMIV. They were given low dose clozapine following four months of inpatient therapeutic treatment to which they did not respond. Bohus 1999 included in a case series people with borderline personality disorder experiencing flashbacks with nalrexone (naltrexone). Participants were split into two groups, with an overlap of five participants, and given different rating scales to complete. In Chengappa 1999 seven white women known to the authors were selected and given clozapine in a before and after analysis of events such as self harm, seclusion, injuries to staff and peers and levels of hospital privileges. Cowdry 1988 randomised 16 women into a crossover study, using alprazolam, carbamazepine, trifluperazine hydrocholride; tranlcypromine or placebo but reported no data for the first arm of the study. Links 1990 also was a randomised, double blind, crossover study and did not report data for the first arm of the study. This study randomised 19 people to lithium, desipramine or placebo.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

mood and extreme sensitivity to rejection and criticism’. People who participated in Serban 1984 were a consecutive sample of patients with borderline and schizotypal personality disorders who came for treatment to a New York walk-in clinic. Trialists used a DSM-III criterion which was checked using the Borderline Syndrome Index. Participants also had to have had the condition for at least the two previous years with no pre-existing psychiatric illness diagnosed as schizophrenia or affective disorder and had to have a mild transient psychotic disorder before admission. Soloff 1989 randomised people who were inpatients or outpatients who were diagnosed by Gunderson’s Diagnostic Interview for Borderlines and sub-typed by DSM-III into borderline personality disorder, schizotypal personality disorder or schizotypal-borderline mixed disorder. Exclusion criteria was based on the Schedule for Affective Disorders and Schizophrenia, including those with chronicity of illness, an organic illness or a diagnosis of schizophrenia, mania and related psychotic disorders. They allowed participants with a comorbid diagnosis of major depression. Scores of 50 or less on the Global Assesment Scale, and either 17 or greater on the Hamilton Depression Scale, or 66 or greater on the Inpatient Multidimensional Rating Scale were required for study inclusion. Soloff 1993 used similar inclusion criteria for people who were all inpatients, with the addition of using Quitkins Atypical Depression Scale (unpublished) and hysteroid dysphoria using criteria of Liebowitz and Klein (Liebowitz 1981). These trialists excluded people with evidence of drug or alcohol-related deficits or physical dependence, evidence of central nervous system disease, having had recent ECT or a formal diagnosis of seizure disorder, physical disorders of known psychiatric consequence or borderline mental retardation. All participants were in the age range of 16 to 36 years. Zanarini 2001 recruited women between the ages of 18 and 40, using advertisements in a local newspaper. These women were disturbed by moodiness, distrustfulness, impulsivity and painful and difficult relationships. Following contact via telephone, people were initially screened for DSM-IV criteria for borderline personality disorder. At the first face to face meeting people were further screened using the SCID and DIB-R. Exclusion criteria were current or lifetime risk for schizophrenia, schizo-affective disorder or bipolar disorder. De la Fuente 1994, Hollander 2001, Leone 1982, Salzman 1995 and Zanarini 2001 only included people with BPD. The majority of people in Goldberg 1986, Montgomery 1983, Serban 1984, Soloff 1989 and Soloff 1993 did have BPD but these studies did include others with a mix of diagnosis. 5.3 Setting All studies were set in outpatients with three exceptions. Montgomery 1983 included people admitted to hospital following a deliberate self harm attempt, Soloff 1989, a mixture of inpatient and outpatients and Soloff 1993, in which all participants were inpatients. 5.4 Study size The largest study included 108 people (Soloff 1993) and the smallest only 16 (Hollander 2001) (mean 52 SD 31, median 51).

5.5. Interventions 5.5.1 Antidepressants Both trycyclic, tetracyclic and MAOI antidepressants have been given within trials relevant to this review (Montgomery 1983, Salzman 1995, Soloff 1989, Soloff 1993). Antidepressants used were mianserin, fluoxetine, amitriptyline and phenelzine sulfate versus placebo or antipsychotic medication. 5.5.2 Antipsychotics The most common intervention given was antipsychotic medication (Goldberg 1986, Leone 1982, Serban 1984, Soloff 1989, Soloff 1993, Zanarini 2001). Antipsychotic medication used was thiothixene, loxapine, haloperidol and olanzapine, compared with placebo, antidepressants and other antipsychotic medication. 5.5.3 Mood stabilisers The remaining studies (De la Fuente 1994, Hollander 2001) used the mood stabilisers carbamazepine or divalproex sodium, both versus placebo. 5.6 Outcomes One study only provides data for those leaving the study early due to the high levels of loss (Zanarini 2001). 5.6.1 Improvement There were as many definitions of improvement as there were studies. De la Fuente 1994 looked for differences between the pre and post-treatment scores on the Global Assessment Scale (GAS), Symptoms Check List (SCL-90), 24 item Hamilton Depression Rating Scale (HDRS-24), Brief Psychiatryic Rating Scale (BPRS) and the Acting Out Scale, (derived from the literature on Borderline Personality disorder), and used repeated measures of analysis of variance. Hollander 2001 pre-stated that efficacy would be measured by the two global scales used in the study, the Clinical Global Impression (CGI-I) and the Global Assessment Scale (GAS). Participants were considered responders if they received a score of two or one on the CGI-I. Last observation carried forward was used. All participants in the placebo group left. Efficacy in the Leone 1982 study was by repeated measures on the BPRS, Systematic Nurses’ Observation of Psychopathology (SNOOP) and the Clinical Global Impression (CGI), with change from baseline being the main indicator of improvement. Salzman 1995 used a 60% improvement in the Personality Disorder Rating Scale (PDRS), 20% on the Profile of Mood States (POMS) and 80% on the Hamilton Depression Scale (HAM-D) as their measure of improvement. Zanarini 2001 used between group differences on the Hopkins Symptom Checklist-90 (SCL-90). Goldberg 1986 used no operationalised definition of improvement, although change from baseline to endpoint, using last observation carried forward for all rating scales was used. Soloff 1989 reports difference between groups on several rating scales rather than identifying a primary outcome. In Soloff 1993 improvement was defined as an increment of 6.6 or more over their baseline GAS score by the end of the trial. Salzman 1995 used analysis of covariance between
11

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

groups on several rating scales. Montgomery 1983 used further suicidal acts as the measure of improvement. 5.6.2 Adverse effects No formal or operationalised measures of adverse effect were taken in Leone 1982, who noted appearance, date of onset, intensity, duration and remedial action taken. Zanarini 2001 reports some adverse effects, but does not seem to use an operationalised checklist. None of the other studies report adverse effects data in a usable format. 5.6.3 Outcome scales - Details of scales that provided usable data are shown below. Reasons for exclusion of data from other instruments are given under ’Outcomes’ in the ’Included studies’ section. 5.6.3.1 Global state a. Global Assessment Scale - GAS (Endicott 1976) Used to evaluate the overall functioning of a person during a specified time period in terms of psychological well-being or sickness. Time period assessed is generally one week prior to evaluation. Scale covers entire range of severity and can be used in any situation or study where an overall evaluation of the severity of the illness or degree of health is needed. b. Clinical Global Impression Scale - CGI Scale (Guy 1976) This is used to assess both severity of illness and clinical improvement, by comparing the conditions of the person standardised against other people with the same diagnosis. A seven-point scoring system is usually used with low scores showing decreased severity and/or overall improvement. 5.6.3.2 Mental state a. Beck Depression Inventory BDI (Beck 1961) A twenty-one question survey completed by participant. Answers scored on 0 to 3 scale (minimal = 0, severe = 3). b. Hamilton Depression Scale - 17 item version HAMD-17 (Hamilton 1960) This instrument should be used only on patients already diagnosed as suffering from affective disorder of depressive type. It is used for quantifying the results of an interview and its value depends entirely on the skill of the interviewer in eliciting the necessary information. The scale contains 17 variables measured on either a five-point or a three-point rating scale, the latter being used where quantification of the variable is either difficult or impossible. Among the variables are: depressed mood, suicide, work and loss of interest, retardation, agitation, gastro-intestinal symptoms, general somatic symptoms, hypochondriasis, loss of insight and loss of weight. It is useful to have two raters independently scoring a patient at the same interview. The scores of the patient are obtained by summing the scores of the two physicians. c. Hamilton Depression Scale - 24 item version - HAM-D-24 or HDRS-24 (Williams 1988) The instrument is designed for patients already diagnosed as suffering from affective disorder of depressive type. It is used for quan-

tifying the results of an interview, and its value depends entirely on the skill of the interviewer in eliciting the necessary information d. Inpatient Multidimensional Rating Scale IMPS (Lorr 1962) Another rating scale used to assess the severity of a range of psychiatric symptoms. Higher scores indicate more symptoms. Further details could not be obtained. e. Profile of Moods State POMS (self rated) (McNair 1971). This instrument was designed to measure mood states in psychiatric outpatients and as a method for assessing change in such people. Has been used in many drug evaluation studies. Mood reactions are to be reported for a specific period of time, such as the previous week. This helps distinguish mood states from enduring personality traits. It contains 65 items, takes about 5 minutes to complete and is designed for use with adults. High score indicates a poor outcome. 5.6.3.3 Behaviour a. Acting Out Scale AOS (Cowdry 1988 a) De la Fuente 1994 used the AOS to measure behavioural dyscontrol. We are unsure as to whether this is a scale, but the authors give the following information: ’Which classified episodes of behavioural dyscontrol into no acting out, plus three categories: mild acting out (exaggerated demanding, angry outbursts, suicidal threats), moderate acting out (throwing objects, physical violence without the intention of causing injury, head or arm banging), severe acting out (suicidal acts, physical violence either the intention to cause injury). When a participant committed an acting out behaviour the were assigned to a category, if they committed further, more serious acting out behaviour, they were then assigned to the next category. For each period of time in the study, they participant was registered as improved, unchanged or worse’. b. Buss Durkee Hostility Inventory - BDHI (Buss 1957) This is a self-rated multidimensional scale of hostility. The tester reads to the respondent some behaviours that people use to handle problems and express feelings. The respondent is asked how often he or she behaved this way during the last week using the categories listed below. Assault (physical violence against others); Indirect Hostility (undirected aggression); Irritability (readiness to explode with negative affect with provocation); Negativism (oppositional behaviour); Resentment (Jealously and hatred of others); Suspicion (projection of hostility toward others) and Verbal Hostility (negative affect expressed in style and content of speech). Variables are graded from 0 to 4, with 4 equating to 5 or more times a week. High score indicates a poor outcome. c. Barratt Impulsiveness Scale BIS (Barratt 1965) An original version of this test was subjected to factor analysis to produce a 30-item test that loaded on 5 factors. This was further refined to a 26-item self-report questionnaire to measure impulsivity. It provides four factor scores: speed of cognitive response; impulsiveness; adventure seeking; risk-taking. High score indicates a poor outcome.
12

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

d. Borderline Syndrome Index BSI (Conte 1980) This is a self-report questionnaire. It is a 52-item, forced choice measure based on phenomena as described. A semi-structured interview is used, in which 24 personality traits are rated on 08 scales. Diagnoses are computed using combinations of trait scores, and fall into four broad groups; the Antisocial group (sociopathic, explosive and sensitive aggressive personality disorders); the Dependent group (passive dependent, histrionic and asthenic personality disorders); the Inhibited group (anankastic, anxious, hypochondriacal and dysthymic personality disorders); and the Withdrawn group (schizoid, paranoid and avoidant personality disorders). Certain combinations of variables correspond to Axis II, DSM-III-R categories. e. Symptoms Check List - SCL90-R (Derogatis 1977) The Symptoms Check List is a self-rated instrument containing 90 symptom related questions. The subject assesses the degree of severity of each symptom: The scale ranges from: 0 (’Not at all’) 1 (’A little bit’) 2 (’Moderately’) 3 (’Quite a bit’) to 4 (’Extremely’). High score=poor. f. Self Report Test of Impulse Control - STIC (Lazzaro 1969) This is a 72 item measurement of impulsiveness. We have been unable to find further details. 6. Redundant data Many studies did record more data than we have been able to present. Much were reported in such a way that rendered the information meaningless or of no use for analyses. For example Serban 1984 used four rating scales, reported average results, but no standard deviations. Soloff 1989 did report usable data on many other rating scales, but as they were unpublished or no reference was given for them and the authors were unable to find further details, and these data had to be left out because of the high risk of introducing bias (Marshall 2000). Two trials potentially interesting studies had to be excluded because they were crossover studies which contained no data on the first arm of the study (Cowdry 1988, Links 1990) 7. Missing outcomes There are almost no data on outcomes such as self harm and none on the outcome of death. We found no data on hospital or service outcomes. What data there were on mental state and behaviour were difficult to interpret from the clinical perspective.

used a random number sequence to allocate medication bottles to participants. 2. Blinding In De la Fuente 1994 one study author was blind to allocation and performed all psychometric tests, the other, who was not blind to allocation monitored blood plasma levels. Participants were requested not to communicate adverse effects to the blinded clinician. Leone 1982 and Goldberg 1986 used identical appearing capsules. The treating clinician in Hollander 2001 was kept blind to intervention group. Salzman 1995 and Serban 1984 also used identical looking capsules as did Zanarini 2001. The latter study also supplied the tablets in numbered bottles, determined by a random number sequence. All participants and clinicians were blinded to assignment until all endpoint data had been acquired. No details are given for Montgomery 1983, Soloff 1989 or Soloff 1993. 3. Follow-up Two of the twenty participants in De la Fuente 1994 dropped out from the chlorpromazine group, whilst nobody left in the placebo group. Forty eight percent of people who participated in Goldberg 1986 left the study early. The small Hollander 2001 study understandably used an Intention to treat analysis using last observation carried forward. However, everyone in the placebo group (n=4) failed to complete and half of the 12 allocated to divalporex sodium. As last observation forward data were available, this study is included. Leone 1982 lost 14% to follow up. It was obvious from which group these people came from. Follow up in Montgomery 1983 was only on people who completed the study. Of the original 58 who participated, 35% left early. It is not possible to identify the original numbers of people who were in each arm of the study, and the authors are impossible to contact to clarify the situation. In Salzman 1995, 22 of the 27 who enrolled, completed this study. No information was available as to which groups the five who left early were in, although it is suspected most were from the placebo group (n=9 vs n=13). Serban 1984 lost 11% of people who enrolled, but again it was clear from which group people had left. Soloff 1989 did not include five people in the final analysis as they did not complete the first two weeks of treatment. However, as all outcomes other than this were reported as continuous results, these could not be added back in, although it was clear which groups they were from. Soloff 1993 lost 32 participants from 108 who were randomised, and it was clear which groups these came from. End point analysis was performed using last observation carried forward. Zanarini 2001 used last observation carried forward. Attrition was high, with 57% having left by week 20 of this 24 week study. 4. Overall The reporting of methods within the ten was not good and leaves all results prone to the influence of bias.
13

METHODOLOGICAL QUALITY 1. Randomisation Eight of the ten included studies reported no details about how randomisation was generated and concealed (De la Fuente 1994, Goldberg 1986, Hollander 2001, Leone 1982, Montgomery 1983, Salzman 1995, Soloff 1989, Soloff 1993). Serban 1984 used computer generated randomisation codes whilst Zanarini 2001

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

RESULTS 1. COMPARISON 1. ANTIDEPRESSANT (TRICYCLIC/ TETRACYCLIC) vs PLACEBO 1.1 Global state: Mean score - by 6 months Soloff 1989 employed and reported the GAS scores. They found no difference between amitriptyline and placebo, although the results look as though they begin to favour those taking placebo (n= 57, MD 3.32 CI -2.91 9.55). This study also used the IMPS and SCL-90 but data were skewed as are those for the IMPS subscale for excitement. 1.2 Mental state For anger, by 6 months, an antidepressant (fluoxetine) seems to offer some improvement for those on antidepressant therapy over placebo (n=22, 1 RCT, RR anger not improved 0.30 CI 0.10 to 0.85, NNT 2 CI 2 to 9). Salzman 1995, using the IMPS, did not detect a statistical significance in anxiety scores (intropunitiveness by 6 months, n=57, MD -0.39 CI -4.87 to 4.09). Also measuring anxiety, this time using the SCL-90 subscales to measure general and phobic anxiety, and interpersonal sensitivity, Soloff 1989 reports uninformative skewed data. Depression was rated in several different ways, for example, Salzman 1995 used two different validated rating scales for this purpose. The HAM-D, did not find a statistical difference (n=22, RR 0.14 CI 0.01 to 2.66) and the POMS did (n=22, RR 0.26 CI 0.09 to 0.72, NNT 2 CI 2 to 5). Also at about 6 months, Soloff 1989 presents skewed data for the BDI, reported as statistically significant (F=6.51, p</=0.05 (A>P) as well as skewed HAMD-17 data (n=29), HAM-D-24 (n=29, F=4.33, p</=0.05 (A>P), and the SCL-90 subscale (n=29). Salzman 1995 rated mood disturbance (not improved by 6 months) using the POMS but did not reach conventional levels of statistical significance when comparing fluoxetine with placebo (n=22, RR 0.35 CI 0.12 to 1.03). Using four sub-scale scores of the IMPS and the SCL-90, Soloff 1989 presents 6 month skewed data on several schizotypal symptoms such paranoid ideation, paranoid projection, perceptual distortions, and psychoticism. This group found no striking differences between amitriptyline and placebo. Finally for mental state within this comparison, Soloff 1989 reports skewed data for obsessive-compulsive symptoms, somatization and unspecific other symptoms. Again, in this small trial (n= 29), they found no differences between amitriptyline and placebo. 1.3 Behaviour The one small study investigating the important outcome of attempted suicide (Montgomery 1983, n=38) found no difference between those allocated to mianserin and people given placebo (RR 0.82 CI 0.44 to 1.54).

For the outcome of hostility there was no statistical difference between groups (n=57, 1 RCT, MD BDHI -0.67 CI -1.57 to 0.23), as was the case for impulsiveness (n=57, 1 RCT, MD -2.56 CI -13.10 to 7.98). Other measures of hostility tended to be skewed and always on very small numbers. There were no convincing differences in scores by 6 months for general hostility using the SCL-90 subscale, hostile belligerence using the IMPS subscale and verbal hostility using the BDHI subscale. Soloff 1989 (n=29) reports skewed data for the outcome of impulsiveness using the WSIAP, but also more clear data on a similar outcome using the BIS to measure self-control. This group found no statistical difference between those on amitriptyline and people allocated to placebo (n=57, 1 RCT, MD -2.56 CI -13.10 to 7.98). Using the STIC they again found no statistical differences (n=57, 1 RCT, MD 1.02 CI -1.13 to 3.17). 1.4 Leaving the study early One study reported good data on attrition. No statistical difference was found between groups for the outcome of leaving the study early (n=59, RR 0.97 CI 0.06 to 14.74). 2. COMPARISON 2. ANTIDEPRESSANT (TRYCYCLIC) vs ANTIPSYCHOTICS This comparison only involves one study (Soloff 1989, n=57). 2.1 Global state Using the GAS, Soloff 1989 found no statistical difference between people allocated to amitriptyline compared with those taking haloperidol (n=57, MD by 6 months -3.87 CI -10.67 to 2.93). The skewed data on the IMPS tended to favour haloperidol but the average SCL-90 scores did not really indicate a difference. Soloff 1989 reported skewed data on excitement at 6 months that were not statistically significant. 2.2 Mental state: 1a. Anxiety score - intropunitiveness by 6 months (IMPS) In Soloff 1989, when using the IMPS to detect intropunitiveness by 6 months, conventional levels of statistical significance were not detected (n=57, MD 2.32 CI -1.93 to 6.57). 2.5 Mental state: 1b. Anxiety scores - by 6 months For measures of general and phobic anxiety and interpersonal sensitivity using the SCL-90 sub-scales, Soloff 1989 reports skewed data which are reported as not reaching statistical significance for both general and phobic anxiety. These data were reported as being significant for interpersonal sensitivity in favour of haloperidol over amitriptyline. 2.6 Mental state: 2. Depression scores - by 6 months Using the BDI, HAMD-17 and 24 and the SCL-90 subscale, Soloff 1989 reports that skewed data are not statistically significant. 2.7 Mental state: 3. Schizotypal symptoms - by 6 months
14

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Soloff 1989 measured paranoid ideation using the SCL-90; paranoid projection and perceptual distortions using IMPS; psychoticism the SCL-90 and schizotypal symptoms using SSI. All returned skewed results. Haloperidol was reported to be statistically significantly better for paranoid ideation, paranoid projection, perceptual distortions, and schizotypal symptoms compared with amitriptyline. 2.8 Mental state: 4. Other symptoms - by 6 months Again using the SCL-90 subscale for obsessive-compulsive symptoms; somatization symptoms and unspecified additional items, Soloff 1989 reported skewed results. 2.9 Behaviour: 1.a. indirect hostility score Soloff 1989 reported an indirect hostility measure using the BDHI which was not statistically significant (n=57, MD-0.09 CI -0.09 to 0.78). 2.10 Behaviour: 1b. Hostility scores - by 6 months For the outcomes of general hostility using the SCL-90 subscale; hostile belligerence using the IMPS subscale and verbal hostility using the BDHI subscale, all results were skewed (Soloff 1989). Outcomes for general hostility favoured haloperidol over amitriptyline. 2.11 Behaviour: 2a. Impulsiveness - behaviour by 6 months Soloff 1989 used the WSIAP and reports skewed data, results favoured haloperidol over amitriptyline. 2.12 Behaviour: Impulsiveness Using the BIS or the STIC to measure and report self control, Soloff 1989 found no statistical difference. 2.13 Leaving the study early Finally, few people left this study with no differences between groups (n=61, RR 0.34 CI 0.04 to 3.13). 3. COMPARISON 3. ANTIDEPRESSANT (MONOAMINE OXIDASE INHIBITER) vs PLACEBO 3.1 Global state: 1. Borderline symptomatology - by 6 months (BSI) Using the BSI to identify change in borderline symptomatology, Soloff 1993 reports skewed data that are not statistically significant. 3.2 Global state: 2. Global functioning score - by 6 months (GAS) Using the GAS, Soloff 1993 finds no difference between those on MAOI’s and placebo (n=64, MD 1.67 CI -4.15 to 7.49). 3.3 Global state: 3a. Mean symptom score - by 6 months (IMPS) Soloff 1993, using the IMPS found no statistical difference for the outcome of global state (n=62, MD 0.10 CI -10.36 to 10.56) 3.4 Global state: 3b. Mean symptom score - by 6 months (SCL-90) Soloff 1993 reports skewed data using the SCL-90 for mean symptom score but found no differences between the monoamine oxidase inhibitor and placebo.

3.5 Mental state: 1. Did not respond - by 6 months (HAM-D-17, 50% improvement) Soloff 1993 detected no difference between people who responded when results were dichotomized (n=72, RR 0.93 CI 0.71 to 1.22). 3.6 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS) Using the IMPS, Soloff 1993 found no statistical difference between those on MAOI’s and those on placebo (n=62, MD -1.51 CI -5.38 to 2.36). 3.7 Mental state: 2b. Anxiety scores - by 6 months Soloff 1993 reports skewed results when using the SCL-90 to measure anxiety and found no statistical difference between those on MAOI’s and those on placebo. 3.8 Mental state: 3b. Depression scores - by 6 months Using the BDI, HAM-D-17, HAM-D-24 and SCL-90 subscale, Soloff 1993 reports non statistically significant skewed results. 3.9 Mental state: 4. Schizotypal symptoms - by 6 months Soloff 1993, measuring paranoid ideation using the SCL-90, paranoid projection using the IMPS, psychoticism using the SCL-90 and schizotypal symptoms using the SSI, found no clear differences between groups. 3.10 Mental state: 5. Other symptoms - by 6 months Soloff 1993 using the ADI subscale to measure hypersomnia, the ADI subscale for leaden paralysis, the SCL-90 subscale to measure obsessive-compulsive symptoms, the ADI subscale to measure reactivity and rejection sensitivity, reports no statistically significant differences in skewed data. 3.11 Behaviour: 1a. Hostility - by 6 months (BDHI) Using the BHDI, Soloff 1993 found a statistical difference favouring people on MAOI’s compared with those allocated to placebo. People given MAOIs were less hostile (n=62, MD -9.19 CI -16.12 to -2.26). 3.12 Behaviour: 1b. Hostility scores - by 6 months Using a different measurement of hostility (SCL-90 subscale) and hostile belligerence (IMPS subscale), Soloff 1993 reports non statistically significant and skewed results. 3.13 Behaviour: 2a. Impulsiveness - behaviour by 6 months WSIAP Using the WSIAP, Soloff 1993 reports non-significant and skewed results. 3.14 Behaviour: 2b. Impulsiveness - traits by 6 months Using the BIS to measure self control (n=62, MD -0.08 CI -19.54 to 19.38) and the STIC, impulsiveness (n=62, MD 2.08 CI -1.83 to 5.99), Soloff 1993 found no difference between people on MAOI’s and placebo. 3.15 Appetite and weight
15

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Using the ADI subscale to measure increased appetite and ADI subscale to record weight gain Soloff 1993 reports no significant differences in the skewed results. 4. COMPARISON 4. ANTIDEPRESSANT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTICS 4.1 Global state 1. Borderline symptomatology - by 6 months (BSI) Soloff 1993 reports non statistically significant skewed results when people on MAOI’s are compared with those on haloperidol. 4.2 Global state: 2. Global functioning score - by 6 months (GAS) Using the GAS, Soloff 1993 reports a significant difference between people on phenelzine over those taking haloperidol (n=64, MD 5.15 CI 0.29 to 10.01). 4.3 Global state 3a. Mean symptom score The mean symptom score on IMPS did not reach conventional levels of statistical difference in Soloff 1993, (n=64, MD -3.57 CI -11.66 to 4.52). 4.4 Global state: 3b. Mean symptom score - by 6 months (SCL-90) Skewed results favouring MAOI’s over haloperidol are reported in Soloff 1993 for people on phenelzine and those haloperidol. 4.5 Mental state: 1. Did not respond - by 6 months (HAM-D-17 50% improvement) No clear difference was found for the outcome of not responding in Soloff 1993 (n=58, RR 0.83 CI 0.65 to 1.05). 4.6 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS) Soloff 1993, using the IMPS, found clear differences between people on phenelzine and those allocated haloperidol with results favouring the phenelzine group (n=64, MD -3.88 CI -7.51 to -0.25). 4.7 Mental state: 2b. Anxiety scores - by 6 months Soloff 1993 using the SCL-90 subscale for anxiety, reported skewed results which favour phenelzine. 4.8 Mental state: 3a Depression by six months Soloff 1993 found clear differences between groups using the HAM-D-17 (n=64, MD -7.86 CI 10.51 to -5.21). 4.9 Mental state: 3b. Depression scores - by 6 months However using the BDI, Soloff 1993 reports no statistically significant findings for skewed data when measuring the same symptom cluster. When using the HAM-D-24 and SCL-90 subscale statistical significance is indeed reported in favour of phenelzine. 4.10 Mental state: 4. Dysphoria - hysteroid dysphoria - by 6 months (HDQ) Soloff 1993 reported no statistically significant findings from skewed data using the HDQ. 4.11 Mental state: 5. Schizotypal symptoms - by 6 months

Soloff 1993, measuring paranoid ideation and psychoticism using the SCL-90, paranoid projection using the IMPS and schizotypal symptoms using the SSI reports no significant findings from the skewed data. Using the SSI to measure schizotypal symptoms, skewed results do favour phenelzine. 4.12 Mental state: 6. Other symptoms - by 6 months Soloff 1993 measured hypersomnia, reactivity, rejection sensitivity and leaden paralysis score using the ADI subscales and reported non statistically significant skewed results. Using the SCL-90 to measure obsessive-compulsive symptoms significance was reported in favour of phenelzine. 4.13 Behaviour: 1a. Hostility - by 6 months (BDHI) Using the BDHI, Soloff 1993 found no clear difference between people taking phenelzine and those allocated placebo (n=64, MD -4.84 CI -11.94 to 2.26). 4.14 Behaviour: 1b. Hostility scores - by 6 months Soloff 1993 reported statistically non-significant skewed results for the outcome of general hostility rated on the SCL-90 subscale. However, for the outcome of hostile belligerence, this time using the IMPS subscale, skewed results do favour haloperidol over phenelzine. 4.15 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP) Using the WSIAP, Soloff 1993, reported statistically significant findings from skewed data favouring haloperidol. 4.16 Behaviour: 2b. Impulsiveness - traits by 6 months Soloff 1993, measuring impulsiveness using the BIS for self control, found no clear difference between people allocated phenelzine and those given placebo (n=64, MD -3.29 95% CI -21.10 to 14.52). Using the STIC, again they found no statistical difference (n=64, MD 0.63 CI -3.35 to 4.61). 4.17 Appetite and weight Using the ADI subscale to measure increased appetite and weight gain scores, Soloff 1993 reports no statistically significant findings for the skewed data. 5. COMPARISON 5. ANTIPSYCHOTICS vs PLACEBO 5.1 Global state: 1. Borderline symptomatology - by 6 months (BSI) Soloff 1993, using the BSI, reports skewed data not favouring haloperidol over placebo for borderline symptoms. 5.2 Global state: 2. Global functioning score - by 6 months (GAS) Soloff 1989 found no clear difference between people on haloperidol and those allocated placebo (n=114, WMD 1.75 CI -2.37 to 5.86). 5.3 Global state: 3a. Mean symptom score - by 6 months (IMPS) Combined, statistical difference is not found using the IMPS mean symptom score which showed heterogeneous results in both Soloff 1989 and Soloff 1993 (n=114, WMD -1.86 CI -10.85 to 7.14;
16

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

chi-squared 5.00 df=1, p=0.03). Soloff 1989 found a difference favouring people on haloperidol over those on placebo (n=56, MD -20.30 95% CI -39-91 to -1.95) whilst Soloff 1993 demonstrates a different direction of effect, which did not reach statistical significance (n=58, MD 3.67 95% CI -6.55 to 13.89). 5.4 Global state: 3b. Mean symptom score - by 6 months (SCL-90) Soloff 1989 and Soloff 1993 reported no difference for the mean symptom score on the SCL 90. 5.5 Mental state: 1. Did not respond by 6 months (HAM-D-17 50% improvement) Soloff 1993 reported non-significant results for the outcome of ’no improvement in mental state’ (n=70, RR 0.77 CI 0.37 to 1.63). 5.6 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS) Soloff 1989 and Soloff 1993 did not detect statistical significance using the IMPS for intropunitiveness (n=114, WMD -0.36 CI -3.30 to 2.58). 5.7 Mental state: 2b. Anxiety scores - by 6 months Soloff 1989 used the SCL-90 to measure general anxiety; Soloff 1993 and Soloff 1989 used the SCL-90 to measure phobic anxiety. All results were skewed and not statistically. Soloff 1989 again used the SCL-90 sub-scale to measure interpersonal sensitivity reports a significant difference favouring people on haloperidol. 5.8 Mental state: 3a. Depression scores - by 6 months Soloff 1989 used the HAM-D-17 to measure depression and did not find any clear difference (n=56, MD -1.44 CI -1.44 CI -4.41 to 1.53). 5.9 Mental state: 3b. Depression scores - by 6 months Two studies report statistically non-significant skewed results measuring depression using the BDI. Soloff 1993, using the HAMD-17 and HAM-D-24 found the same. Soloff 1989, however, reported skewed, statistically significant results for depression scores in favour of haloperidol, and Soloff 1989 and Soloff 1993) reported skewed statistically significant results using the SCL-90 sub-scale. 5.10 Mental state: 5. Schizotypal symptoms - by 6 months Soloff 1993 used the SCL-90 to measure paranoid ideation and the IMPS to measure paranoid projection and psychoticism and the SSI for schizotypal symptoms. All results were skewed and not statistically significant. However, Soloff 1989, using the SCL-90 to measure paranoid ideation, the IMPS to measure paranoid projection and perceptual distortions, the SCL-90 psychoticism and the SSI schizotypal symptoms reported statistically significant skewed results all favouring haloperidol over placebo. 5.11 Mental state: 6. Other symptoms Measures of hypersomnia (ADI subscale), leaden paralysis (ADI subscale), obsessive-compulsive symptoms (SCL-90 subscale), reactivity (ADI subscale), rejection sensitivity (ADI subscale), somatization (SCL-90 subscale) and various additional items (SCL-90

subscale)were not statistically significant. One study reported statistically significant skewed results on the SCL-90 sub-scale for obsessive-compulsive symptoms in favour of haloperidol over placebo. 5.12 Behaviour: 1a. Hostility - by 6 months Using the BDHI, one study reported statistically non-significant results for hostility (n=58, MD -4.35 CI -11.03 to 2.33). Another trial measured indirect hostility using the BDHI subscore (n=56 MD -0.58 CI -1.60 to 0.44) and verbal hostility using the BDHI subscore (n=56 MD 0.12 CI -1.50 to 1.74). 5.13 Behaviour: 1b. Hostility scores - by 6 months Soloff 1993 reported statistically non significant skewed results for general hostility using the SCL-90 subscale and hostile belligerence using the IMPS subscale, whilst Soloff 1989 reported statistically significant skewed results for general hostility, and hostile beligerence in favour of haloperidol. 5.14 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP) One study reports non statistically significant skewed results for impulsiveness. However, statistically significant skewed results were reported in Soloff 1989 favouring haloperidol. 5.15 Behaviour: 2b. Impulsiveness - traits by 6 months Self-control measured using the BIS in Soloff 1989 and Soloff 1993 did not reach conventional levels of statistical significance (n=114, WMD -1.38 CI -7.51 to 10.27). Ratings on the STIC found similar results (n=114, 2 RCTs, WMD 1.12 CI .82 to 3.07). 5.16 Leaving the study early No differences were found between people allocated antipsychotics and those on placebo for the outcome of leaving the study early (n=110, 2 RCTs, RR 1.42 CI 0.80 to 2.51). 5.17 Appetite and weight No clear differences were found by Soloff 1993 for weight gain on the ADI. 6. COMPARISON 6. ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE 6.1 Leaving the study early - before 3 weeks of treatment Leone 1982 found no statistical differences between groups for those taking loxapine compared with people allocated chlorpromazine for the outcome of leaving the study early (n=80, RR 1.20 CI 0.40 to 3.62) 6.2 Adverse effects: One or more adverse effects reported Leone 1982 found no difference between groups for reporting one or more adverse effects (n=80, RR 1.27 CI 0.66 to 2.45). 7. COMPARISON 7. ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL 7.1 Leaving the study early - by 6 weeks
17

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Serban 1984 found no difference for those leaving the study early between people taking thiothixene and those on haloperidol, (n= 52, RR 0.50 CI 0.10 to 2.50). 7.2 Adverse effects: Any adverse effect Serban 1984 did not detect any differences between groups for experiencing any adverse event (n=52, RR 0.95 CI 0.74 to 1.22). 8. COMPARISON 8. MOOD STABILISERS vs PLACEBO 8.1 Leaving the study early No difference was evident for the outcome of leaving the study early between those taking carbemazepine compared with people allocated to placebo in De la Fuente 1994 (n=20, RR 5.0 CI 0.27 to 92.62). For those in the small Hollander 2001 study (n=16), who took divalporex compared with people who took placebo, a significant difference was detected (n=16, RR 0.50 CI 0.28 to 0.88, NNT 2 CI 2 to 9) in favour of those taking the mood stabiliser. 8.2 Mental state: 1. Not responsive - by 6 months (> 2 on CGI-I) In the small Hollander 2001 study, for the outcome of did not respond measured by a score of greater than two on the CGI, none of those on placebo responded (n=4), whilst seven of those on divalporex did (n=12). This finding is statistically significant (n= 16, RR 0.58 CI 0.36 to 0.94, NNT 3 CI 2 to 17). 8.3 Mental state: 2. Depression scores - by 6 months (BDI) Hollander 2001 reports skewed results for the outcome of depression using the BDI. These data are not statistically significant. 8.4 Behaviour: 1. Acting out (AOS behavioural dyscontrol) De la Fuente 1994 measured acting out using the AOS, but results did not reach statistical significance (n=20, RR 1.33 CI 0.74 to 2.41) for those taking carbemazepine compared with people allocated placebo. 8.5 Behaviour: 2a. Aggression score - by 6 months (AQ) Hollander 2001 measured aggression using the AQ but results did not reach conventional levels of statistical significance in this small study, (n=16, MD -9.20 CI -24.46 to 6.06). 8.6 Behaviour: 2b. Aggression scores - by 6 months Using the OAS-M aggression, irritability and suicidality subscale, Hollander 2001 reports statistically non significant skewed results. 8.7 Behaviour: 3. Not improved behavioural dyscontrol (AOS) De la Fuente 1994 recorded behavioural dyscontrol recorded using the AOS, but results were not statistically significant (n=20, RR 0.88 CI 0.53 to 1.46).

able to find ten, small studies of moderate quality. This is similar to other reviews in this area. It is possible that we have failed to identify relevant work and we would be grateful for any contact from anyone with knowledge of any relevant trials. This is a difficult area of research but the production of these ten studies does demonstrate that relevant evaluative research is feasible. The small numbers of included participants in this review make type 1 errors more possible, as well as multiple testing ,using a variety of scales makes it probable that only positive findings are reported in reports of studies, whilst statistically insignificant, or neagtive results, do not get reported, leading to reporting bias. Data were lost because of unclear reporting. Should the studies we identified have reported as clearly as is now expected after the CONSORT guidelines, first formulated in 1996 (Moher 2001) considerably more might have been known on the effects of treatment of people with a borderline personality disorder. The use of scales may be of value for generating or investigating hypotheses but they are often of little clinical utility. Binary outcomes such as ’improved or not’, ’self harm or not’ should be possible to collect in this field. In scale data carrying the last observation forward is not altogether useful as this generates such large assumptions about the data that often findings are not sensible to include. 2. Applicability Participants in the included studies were recognisable to most people working in this field and the interventions potentially accessible to most health service providers. Outcomes such as admission to hospital and abstinence from drugs are also clinically important. The difficulty in interpretation of most of the scale-derived results, however, does limit clinical utility. 3. COMPARISON 1. ANTIDEPRESSANT (TRICYCLIC/ TETRACYCLIC) vs PLACEBO Antidepressants do not seem to have any over riding effect over and above placebo when it comes to improving mean scores on global levels of functioning and we found it difficult to know what the skewed results on the IMPS, SCL-90 and the IMPS subscale mean. For those taking antidepressants there are some suggestions that scores on anger are improved (NNT 2 CI 2 to 9) and perhaps depression but all findings could be the result of the play of chance and need replicated. None should be considered as conclusive. There is no conclusive indication that antidepressants prevent people from attempting suicide or reduce their hostility or impulsiveness. Having said this the recoding of the outcome ’attempted suicide by 6 months’ in Montgomery 1983 was most welcome and rare. Findings in this small study (n=38) were equivocal but if they represented a real effect this would be most important. If antidepressants were to reduce self harm by nearly 20% this would be a potent accessible treatment (see Implications for research below). Although there is no clear benefit for the antidepressant group Soloff 1989 keeps most people who started the study in treatment.
18

DISCUSSION 1. Limitation of data Despite a comprehensive search for randomised controlled trials for pharmacological therapies for people with BPD, and the increasing prevalence of this approach, this review has only been

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

This is an achievement in its own right and sets a standard for other trialists. It would seem that this question should be revisited and more trials conducted. 4. COMPARISON 2. ANTIDEPRESSANT (TRYCYCLIC) vs ANTIPSYCHOTICS The Soloff 1989 comparison only involved 62 people so it is unsurprising that no firm conclusions can be drawn. For changes to global functioning no outcome reached statistical significance. Neither amitriptyline nor haloperidol offers any advantage in terms of mental state or behaviour. Again Soloff 1989 seems to keep participants involved with their treatment with little loss to follow-up from either group. 5. COMPARISON 3. ANTIDEPRESSANT (MONOAMINE OXIDASE INHIBITER) vs PLACEBO Again this comparison only involved a total of 64 people. MAOI antidepressants did not seem any the more effective than placebo at reducing borderline symptoms, global functioning or catergorisation of type of presentation. This group of drugs give no impression of advantage in mental state response rates over placebo. However, people whose problems are related to hostility may benefit from MAOIs. 6. COMPARISON 4. ANTIDEPRESSANT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC This comparison involves 74 people. There is no advantage in taking MAOIs over haloperidol for overall borderline symptoms, but mean scores do improve though for people on MAOIs over haloperidol for general anxiety. However, there seemed to be some benefit in taking haloperidol over MAOIs for general global functioning. Neither class of drug conferred great advantage in terms of mental state measures. However, people taking phenelzine do seem to improve more than those on haloperidol for general anxiety, intropunitiveness, anxiety, depression, and schizotypal symptoms and also paranoid ideation and psychoticism and obsessive compulsive symptoms, but not hypersomnia, reactivity, rejection sensitivity and leaden paralysis. Phenelzine did not confer convincing benefits over haloperidol for outcomes of hostility with two measures not detecting any difference in hostility and one reporting that hostility improves in those taking phenelzine. However, for those with problems of hostile belligerence, phenelzine may offer an advantage over haloepridol. Phenelzine did not produce clear effects on people’s appetite or weight. 7. COMPARISON 5. ANTIPSYCHOTICS vs PLACEBO This comparison involved 180 people in three studies. There is no distinguishable benefit to taking haloperidol compared with placebo for borderline symptoms and mixed results for improvement in global functioning, with one small study reporting improvement, whilst a second finding none, with no significance overall.

In terms of mental state haloperidol does not offer any benefit to response rates over placebo for outcomes of anxiety and intropunitiveness, with mixed unconvincing results for depression. There does not seem to be any benefit for those with perceptual distortions, hypersomnia or leaden paralysis, reactivity, rejection sensitivity somatization and some unspecified symptoms, with mixed, inconclusive results for obsessive compulsive problems. However, haloperidol may offer some benefits for people with schizotypal symptoms, paranoid projections and psychotism. Behaviour as reported in ratings of hostility, hostile belligerence and impulsiveness, seems unaffected by haloperidol and there were no clear differences for leaving the study early for either haopleridol or placebo (about 25% in both groups). 8. COMPARISON 6. ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE This comparison involves 80 participants. Loxapine offers no advantage over chlorpromazine for keeping people involved with the study, and about a quarter of both groups experience at least one adverse effect. Data relevant to positive effects are lacking. 9. COMPARISON 7. ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL This comparison involves only 52 people. Thiothixene offers no advantage over haloperidol for keeping people involved with the study and most people in both groups experience adverse effects by 6 months. Data relevant to positive effects are lacking. 10. COMPARISON 8. MOOD STABILISERS vs PLACEBO This comparison involved only 34 people in two studies. In terms of mental state divalporex did seem to offer advantage over placebo for change in clinical global scores but this study involves only 16 people. For behaviour the one short, small study detected no benefits for carbemazepine for acting out, aggression, irritability or suicidality, or behavioural dyscontrol. Taking carbamazepine did not seem to offer any advantage over placebo in keeping a person in contact with the trial, whilst people taking divalporex did stay in contact more than those on placebo, but again this tiny study must be seen as hypothesis-generating rather than hypothesis-testing.

AUTHORS’ CONCLUSIONS Implications for practice 1. For people with Borderline Personality Disorder If offered medication, people with BPD should know that this is not based on good evidence from trials. That does not mean it may not do considerable good and there is no indication of significant harm. People with BPD or their carers are in a position to lobby for and facilitate good research in this area. 2. For clinicians Trial-based evidence generates as many questions as it answers. Different types of antidepressants, antipsychotics and even mood
19

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

stabilisers may be of use. Largely, trials have been small, short and poorly reported. However, any of these treatments could help but it is arguable that their use, routine or not, should be from within randomised trials. 3. For managers and policy makers For those guiding or directing care this review suggests that data are too sparse for anyone to be confident about specific direction except, perhaps, that this large important area should be the focus of much more research in order to ensure that treatment is not based on impression and fashion rather than good cleaer evidence. Implications for research 1. General 1.1 Reporting As with many similar studies, even though the trials were, by and large, too small and short to really provide conclusive outcomes, if future studies were to comply with CONSORT guidance (Moher 2001) more data would be made available for those wishing to inform practice. 1.2 Trails Although considerable criticism can be levelled at the included studies these trials are pioneering and most treatment is carried on outside of, and regardless to such evaluative work. Randomised trials with good follow up have been shown to be possible. Collaborative work in order to produce large clinically relevant data would seem an urgent necessity. 2. Specific This review has highlighted that pharmacological interventions can be used for people with borderline personality disorder and that some may influence feelings and behaviour in the short term. It would seem that more drug trials in this area are warranted. The

remaining question is what particular drugs? This group often is not treated with pharmacological interventions at all. Standard care could be seen as the placebo and it would seem that the absolute effects of any class of drugs is established rather than comparative efficacy of one against another. If the finding that antidepressants really reduce self harm by nearly 20% by six months is real (Montgomery 1983) this would be a hugely important finding. Trials of adequate power and duration comparing a safe antidepressant with placebo for clinically meaningful outcomes is needed.

POTENTIAL CONFLICT OF INTEREST None known.

ACKNOWLEDGEMENTS The authors acknowledge the UK’s National Health System Forensic R&D for supporting this work.

SOURCES OF SUPPORT External sources of support • NHS National R&D Programme on Forensic Mental Health UK Internal sources of support • University of Leeds UK • University of Leicester UK

REFERENCES

References to studies included in this review
De la Fuente 1994 {published data only} De la Fuente JM, Lotstra F. A trial of carbamazepine in borderline personality disorder. European Neuropsychopharmacology 1994; 4:479–86. De la Fuente JM, Lotstra F. Carbamazepine in borderline personality disorder. European Neuropsychopharmacology 1993;3(3):350. Goldberg 1986 {published data only} Goldberg SC, Schulz SC, Resnick RJ, Hamer RM, Schulz PM. Differential prediction of response to thiothixene and placebo in borderline and schizotypal personality disorders. Psychopharmacology Bulletin 1987;23(3):342–6. [MedLine: 3324147]. Goldberg SC, Schulz SC, Schulz PM, Resnick RJ, Hamer RM, Friedel R.O. Borderline and schizotypal personality disorders treated

with low-dose thiothixene vs. placebo. Archives of General Psychiatry 1986;43:680–6. [MedLine: 3521531]. Hollander 2001 {published data only} Hollander E, Allen A, Lopez RP, Bienstock CA, Grossman R, Siever LJ, Merkatz L, Stein DJ. A preliminary double-blind, placebo-controlled trial of divalproex sodium in borderline personality disorder. Journal of Clinical Psychiatry 2001;62(3):199–203. Leone 1982 {published data only} Leone NF. Response of borderline patients to loxapine and chlorpromazine. Journal of Clinical Psychiatry 1982;43(4):148–50. Montgomery 1983 {published data only} Montgomery SA, Roy D, Montgomery DB. The prevention of recurrent suicidal acts. British Journal of Clinical Pharmacology 1983; 15:183–8S.
20

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Salzman 1995 {published data only} Salzman C, Wolfson AN, Schatzberg A, Looper J, Henke R, Albanese M, Schwartz J, Miyawaki E. Effect of fluoxetine on anger in symptomatic volunteers with borderline personality disorder. Journal of Clinical Psychopharmacology 1995;15(1):23–9. Serban 1984 {published data only} Serban G, Siegel S. Response of borderline and schizotypal patients to small doses of thiothizene and haloperidol. American Journal of Psychiatry 1984;141(11):1455–8. Soloff 1989 {published data only} Soloff P, George A, Nathan RS, Schulz PM, Ulrich RF, Perel JM. Progress in pharmacotherapy of borderline disorders: A double-blind study of amitryptyline, haloperidol, and placebo. Archives of General Psychiatry 1986;43:691–7. Soloff PH, George A, Nathan RS. Paradoxical effects of amitriptyline on borderline patients. American Journal of Psychiatry 1986;143(12): 1603–5. Soloff PH, George A, Nathan RS, Schulz PM, Cornelius JR, Herring J, Perel JM. Amitriptyline versus haloperidol in borderlines: Final outcomes and predictors of response. Journal of Clinical Psychopharmacology 1989;9(4):238–46. Soloff PH, George A, Nathan RS, Schulz PM, Perel JM. Behavioral dyscontrol in borderline patients treated with amitriptyline. Psychopharmacology Bulletin 1987;23(1):177–81. Soloff PH, George A, Nathan S, Schulz PM, Ulrich RF, Perel JM. Amitriptyline and haloperidol in unstable and schizotypal borderline disorders. Psychopharmacology Bulletin 1986;22(1):177–82. Soloff 1993 {published data only} Cornelius JR, Soloff PH, George A, Ulrich RF, Perel J. Haloperidol vs. phenelzine in continuation therapy of borderline disorder. Psychopharmacology Bulletin 1993;29(2):333–7. Cornelius JR, Soloff PH, Perel JM, Ulrich RF. Continuation pharmacotherapy of borderline personality disorder with haloperidol and phenelzine. American Journal of Psychiatry 1993;150(12):1843–8. Soloff PH, Cornelius J, Georger A, Swami N, Perel JM, Ulrich RF. Efficacy of phenelzine and haloperidol in borderline personality disorder. Archives of General Psychiatry 1993;50(5):377–85. Zanarini 2001 {published data only} ∗ Zanarini MC, Frankenburg FR. Olanzapine treatment of female borderline personality disorder patients: A double-blind, placebocontrolled pilot study. Journal of Clinical Psychiatry 2001;62(11): 849–54.

Chengappa 1999 Chengappa KN, Ebeling T, Kang JS, Levine J, Parepally H. Clozapine reduces severe self-mutilation and aggression in psychotic patients with borderline personality disorder. Journal of Clinical Psychiatry 1999;60(7):477–84. [MedLine: 10453803]. Cowdry 1988 Cowdry RW, Gardner DL. Pharmacotherapy of borderline personality disorder. Alprazolam, carbamazepine, trifluoperazine, and tranylcypromine. Archives of General Psychiatry 1988;45(2):111–9. Gardner DL, Cowdry RW. Alprazolam-induced dyscontrol in borderline personality disorder. American Journal of Psychiatry 1985;142 (1):98–100. Gardner DL, Cowdry RW. Positive effects of carbamazepine on behavioral dyscontrol in borderline personality disorder. American Journal of Psychiatry 1986;143(4):519–22. Links 1990 ∗ Links PS, Steiner M, Boiago I, Irwin D. Lithium therapy for borderline patients: Preliminary findings. Journal of Personality Disorders 1990;4(2):173–81.

Additional references
Altman 1996 Altman DG, Bland JM. Detecting skewness from summary information. BMJ 1996;313(7066):1200. APA 1994 American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV). 4th Edition. Washington DC: American Psychiatric Association, 1994. Barratt 1965 Barratt EL. Factor analysis of some psychometric measures of impulsiveness and anxiety. Psychological reports 1965;16:547–54. Beck 1961 Beck AT, Ward CH, Mendelsohn M, Mock J, Erbaugh J. An inventory for measuring depression. Archives of General Psychiatry 1961;4: 561–71. Benjamin 1993 Benjamin LS. Interpersonal Diagnosis and Treatment of Personality Disorders. 2nd Edition. New York: Guilford, 1998. Bland 1997 Bland JM. Statistics notes. Trials randomised in clusters. BMJ 1997; 315:600. BNF 2003 British Medical Association and Royal Pharmaceutical Society of Great Britain. BNF. Vol. 46, London: British Medical Association and Royal Pharmaceutical Society of Great Britain, 2003. Buss 1957 Buss AH, Durkee A. An inventory for assessing different kinds of hostility. Journal of Consulting Psychology 1957;21:343–9. Clarke 2002 Clarke M, Oxman AD. Cochrane Collaboration Handbook. Oxford: Update Software, 2002. Coid 1993 Coid JW. An affective syndrome in psychopaths with borderline personality?. British Journal of Psychiatry 1993;163:641–50.
21

References to studies excluded from this review
Benedetti 1998 Bendetti F, Sforzini L, Colombo C, Maffei C, Smeraldi E. Low-dose clozapine in acute and continuation treatment of severe borderline personality disorder. Journal of Clinical Psychiatry 1998;59(3):103–7. Bohus 1999 Bohus MJ, Landwehrmeyer GB, Stiglmayr CE, Limberger MF, Bohme R, Schmahl CG. Naltrexone in the treatment of dissociative symptoms in patients with borderline personality disorder: an openlabel trial. Journal of Clinical Psychiatry 1999;60(9):598–603. [MedLine: 10520978].

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Conte 1980 Conte HR, Plutchik R, Karasu TB, Jerrett I. A self-report borderline scale: discriminative validity and preliminary norms. Journal of Nervous and Mental Disease 1980;168:428–35. [MedLine: 7400793]. Cowdry 1988a Cowdry RW, Gardner DL. Pharmacotherapy of borderline personality disorder. Alprazolam, carbamazepine, trifluoperazine, and tranylcypromine. Archives of General Psychiatry 1988;45(2):111–9. [MedLine: 3276280]. Derogatis 1977 Derogatis LR. SCL-90-R: Administration Scoring and Procedures Manual. Baltimore: Clinical Psychometrics Research, 1977. Divine 1992 Divine GW, Brown JT, Frazier LM. The unit of analysis error in studies about physicians’ patient care behavior. Journal of General Internal Medicine 1992;7(6):623–9. Donner 2002 Donner A, Klar N. Issues in the meta-analysis of cluster randomized trials. Statistics in Medicine 2002;21:2971–80. Egger 1997 Egger M, Davey-Smith G, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315:629–34. Endicott 1976 Endicott J, Spitzer RL, Fleiss JL, Cohen J. The Global Assessment Scale: a procedure for measuring overall severity of psychiatric disturbance. Archives of General Psychiatry 1976;33(6):766–71. First 1997 First MB. Spitzer RL, Gibbon M, Williams JBW. Structured Clinical Interview for DSM-IV Personality Disorders, (SCID-II). Washington, DC: American Psychiatric Press Inc, 1997. Gulliford 1999 Gulliford MC. Components of variance and intraclass correlations for the design of community-based surveys and intervention studies: data from the Health Survey for England 1994. American Journal of Epidemiology 1999;149:876–83. Gunderson 1981 Gunderson JG, Kolb JE, Austin V. The diagnostic Interview for borderline patients. American Journal of Psychiatry 1981;138:896–903. Gunderson 1983 Gunderson JG. Empirical studies of the borderline diagnosis. In: GrinspoonL editor(s). Psychiatry Annual Review. Washington, DC: American Psychiatric Association, 1983:415–36. Guy 1976 Guy W. Early clinical drug evaluation (ECDEU) assessment manual for psychopharmacology. Washington, DC: National Institute of Mental Health, 1976:217–22. Hamilton 1960 Hamilton M. A rating scale for depression. Journal of Neurology Neurosurgery and Psychiatry 1960;23:56–62. Higgins 2003 Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60.

ICD-10 1992 World Health Organization. The ICD-10 Classification of Mental and Behavioural Disorders. Geneva: World Health Organization, 1992. Kernberg 1967 Kernberg OF. Borderline personality organisation. Journal of the American Psychoanalytic Association 1967;15:641–85. Kernberg 1975 Kernberg OF. Borderline conditions and primary narcissism. New York: Jason Aronson, 1975. Lazzaro 1969 Lazzaro TA, Biggs DL, McNeil KA. The development and validation of the self-report test of impulse control. Journal of Clinical Psychology 1969;25(4):434–8. Liebowitz 1981 Liebowitz MR, Klein DF. Interrelationship of hysteroid dysphoria and borderline personality disorder. The Psychiatric clinics of North America 1981;4:67–87. Linehan 1997 Linehan MM. Behavioral treatments of suicidal behaviors. Definitional obfuscation and treatment outcomes. Annals of the New York Academy of Sciences 1997;836:302–28. [MedLine: 9616806]. Lorr 1962 Lorr M, McNair DM, Klett CJ, Lasky JJ. Evidence of ten psychotic symptoms. Journal of Consulting Psychology 1962;26:185. Marshall 2000 Marshall M, Lockwood A, Bradley C, Adams C, Joy C, Fenton M. Unpublished rating scales: a major source of bias in randomised controlled trials of treatments for schizophrenia. British Journal of Psychiatry 2000;176:249–52. [MedLine: 10755072]. McGlashan 1986 McGlashan TH. The Chestnut Lodge follow-up study 111: longterm outcome of borderline personalities. Archives of General Psychiatry 1986;43(1):2–30. [MedLine: 3942471]. McNair 1971 McNair D, Lorr M, Droppleman L. EITS Manual for the profile of mood states educational and training testing service. San Diego, CA: Educational and Testing Services, 1971. Moher 2001 Moher D, Schulz KF, Altman D. The CONSORT statement: revised recommendations for improving the quality of reports of parallelgroup randomized trials. JAMA 2001;285(15):1987–91. NIMHE 2003 National Institute for Mental Health for England. Personality Disorder: No longer a diagnosis of exclusion. London: NIMHE:DOH, 2003. Paris 1987 Paris J, Brown R, Nowlis D. Long-term follow-up of borderline patients in a general hospital. Comprehensive Psychiatry 1987;28(6):47– 115. [MedLine: 3691077]. Perry 1993 Perry JC. Longitudinal studies of personality disorders. Journal of Personality Disorders 1993;Suppl.1:63–85.
22

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Plakum 1985 Plakun EM, Burkhardt PE, Muller JP. 14 year follow-up of borderline and schizotypal personality disorders. Comprehensive Psychiatry 1985; 26(5):448–55. [MedLine: 4028695]. Soloff 1998 Soloff PH. Symptom-oriented psychopharmacology for personality disorders. Journal of Practical Psychiatry and Behavioral Health 1998; 4:3–11. Stone 1990 Stone MH. The fate of borderline patients: successful outcome and psychiatric practice. New York: Guilford, 1990. Stone 1993 Stone MH. Abnormal personalities: within and beyond the realm of treatment. New York: WW Norton and Co, 1997. Ukoumunne 1999 Ukoumunne OC, Gulliford MC, Chinn S, Sterne JAC, Burney PGJ. Methods for evaluating area-wide and organisation-based interventions in health and health care: a systematic review. Health Technology Assessment 1999;3(5):iii–92. [MedLine: 10982317]. Williams 1988 Williams JB. A structured interview guide for the Hamilton Depression Rating Scale. Archives of General Psychiatry 1988;45(8):742–7. [MedLine: 3395203].
∗

Indicates the major publication for the study

TABLES

Characteristics of included studies
Study Methods De la Fuente 1994 Allocation: randomised, no further details. Blinding: double, no further details. Duration: 32 days (after 10+ days washout). Setting: inpatient. Diagnosis: borderline personality disorder (DSM-III-R criteria, DIB). N=20. Age: mean ~ 33 years. Sex: 6M, 14F. History: not stated. Exclusions: DSM-III-R Axis I disorders (although patients depressed for <2 weeks were included), history of epilepsy or EEG traits of epilepsy, inability to stop using alcohol and illicit substances and suspected poor treatment compliance. neuroleptic medication taken in two months prior to study. 1. Carbamazepine: administered orally to obtain plasma drug levels usually required for the management of epileptic and affectively ill patients. N=10. 2. Placebo: N=10. Also received supportive atheoretical psychotherapy Outcomes Behaviour: behavioural dyscontrol (acting out, AOS).
23

Participants

Interventions

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of included studies (Continued )
Leaving the study early. Unable to use Global state: GAS (N values not expllicitly stated). Mental state: BPRS, HDRS-24, SCL-90 (N values not explicitly stated). Notes Allocation concealment Study Methods Reviewers assume poor outcome when data missing on dichotomous variables. A – Adequate Goldberg 1986 Allocation: randomised, no further details. Blinding: double, no further details. Duration: 12 weeks (after 1 week washout). Setting: outpatient. Diagnosis: borderline and/or schizotypal personality disorder (17 borderline personality disorder, 13 schizotypal personality disorder, and 20 both diagnoses(SIB)). N=50. Age: mean ~ 32 years. Sex: 21M, 29F. History: present illness for one year or more, at least one psychotic symptom, 88% never been hospitalised before. Exclusions: schizophrenia, mania, melancholia, organic brain syndrome/mental retardation, current alcoholism or drug addiction and a variety of medical disorders. 1. Thiothixene hydrochloride: initial dose 5mg capsule, increased on each succeeding visit by 5mg capsule until marked improvement or side effects. Max dose 40mg (8 capsules). N=24. 2. Placebo: identical capsules containing lactose. administration schedule same as for thiothixene hydrochloride, up to maximum of 8 capsules. N=26. Benztropine mesylate (1mg tablets) provided in event of extrapyramidal symptoms. Outcomes Leaving the study early. Unable to use Global state: GAS (no SD) Mental state: SCL-90, SIB (no SD). Notes Allocation concealment Study Methods A – Adequate Hollander 2001 Allocation: randomised (3:1 assignment). Blinding: double, no further details. Duration: 10 weeks. Setting: outpatient. Diagnosis: 16 borderline personality disorder (SCID-II, DSM-IV). N=21* Age: mean ~ 39 years. Sex: 10M, 11F. History: not stated. Exclusion: other medical or neurological illness, psychotic disorders, current substance abuse, bipolar disorder (type I and II), current major depression, current suicidal ideation , pregnancy. 1. Divalproex sodium: initial dose 250mg at bedtime, increased gradually to maintain a blood valproate level of 80 ug/mL or highest tolerated dose. N=12*. 2. Placebo. N=4*.
24

Participants

Interventions

Participants

Interventions

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of included studies (Continued )
Outcomes Global state: CGI-I. Mental state: BDI. Behaviour: OAS-M, AQ. Leaving the study early. Unable to use Global state: GAS (mean and SD only reported for divalproex sodium group. Notes Allocation concealment Study Methods * initial sample of 21, only 16 randomised to treatment. A – Adequate Leone 1982 Allocation: randomised, no further details. Blinding: double, identical capsules. Duration: 6 weeks. Setting: outpatient. Diagnosis: borderline personality disorder (Gunderson and Kolb 1978). N=80. Age: mean ~ 31 years. Sex: 32M, 48F. History: acute disruptive symptoms. Exclusions: known allergy / hypersensitivity to study medication, moderate to severe mental retardation or organic brain syndrome, severe medical condition, use of sedatives or tranquilisers, treated with other psychotropic drugs in 48 hours prior to study. 1. Loxapine succinate: mean dose 14.5mg/day. N=40. 2. Chlorpromazine: mean dose 110mg/day. N=40. No other psychotropic medication, fluorazepam and chloral hydrate permitted as sedatives. Outcomes Adverse effects: various side effects. Leaving the study early. Unable to use Global state: CGI (no mean, SD). Mental state: BPRS, POMS, SNOOP (no mean, SD). Notes Allocation concealment Study Methods A – Adequate Montgomery 1983 Allocation: randomised, no further details. Blinding: double, no further details. Duration: 6 months. Setting: outpatient. Diagnosis: 26 borderline personality disorder, 8 histrionic personality disorder and 4 both diagnoses (ICD-9 and DSM-III).** N=58.* Sex: 20M, 38F. Age: mean ~ 35 years. History: admission to a medical ward following a suicidal act, history of at least two acts of deliberate selfharm Exclusions: other physical illness, overt schizophrenia or depression. 1. Mianserin: 30mg dose. N=17**. 2. Placebo. N=21**.
25

Participants

Interventions

Participants

Interventions

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of included studies (Continued )
All followed-up in daily clinic with access to social workers, community nurses and a crisis intervention team. Outcomes Behaviour: suicide attempt. Unable to use Mental state: MADRS (data not reported). Leaving the study early (no data for mianserin group). Notes Allocation concealment Study Methods * numbers randomised to treatment. ** numbers of people who completed, data only reported for those finishing study. A – Adequate Salzman 1995 Allocation: randomised, no further details. Blinding: double, identical capsules. Duration: 12 weeks (after 1 week washout). Setting: community. Diagnosis: borderline personality disorder or significant traits without meeting full criteria (DIB-R, SCIDII, DSM-III-R). N=27*. Age: mean ~ 36 years. Sex: 8M, 14F.** History: not stated. Exclusions: inpatient or history of hospitalisation, major depression or other Axis I disorder, current or recent severe symptoms (e.g.suicidal behaviour), additional secondary Axis II disorder, self-mutilating behaviours in past four years, present history of substance abuse, use of other psychotropic medication. 1. Fluoxetine: intial dose 20mg capsule, titrated up to a maximum 3 capusles (60mg/day) according to need. N=13**. 2. Placebo: identical capsules. N=9**. Mental State: improved/not improved HAM-D, PDRS, POMS. Unable to use: Global state: GAS (no mean, SD). Mental state: PDRS, POMS, HAM-D (no mean, SD). Behaviour: OAS-R (reviewers uncertain as to whether scale is a published rating scle and/or in the public domian for two years prior to its reporting). Leaving the study early: (data not reported separately by treatment group). Notes Allocation concealment Study Methods * number randomised to treatment. ** number of people who completed, data only reported for those finishing study. A – Adequate Serban 1984 Allocation: randomised, computer generated code. Blinding: double, no further details. Duration: 3 months (after 1 week washout). Setting: outpatient. Diagnosis: 16 borderline personality disorder, 14 schizotypal personality disorder, 16 mixed types of these personality diagnoses (DSM-III criteria, with diagnosis rechecked using BSI). N=52. Age: mean ~ 32 years. Sex: 36M, 16F.
26

Participants

Interventions

Outcomes

Participants

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of included studies (Continued )
History: personality disorder diagnosis for at least 2 years, no pre-existing diagnosis of schizophrenia or affective illness, experienced mild transient psychotic episode prior to admission. Exclusions: not stated. 1. Thiothixene: 2mg b.i.d, dose varied throughout the study according to therapeutic need of patient, mean dose 9.4mg/day. N=26. 2. Haloperidol: 0.8mg b.i.d, dose varied throughout the study according to therapeutic need of patient, mean dose 3mg/day. N=26. Chloral hydrate or antiparkinsonian medication if needed. Outcomes Adverse effects: various side effects. Unable to use: Mental state: BSI (no SD) PAI, HAM-D (no N values reported explicitly, although implied). Leaving the study early: (data not fully) reported. Notes Allocation concealment Study Methods D – Not used Soloff 1989 Allocation: randomised, no further details. Blinding: double, no further details. Duration: 5 weeks (after 1 week washout). Setting: inpatient (after 3 weeks some allowed to complete as outpatients). Diagnosis: borderline personality disorder (DIB), subtyped by DMS-III into 35 borderline personality disorder, 4 schizotypal personality disorder and 51 schizotypal-borderline mixed disorder. N=90. Age: mean ~ 25 years. Sex: 22M, 68F. History: not chronically ill. Exclusions: schizophrenia, mania or related psychotic disorders. 1. Haloperidol: 4-16mg/day, mean dose 4.8mg/day by day 35. N=31. 2. Amitriptyline: 100-175mg/ day. mean dose 149mg/day by day 35. N=30. 3. Placebo: up to 6 tablets/day. N=29. In addition to usual group milieu or individual therapies on the inpatient unit and biperiden hydrochloride (2mg) as needed for extrapyramidal symptoms. Outcomes Global state: GAS. Mental state: BDHI, HAM-D-24, HAM-D-17, BDI, IMPS, SCL-90, SSI. Behaviour: impulsivity:, WSIAP, BIS, STIC. Unable to use Mental state: Hysteroid Dysphoria Questionnaire (unable to find any details of validity or relaibility or publication). Behaviour: behavioural dyscontrol (incomplete data). Leaving the study early (incomplete data). Notes Allocation concealment Study Methods A – Adequate Soloff 1993 Allocation: randomised, no further details. Blinding: double, no further details. Duration: 5 weeks (after 1 week washout)*.
27

Interventions

Participants

Interventions

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of included studies (Continued )
Setting: inpatients for minimum of 2 weeks before discharge to an outpatient borderline disorders research clinic. Diagnosis: borderline personality disorder (DIB), subtyped by DSM-III-R into 42 borderline personality disorder and 66 schizotypal/ borderline personality disorder. N=108. Age: 16 - 36 years, mean ~ 27 years. Sex: 26M, 82F. History: not stated. Exclusions: current and/or lifetime diagnoses of schizophrenia, schizoaffective disorder, manic disorder, bipolar disorder with mania or hypomania or psychotic major depressive disorder; substance abuse, evidence of CNS disease (includes recent ECT or seizure disorder); physical disorders of known psychiatric consequence; or borderline mental retardation. 1. Haloperidol: titrated up to doses of 4mgday within 1 week, could be increased up to 6mg/day after two weeks according to patient need, mean dose after three weeks treatment ~ 3.93mg/day. N=36. 2. Phenelzine sulfate: titrated up to doses of 60mg/day within 1 week, could be increased up to 90mg/day after two weeks according to patient need, mean dose after three weeks ~ 60.45mg/day. N=38. 3. Placebo: titrated up to doses of 4 tablets/day of placebo within 1 week, could be increased up to six tablets/day after two weeks intitial medication according to patient need, mean dose after three weeks treatment ~ 4.31 tablets/day. N=34. Outpatients seen weekly for supportive psychotherapy. Outcomes Global state: GAS. Mental state: ADI, BDHI, BDI, BIS, BSI, HAM-D-17, HAM-D-24, IMPS, SCL-90, SSI, STIC, WSIAP. Leaving the study early. Unable to use Mental state: Hysteroid Dysphoria Questionnaire (unable to find any details of validity or relaibility or publication). Notes Allocation concealment Study Methods * after study 54 participants entered a continuation pharmacotherapy trial lasting 16 weeks. Only medication responders entered the continuation study, consequently these data not entered into analysis. A – Adequate Zanarini 2001 Allocation: randomised, 2:1 random number sequence. Blinding: double, identical capsules. Duration: six months. Setting: community. Diagnosis: borderline personality disorder (DIB-R, SCID-II). N=28. Age: mean ~ 27 years. Sex: all female. History: recently disturbed by moodiness, distrustfellness, impulsivity and painful and difficult relationships. Exclusions: major depression, schizophrenia, scizoaffective disorder, or bipolar disorder, previous tretment with olanzapine, medically ill, seizure disorders, current use of psychotropic medication the participants thought were helpful, substance abuse, at risk of sucidal, pregnancy, at risk of pregnancy or breastfeeding. 1. Olanzapine: starting dose 2.5mg/day, dose adjusted according to perceived response and side effects, mean dose ~ 5.33mg/day. N=19. 2. Placebo: N=9. Leaving the study early. Unable to use Global state: GAF (>50% loss)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 28

Participants

Interventions

Participants

Interventions

Outcomes

Mental state: HDI, PANSS; SCL-90 (>50% loss). Adverse effects: AIMS, Barnes, constipation, sedation, weight gain (>50% loss). Notes Allocation concealment A – Adequate
ADI - Atypical Depression Inventory AIMS - Abnormal Involuntary Movement Scale AOS - Acting Out Scale AQ - Aggression Questionnaire BARNES- Barnes Akathisia Scale BDHI - Buss Durkee Hostility Inventory BDI - Beck Depression Inventory BIS - Barratt Impulsiveness Scale BPRS - Brief Psychiatric Rating Scale BSI - Borderline Syndrome Index CGI - Clinical Global Impression Scale CGI-I - Clinical Global Impressions - Improvement Scale DES - Dissociative Experience Scale DIB - Diagnostic Interview for Borderlines DIB-R - Revised Diagnostic Interview for Borderlines DSM-III - Diagnostic and Statistical Manual of Mental Disorders, 3rd edition DSM-III-R - Diagnostic and Statistical Manual of Mental Disorders, 3rd edition, revised GAF - Global Assessment of Functioning GAS - Global Assessment Scale HAM-D - Hamilton Rating Scale for Depression HAM-D-17 - Hamilton Depression Scale - 17 item version HAM-D-24 - Hamilton Depression Scale - 24 item version HDI - Hamilton Depression Inventory HDQ - Hysteroid Dysphoria Questionnaire ICD-9 - International Classification of Disease, 9th edition IMPS - Inpatient Multidimensional Rating Scale MADRS - Montgomery and Asberg Depression Rating Scale OAS-M - Overt Aggression Scale - Modified OAS-R - McLean Hospital Overt Aggression Symptom Checklist PAI - Psychiatric Assessment Interview PANSS - Positive and Negative Syndrome Scale PDRS - Personality Disorder Rating Scale POMS - Profile of Mood States SCID-II (DSM-III-R) - Structured Clinical Interview for DSM-III-R Axis II disorders SCID-II (DSM-IV) - Structured Clinical Interview for DSM-IV Axis II disorders SCL-90 - Hopkins Symptom Checklist-90 SIB - Schedule for Interviewing Borderlines SNOOP - Systematic Nurses’ Observation of Psychopathology SSI - Schizotypal Symptom Inventory STIC - Self Report Test of Impulse Control WSIAP - Ward Scale of Impulse Action Patterns

Characteristics of excluded studies
Study Benedetti 1998 Bohus 1999 Chengappa 1999 Cowdry 1988 Reason for exclusion Allocation: not randomised, case series. Allocation: not randomised. Allocation: not randomised, case series. Allocation: randomised, crossover trial.
29

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Characteristics of excluded studies (Continued )
Participants: people with borderline personality disorder primary diagnosis, other personality disorder diagnoses included: dependent (N=15), avoidant (N=13), histrionic (N=10), schizotypal (N=6), narcissistic (N=3), and paranoide (N=2). N= 16. Interventions: alprazolam, carbamazepine, trifluoperazine, tranylcypromine and placebo. Outcome: leaving the study early, depression, anxiety, rejection sensitivity, euphoria, capacity for pleasure, impulsivity, suicidality and behvaioural dyscontrol: data not usuable - no separate data for first arm of cross-over trial. Links 1990 Allocation: randomised crossover trial. Participants: people with borderline personality disorder (DIB). N=unclear. Interventions: lithium carbonate, desipramine and placebo. Outcome: leaving the study early, depression, anger and suicide symptoms: data not usable - no separate data for first arm of crossover trial.

DIB - Diagnostic Interview for Borderlines. DIB-R - Revised Diagnostic Interview for Bordelines. SCID-II - Structured Clinical Interview for DSM-IV AXIS II disorders.

ANALYSES

Comparison 01. ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO
No. of studies 1 No. of participants 57

Outcome title 01 Global state: 1. Mean score - by 6 months (GAS, higher scores = better) 02 Global state: 2. Mean symptom score - by 6 months (skewed data) 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data) 04 Mental state: 1. Anger - not improved by 6 months (POMS not 20% over baseline score. high score=poor) 05 Mental state: 2a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) 06 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data) 07 Mental state: 3a. Depression not improved - by 6 months 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) 09 Mental state: 4. Mood disturbance - not improved by 6 months (POMS, not 20% over baseline score)

Statistical method Weighted Mean Difference (Fixed) 95% CI

Effect size 3.32 [-2.91, 9.55]

Other data

No numeric data

Other data

No numeric data

1

22

Relative Risk (Fixed) 95% CI

0.30 [0.10, 0.85]

1

57

Weighted Mean Difference (Fixed) 95% CI

-0.39 [-4.87, 4.09]

Other data

No numeric data

Relative Risk (Fixed) 95% CI Other data

Subtotals only No numeric data

1

22

Relative Risk (Fixed) 95% CI

0.35 [0.12, 1.03]

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

30

10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) 11 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed) 12 Behaviour: 1. Attempting suicide within 6 months 13 Behaviour: 2a. Hostility indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) 14 Behaviour: 2b. Hostility scores - by 6 months (higher scores = poor, skewed data) 15 Behaviour: 3a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) 16 Behaviour: 3b. Impulsiveness traits by 6 months 17 Leaving the study early: Completed <2 weeks of treatment

Other data

No numeric data

Other data

No numeric data

1 1

38 57

Relative Risk (Fixed) 95% CI Weighted Mean Difference (Fixed) 95% CI

0.82 [0.44, 1.54] -0.67 [-1.57, 0.23]

Other data

No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI 1 59 Relative Risk (Fixed) 95% CI

Subtotals only 0.97 [0.06, 14.74]

Comparison 02. ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC
No. of studies 1 No. of participants 57

Outcome title 01 Global state: 1. Global functioning score - by 6 months (GAS, higher scores = better) 02 Global state: 2. Mean symptom score - by 6 months (skewed data) 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data) 04 Mental state: 1a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) 05 Mental state: 1b. Anxiety scores - by 6 months (higher scores = poor, skewed data) 06 Mental state: 2. Depression scores - by 6 months (higher scores = poor, skewed data) 07 Mental state: 3. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)

Statistical method Weighted Mean Difference (Fixed) 95% CI

Effect size -3.87 [-10.67, 2.93]

Other data

No numeric data

Other data

No numeric data

1

57

Weighted Mean Difference (Fixed) 95% CI

2.32 [-1.93, 6.57]

Other data

No numeric data

Other data

No numeric data

Other data

No numeric data

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

31

08 Mental state: 4. Other symptoms - by 6 months (high scores = poor, data skewed) 09 Behaviour: 1a. Hostility indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) 10 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) 11 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) 12 Behaviour: 2b. Impulsiveness traits by 6 months 13 Leaving the study early: Completed <2 weeks of treatment

Other data

No numeric data

1

57

Weighted Mean Difference (Fixed) 95% CI

-0.09 [-0.96, 0.78]

Other data

No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI 1 61 Relative Risk (Fixed) 95% CI

Subtotals only 0.34 [0.04, 3.13]

Comparison 03. ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO
Outcome title 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data) 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) 06 Mental state: 2a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90 subscale higher scores = poor, skewed data) 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) No. of studies No. of participants Other data Statistical method Effect size No numeric data

1

64

Weighted Mean Difference (Fixed) 95% CI

1.67 [-4.15, 7.49]

1

62

Weighted Mean Difference (Fixed) 95% CI

0.10 [-10.36, 10.56]

Other data

No numeric data

1

72

Relative Risk (Fixed) 95% CI

0.93 [0.71, 1.22]

1

62

Weighted Mean Difference (Fixed) 95% CI

-1.51 [-5.38, 2.36]

Other data

No numeric data

Other data

No numeric data

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

32

09 Mental state: 4. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) 10 Mental state: 5. Other symptoms - by 6 months (high scores = poor, data skewed) 11 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor) 12 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) 13 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) 14 Behaviour: 2b. Impulsiveness traits by 6 months 15 Appitite and weight (higher scores = poor, skewed data)

Other data

No numeric data

Other data

No numeric data

1

62

Weighted Mean Difference (Fixed) 95% CI

-9.19 [-16.12, -2.26]

Other data

No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI Other data

Subtotals only No numeric data

Comparison 04. ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC
No. of studies No. of participants Other data

Outcome title 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data) 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) 06 Mental state: 2a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90, higher scores = poor, skewed data)

Statistical method

Effect size No numeric data

1

64

Weighted Mean Difference (Fixed) 95% CI

5.15 [0.29, 10.01]

1

64

Weighted Mean Difference (Fixed) 95% CI

-3.57 [-11.66, 4.52]

Other data

No numeric data

1

74

Relative Risk (Fixed) 95% CI

0.83 [0.65, 1.05]

1

64

Weighted Mean Difference (Fixed) 95% CI

-3.88 [-7.51, -0.25]

Other data

No numeric data

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

33

08 Mental state: 3a. Depression scores - by 6 months (HAMD-17, high score=poor)) 09 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) 10 Mental state: 4. Dysphoria - hysteroid dysphoria - by 6 months (HDQ, higher scores = poor) 11 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) 12 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed) 13 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor) 14 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) 15 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) 16 Behaviour: 2b. Impulsiveness traits by 6 months 17 Appetite and weight (higher scores = poor, skewed data)

1

64

Weighted Mean Difference (Fixed) 95% CI

-7.86 [-10.51, -5.21]

Other data

No numeric data

1

64

Weighted Mean Difference (Fixed) 95% CI

-0.82 [-2.51, 0.87]

Other data

No numeric data

Other data

No numeric data

1

64

Weighted Mean Difference (Fixed) 95% CI

-4.84 [-11.94, 2.26]

Other data

No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI Other data

Subtotals only No numeric data

Comparison 05. ANTIPSYCHOTICS vs PLACEBO

Outcome title 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data) 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)

No. of studies

No. of participants Other data

Statistical method

Effect size No numeric data

2

114

Weighted Mean Difference (Fixed) 95% CI

1.75 [-2.37, 5.86]

2

114

Weighted Mean Difference (Fixed) 95% CI

-1.86 [-10.85, 7.14]

Other data

No numeric data

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

34

05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) 06 Mental state: 2a. Anxiety score intropunitiveness by 6 months (IMPS, higher scores = poor) 07 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data) 08 Mental state: 3a. Depression scores - by 6 months (HAMD-17 higher scores = poor) 09 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data) 10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor) 11 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed) 12 Behaviour: 1a. Hostility - by 6 months (higher scores = poor) 13 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data) 14 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed) 15 Behaviour: 2b. Impulsiveness traits by 6 months 16 Leaving the study early 17 Weight gain score (ADI subscale, higher scores = poor, skewed data)

1

70

Relative Risk (Fixed) 95% CI

0.77 [0.37, 1.63]

2

114

Weighted Mean Difference (Fixed) 95% CI

-0.36 [-3.30, 2.58]

Other data

No numeric data

1

56

Weighted Mean Difference (Fixed) 95% CI

-1.44 [-4.41, 1.53]

Other data

No numeric data

Other data

No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI Other data

Subtotals only No numeric data

Other data

No numeric data

Weighted Mean Difference (Fixed) 95% CI 2 110 Relative Risk (Fixed) 95% CI Other data

Subtotals only 1.42 [0.80, 2.51] No numeric data

Comparison 06. ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE

Outcome title

No. of studies

No. of participants

Statistical method

Effect size

01 Leaving the study early - before 3 weeks of treatment 02 Adverse effects: One or more adverse effects reported

1 1

80 80

Relative Risk (Fixed) 95% CI Relative Risk (Fixed) 95% CI

1.20 [0.40, 3.62] 1.27 [0.66, 2.45]
35

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Comparison 07. ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL
Outcome title 01 Leaving the study early - by 6 weeks 02 Adverse effects: Adverse reactions to medication by 6 months No. of studies 1 1 No. of participants 52 52 Statistical method Relative Risk (Fixed) 95% CI Relative Risk (Fixed) 95% CI Effect size 0.50 [0.10, 2.50] 0.95 [0.74, 1.22]

Comparison 08. MOOD STABILISERS vs PLACEBO
Outcome title 01 Leaving the study early 02 Mental state: 1. Not responsive - by 6 months (> 2 on CGI-I) 03 Mental state: 2. Depression scores - by 6 months (BDI, higher scores = poor, skewed data) 04 Behaviour: 1. Acting out by 6 months (AOS behavioural dyscontrol) 05 Behaviour: 2a. Aggression score - by 6 months (AQ, higher scores = poor) 06 Behaviour: 2b. Aggression scores - by 6 months (higher scores = poor, skewed data). 07 Behaviour: 3. Not improved behavioural dyscontrol (AOS) No. of studies 1 No. of participants 16 Statistical method Relative Risk (Fixed) 95% CI Relative Risk (Fixed) 95% CI Other data Effect size Subtotals only 0.58 [0.36, 0.94] No numeric data

1

20

Relative Risk (Fixed) 95% CI

1.33 [0.74, 2.41]

1

16

Weighted Mean Difference (Fixed) 95% CI

-9.20 [-24.46, 6.06]

Other data

No numeric data

1

20

Relative Risk (Fixed) 95% CI

0.88 [0.53, 1.46]

INDEX TERMS Medical Subject Headings (MeSH) Antidepressive Agents [therapeutic use]; Antipsychotic Agents [∗ therapeutic use]; Borderline Personality Disorder [∗ drug therapy; psychology]; Randomized Controlled Trials MeSH check words Humans

COVER SHEET Title Authors Contribution of author(s) Pharmacological interventions for people with borderline personality disorder Binks CA, Fenton M, McCarthy L, Lee T, Adams CE, Duggan C Clive Adams - sought funds, helped write the protocol, helped formulate searches, formatted the review and wrote the dicussion. Claire Binks - helped select studies, extract data and write the report. Conor Duggan - sought funds, helped write the protocol, helped formulate searches, read selection of abstracts, corrected final report.
36

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Mark Fenton - helped find funds, helped write the protocol, helped formulate searches, undertook and formatted the electronic searches, helped select studies, extracted data, and helped write the report. Tracy Lee - helped write the protocol, formulate searches and obtained papers. Lucy McCarthy - sought funds, helped write the protocol, helped formulate searches, read selection of abstracts, corrected final report. Issue protocol first published Review first published Date of most recent amendment Date of most recent SUBSTANTIVE amendment What’s New / 2006/1 17 November 2005 02 September 2005

The editorial base of the Cochrane Developmental, Psychosocial and Learning Problems Group would like to extend its sincere thanks to the Cochrane Schizophrenia Group, particularly Mark Fenton and Clive Adams, who did so much to make this review happen. The first version of this review is being published at Issue 1, 2006, whilst new authors are preparing an update to incorporate new data, which should be published shortly. Information not supplied by author

Date new studies sought but none found Date new studies found but not yet included/excluded Date new studies found and included/excluded Date authors’ conclusions section amended Contact address

Information not supplied by author

Information not supplied by author

Information not supplied by author

Ms Claire Binks C/o Jane Dennis SPS, CDPLPG, University of Bristol 8 Priory Bristol BS8 1TZ UK E-mail: clairebinks@hotmail.com,J.Dennis@bris.ac.uk Tel: 44 117 954 6782 Fax: 44 117 954 6623 10.1002/14651858.CD005653 CD005653 Cochrane Developmental, Psychosocial and Learning Problems Group HM-BEHAV
37

DOI Cochrane Library number Editorial group Editorial group code

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Analysis 01.01.
Review:

GRAPHS AND OTHER TABLES Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 01 Global state: 1. Mean score - by 6 months (GAS, higher scores = better)

Pharmacological interventions for people with borderline personality disorder

Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 01 Global state: 1. Mean score - by 6 months (GAS, higher scores = better) Study N Soloff 1989 Total (95% CI) 29 29 p=0.3 Antidepressant Mean(SD) 51.48 (13.80) N 28 28 Placebo Mean(SD) 48.16 (9.95) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI 3.32 [ -2.91, 9.55 ] 3.32 [ -2.91, 9.55 ]

Test for heterogeneity: not applicable Test for overall effect z=1.04

-10.0

-5.0

0

5.0

10.0

Favours control

Favours treatment

Analysis 01.02.

Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 02 Global state: 2. Mean symptom score - by 6 months (skewed data)

IMPS (higher scores = poor)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 99.96 97.79 SD 57.83 38.00 Notes F=0.001

SCL-90 (higher scores = poor)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.07 1.35 SD 0.73 0.87 Notes Nonsignificant difference.

Analysis 01.03. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data)

Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data)
Study Soloff 1989 Soloff 1989 Interventions Amitriptyline Placebo N 29 28 Mean 3.36 4.74 SD 4.55 5.60 Notes F=3.89, p</=0.1 (A>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

38

Analysis 01.04. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 04 Mental state: 1. Anger - not improved by 6 months (POMS not 20% over baseline score. high score=poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 04 Mental state: 1. Anger - not improved by 6 months (POMS not 20% over baseline score. high score=poor) Study Fluoxetine n/N Salzman 1995 Total (95% CI) 3/13 13 Placebo n/N 7/9 9 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.30 [ 0.10, 0.85 ] 0.30 [ 0.10, 0.85 ]

Total events: 3 (Fluoxetine), 7 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=2.26 p=0.02

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

39

Analysis 01.05. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 05 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 05 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) Study N Soloff 1989 Total (95% CI) 29 29 p=0.9 Antidepressant Mean(SD) 22.67 (9.12) N 28 28 Placebo Mean(SD) 23.06 (8.11) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -0.39 [ -4.87, 4.09 ] -0.39 [ -4.87, 4.09 ]

Test for heterogeneity: not applicable Test for overall effect z=0.17

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 01.06. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 06 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data)

anxiety - general (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.39 1.54 SD 0.96 1.07 Notes Nonsignificant difference.

anxiety - phobic (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 0.86 1.00 SD 0.87 0.97 Notes Nonsignificant difference.

interpersonal sensitivity (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.34 1.74 SD 0.95 1.06 Notes F=0.68.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

40

Analysis 01.07.
Review:

Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 07 Mental state: 3a. Depression - not improved - by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 07 Mental state: 3a. Depression - not improved - by 6 months Study Fluoxetine n/N 01 HAM-D total (not 80% over baseline score) Salzman 1995 Subtotal (95% CI) 0/13 13 2/9 9 100.0 100.0 0.14 [ 0.01, 2.66 ] 0.14 [ 0.01, 2.66 ] Placebo n/N Relative Risk (Fixed) 95% CI Weight (%) Relative Risk (Fixed) 95% CI

Total events: 0 (Fluoxetine), 2 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=1.30 p=0.2

02 POMS (not 20% over baseline score) Salzman 1995 Subtotal (95% CI) 3/13 13 8/9 9 100.0 100.0 0.26 [ 0.09, 0.72 ] 0.26 [ 0.09, 0.72 ]

Total events: 3 (Fluoxetine), 8 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=2.59 p=0.009

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 01.08. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data)

BDI
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 15.34 23.04 SD 10.71 14.88 Notes F=6.51, p</=0.05 (A>P)

HAM-D-17
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 11.36 13.86 SD 5.87 5.65 Notes Nonsignificant difference.

HAM-D-24
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 16.08 20.96 SD 9.04 9.05 Notes F=4.33, P</=0.05 (A>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

41

SCL-90 subscale
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.37 1.99 SD 0.96 1.11 Notes F=2.43

Analysis 01.09. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 09 Mental state: 4. Mood disturbance - not improved by 6 months (POMS, not 20% over baseline score)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 09 Mental state: 4. Mood disturbance - not improved by 6 months (POMS, not 20% over baseline score) Study Fluoxetine n/N Salzman 1995 Total (95% CI) 3/13 13 Placebo n/N 6/9 9 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.35 [ 0.12, 1.03 ] 0.35 [ 0.12, 1.03 ]

Total events: 3 (Fluoxetine), 6 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=1.90 p=0.06

0.1 0.2

0.5

1

2

5

10

Favours control

Favours treatment

Analysis 01.10. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)

paranoid ideation (SCL-90)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.03 1.44 SD 0.86 1.00 Notes F=2.75.

paranoid projection (IMPS)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 3.12 2.62 SD 5.16 3.78 Notes F=0.47.

perceptual distortions (IMPS)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 2.82 2.62 SD 3.68 3.71 Notes F=0.26.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

42

psychoticism (SCL-90)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 0.70 1.05 SD 0.60 0.73 Notes F=2.59.

Analysis 01.11. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 11 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed)

obsessive-compulsive score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.14 1.50 SD 0.84 1.06 Notes F=2.79

somatization score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 0.75 0.91 SD 0.64 0.78 Notes Nonsignificant difference.

unspecified additional items (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.06 1.53 SD 0.77 0.86 Notes Nonsignificant difference.

Analysis 01.12.
Review:

Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 12 Behaviour: 1. Attempting suicide within 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 12 Behaviour: 1. Attempting suicide within 6 months Study Mianserin n/N Montgomery 1983 Total (95% CI) 8/17 17 Placebo n/N 12/21 21 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.82 [ 0.44, 1.54 ] 0.82 [ 0.44, 1.54 ]

Total events: 8 (Mianserin), 12 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.61 p=0.5

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

43

Analysis 01.13. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 13 Behaviour: 2a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 13 Behaviour: 2a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) Study N Soloff 1989 Total (95% CI) 29 29 p=0.1 Amitriptyline Mean(SD) 5.98 (1.40) N 28 28 Placebo Mean(SD) 6.65 (2.00) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -0.67 [ -1.57, 0.23 ] -0.67 [ -1.57, 0.23 ]

Test for heterogeneity: not applicable Test for overall effect z=1.46

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 01.14. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 14 Behaviour: 2b. Hostility scores - by 6 months (higher scores = poor, skewed data)

hostility - general (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 1.12 1.39 SD 1.01 1.05 Notes F=0.84.

hostile belligerence (IMPS subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 8.54 9.33 SD 7.82 9.71 Notes F=1.64.

verbal hostility (BDHI subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 8.52 8.22 SD 5.95 3.11 Notes Nonsignificant difference.

Analysis 01.15. Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 15 Behaviour: 3a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)

Behaviour: 3a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Placebo N 29 28 Mean 9.38 8.28 SD 10.59 9.45 Notes F=0.05.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

44

Analysis 01.16.
Review:

Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 16 Behaviour: 3b. Impulsiveness - traits by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 16 Behaviour: 3b. Impulsiveness - traits by 6 months Study N AMITRIPTYLINE Mean(SD) N PLACEBO Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

01 BIS:self-control (higher scores = poor) Soloff 1989 Subtotal (95% CI) 29 29 p=0.6 100.50 (19.01) 28 28 103.06 (21.48) 100.0 100.0 -2.56 [ -13.10, 7.98 ] -2.56 [ -13.10, 7.98 ]

Test for heterogeneity: not applicable Test for overall effect z=0.48

02 STIC (higher scores = poor) Soloff 1989 Subtotal (95% CI) 29 29 p=0.4 10.22 (3.89) 28 28 9.20 (4.36) 100.0 100.0 1.02 [ -1.13, 3.17 ] 1.02 [ -1.13, 3.17 ]

Test for heterogeneity: not applicable Test for overall effect z=0.93

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

45

Analysis 01.17.
Review:

Comparison 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO, Outcome 17 Leaving the study early: Completed <2 weeks of treatment

Pharmacological interventions for people with borderline personality disorder

Comparison: 01 ANTIDEPRESSANT (TRICYCLIC / TETRACYCLIC) vs PLACEBO Outcome: 17 Leaving the study early: Completed <2 weeks of treatment Study Amitryptiline n/N Soloff 1989 Total (95% CI) 1/30 30 Placebo n/N 1/29 29 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.97 [ 0.06, 14.74 ] 0.97 [ 0.06, 14.74 ]

Total events: 1 (Amitryptiline), 1 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.02 p=1

0.01

0.1

1

10

100

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

46

Analysis 02.01. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 01 Global state: 1. Global functioning score - by 6 months (GAS, higher scores = better)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC Outcome: 01 Global state: 1. Global functioning score - by 6 months (GAS, higher scores = better) Study N Soloff 1989 Total (95% CI) 29 29 p=0.3 Amitriptyline Mean(SD) 51.48 (13.80) N 28 28 Haloperidol Mean(SD) 55.35 (12.36) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -3.87 [ -10.67, 2.93 ] -3.87 [ -10.67, 2.93 ]

Test for heterogeneity: not applicable Test for overall effect z=1.12

-10.0

-5.0

0

5.0

10.0

Favours control

Favours treatment

Analysis 02.02.

Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 02 Global state: 2. Mean symptom score - by 6 months (skewed data)

IMPS (higher scores = poor)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 99.96 76.86 SD 57.83 34.37 Notes F=3.08, p</=0.1

SCL-90 (higher scores = poor)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.07 1.02 SD 0.73 0.74 Notes Nonsignificant difference.

Analysis 02.03. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 03 Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data)

Global state: 3. Excitement score - by 6 months (IMPS subscale, higher scores = poor, skewed data)
Study Soloff 1989 Soloff 1989 Interventions Amitriptyline Haloperidol N 29 28 Mean 3.36 2.93 SD 4.55 2.87 Notes F=0.15.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

47

Analysis 02.04. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 04 Mental state: 1a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC Outcome: 04 Mental state: 1a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) Study N Soloff 1989 Total (95% CI) 29 29 p=0.3 Amitriptyline Mean(SD) 22.67 (9.12) N 28 28 Haloperidol Mean(SD) 20.35 (7.18) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI 2.32 [ -1.93, 6.57 ] 2.32 [ -1.93, 6.57 ]

Test for heterogeneity: not applicable Test for overall effect z=1.07

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 02.05. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 05 Mental state: 1b. Anxiety scores - by 6 months (higher scores = poor, skewed data)

anxiety - general (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.39 1.21 SD 0.96 0.98 Notes Nonsignificant difference

anxiety - phobic (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 0.86 0.83 SD 0.87 0.99 Notes Nonsignificant difference.

interpersonal sensitivity (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.34 1.21 SD 0.95 0.92 Notes F=5.91, p</=0.05. (H>A)

Analysis 02.06. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 06 Mental state: 2. Depression scores - by 6 months (higher scores = poor, skewed data)

BDI
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 15.34 16.22 SD 10.71 12.32 Notes F=0.36

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

48

HAM-D-17
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 11.36 12.42 SD 5.87 5.70 Notes Nonsignificant difference

HAM-D-24
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 16.08 16.12 SD 9.04 8.34 Notes F=0.02.

SCL-90 subscale
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.37 1.50 SD 0.96 1.07 Notes F=0.94

Analysis 02.07. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 07 Mental state: 3. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)

paranoid ideation (SCL-90)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.03 0.75 SD 0.86 0.73 Notes F=4.37, p</=0.05 (H>A)

paranoid projection (IMPS)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 3.12 0.57 SD 5.16 2.10 Notes F=5.37, p</=0.05 (H>A)

perceptual distortions (IMPS)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 2.82 0.93 SD 3.68 2.32 Notes F=5.58, p</=0.05 (H>A)

psychoticism (SCL-90)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 0.70 0.66 SD 0.60 0.55 Notes F=0.70

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

49

schizotypal symptoms (SSI)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 5.33 4.15 SD 3.30 2.48 Notes F=4.05, p</=0.05 (H>A)

Analysis 02.08. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 08 Mental state: 4. Other symptoms - by 6 months (high scores = poor, data skewed)

obsessive-compulsive score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.14 1.12 SD 0.84 0.80 Notes F=0.89.

somatization score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 0.75 0.74 SD 0.64 0.68 Notes Nonsignificant difference.

unspecified additional items (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.06 1.21 SD 0.77 0.84 Notes Nonsignificant difference.

Analysis 02.09. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 09 Behaviour: 1a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC Outcome: 09 Behaviour: 1a. Hostility - indirect hostility score - by 6 months (BDHI subscale, higher scores = poor) Study N Soloff 1989 Total (95% CI) 29 29 p=0.8 Amitriptyline Mean(SD) 5.98 (1.40) N 28 28 Haloperidol Mean(SD) 6.07 (1.90) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -0.09 [ -0.96, 0.78 ] -0.09 [ -0.96, 0.78 ]

Test for heterogeneity: not applicable Test for overall effect z=0.20

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 02.10. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 10 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 50

hostility - general (SCL-90 subscale)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

51

hostility - general (SCL-90 subscale) (Continued )

Study

Intervention

N

Mean

SD

Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

52

hostility - general (SCL-90 subscale) (Continued )

Study

Intervention

N

Mean

SD

Notes

Study

Intervention

N

Mean

SD

Notes
53

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

hostility - general (SCL-90 subscale) (Continued )

Study

Intervention

N

Mean

SD

Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

54

hostility - general (SCL-90 subscale) (Continued )
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 1.12 0.78 SD 1.01 0.82 Notes F=4.74, p</=0.05 (H>A)

hostile belligerence (IMPS subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 8.54 5.02 SD 7.82 5.55 Notes F=3.98, p</=0.1 (H>A)

verbal hostility (BDHI subscale)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 8.52 8.34 SD 5.95 3.07 Notes Nonsignificant difference

Analysis 02.11. Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 11 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)

Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)
Study Soloff 1989 Soloff 1989 Intervention Amitriptyline Haloperidol N 29 28 Mean 9.38 4.46 SD 10.59 3.86 Notes F=5.31, p</=0.05 (H>A)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

55

Analysis 02.12.
Review:

Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 12 Behaviour: 2b. Impulsiveness - traits by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC Outcome: 12 Behaviour: 2b. Impulsiveness - traits by 6 months Study N Amitriptyline Mean(SD) N Haloperidol Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

01 BIS:self-control (higher scores = poor) Soloff 1989 Subtotal (95% CI) 29 29 p=0.4 100.50 (19.01) 28 28 104.02 (15.71) 100.0 100.0 -3.52 [ -12.56, 5.52 ] -3.52 [ -12.56, 5.52 ]

Test for heterogeneity: not applicable Test for overall effect z=0.76

02 STIC (higher scores = poor) Soloff 1989 Subtotal (95% CI) 29 29 p=1 10.22 (3.89) 28 28 10.23 (4.08) 100.0 100.0 -0.01 [ -2.08, 2.06 ] -0.01 [ -2.08, 2.06 ]

Test for heterogeneity: not applicable Test for overall effect z=0.01

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 02.13.
Review:

Comparison 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC, Outcome 13 Leaving the study early: Completed <2 weeks of treatment

Pharmacological interventions for people with borderline personality disorder

Comparison: 02 ANTIDEPRESSANT (TRICYCLIC) vs ANTIPSYCHOTIC Outcome: 13 Leaving the study early: Completed <2 weeks of treatment Study Amitryptiline n/N Soloff 1989 Total (95% CI) 1/30 30 Haloperidol n/N 3/31 31 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.34 [ 0.04, 3.13 ] 0.34 [ 0.04, 3.13 ]

Total events: 1 (Amitryptiline), 3 (Haloperidol) Test for heterogeneity: not applicable Test for overall effect z=0.95 p=0.3

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 03.01. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 56

Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 18.30 20.08 SD 11.24 12.44 Notes Nonsignificant difference

Analysis 03.02. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) Study N Soloff 1993 Total (95% CI) 34 34 p=0.6 MAOI Mean(SD) 60.10 (10.70) N 30 30 Placebo Mean(SD) 58.43 (12.80) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI 1.67 [ -4.15, 7.49 ] 1.67 [ -4.15, 7.49 ]

Test for heterogeneity: not applicable Test for overall effect z=0.56

-10.0

-5.0

0

5.0

10.0

Favours control

Favours treatment

Analysis 03.03. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) Study N Soloff 1993 Total (95% CI) 34 34 p=1 MAOI Mean(SD) 72.56 (17.66) N 28 28 Placebo Mean(SD) 72.46 (23.26) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI 0.10 [ -10.36, 10.56 ] 0.10 [ -10.36, 10.56 ]

Test for heterogeneity: not applicable Test for overall effect z=0.02

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 03.04. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 57

Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.99 1.16 SD 0.64 0.84 Notes Nonsignificant difference

Analysis 03.05. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) Study MAOI n/N Soloff 1993 Total (95% CI) 27/38 38 Placebo n/N 26/34 34 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.93 [ 0.71, 1.22 ] 0.93 [ 0.71, 1.22 ]

Total events: 27 (MAOI), 26 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.52 p=0.6

0.1 0.2

0.5

1

2

5

10

Favours control

Favours treatment

Analysis 03.06. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) Study N Soloff 1993 Total (95% CI) 34 34 p=0.4 MAOI Mean(SD) 16.85 (6.48) N 28 28 Placebo Mean(SD) 18.36 (8.63) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -1.51 [ -5.38, 2.36 ] -1.51 [ -5.38, 2.36 ]

Test for heterogeneity: not applicable Test for overall effect z=0.77

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 03.07. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90 subscale higher scores = poor, skewed data)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 58

Mental state: 2b. Anxiety general - by 6 months (SCL-90 subscale higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenlzine Placebo N 34 28 Mean 1.05 1.18 1.00 SD 0.78 Notes

Analysis 03.08. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 08 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data)

BDI
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 15.69 19.54 SD 9.46 13.04 Notes Nonsignificant difference

HAM-D-17
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 10.71 11.18 SD 5.32 6.43 Notes Nonsignificant difference

HAM-D-24
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 13.50 15.04 SD 7.71 9.02 Notes Nonsignificant difference

SCL-90 subscale
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 1.22 1.58 SD 0.83 1.08 Notes Nonsignificant difference

Analysis 03.09. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 09 Mental state: 4. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)

paranoid ideation (SCL-90)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.92 1.18 0.98 SD 0.87 Notes

paranoid projection (IMPS)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 1.76 2.18 4.53
59

SD 3.56

Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

paranoid projection (IMPS)
Study

(Continued )
N Mean SD Notes

Intervention

psychoticism (SCL-90)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.60 1.00 SD 0.63 0.88 Notes Nonsignificant difference

schizotypal symptoms (SSI)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 6.71 7.46 SD 5.96 7.06 Notes Nonsignificant difference

Analysis 03.10. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 10 Mental state: 5. Other symptoms - by 6 months (high scores = poor, data skewed)

hypersomnia score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.57 0.71 0.85 SD 0.71 Notes

leaden paralysis score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.90 0.66 0.72 SD 0.86 Notes

obsessive-compulsive score (SCL-90 subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 1.02 1.19 SD 0.74 0.86 Notes Nonsignificant difference

reactivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 2.37 2.46 1.39
60

SD 1.52

Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

rejection sensitivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.90 0.73 0.69 SD 0.71 Notes

Analysis 03.11. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 11 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder by 6 months (BDHI total, higher scores = poor) Placebo N 28 28 p=0.009 Mean(SD) 47.88 (12.41) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -9.19 [ -16.12, -2.26 ] -9.19 [ -16.12, -2.26 ] Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 11 Behaviour: 1a. Hostility Study N Soloff 1993 Total (95% CI) 34 34 MAOI Mean(SD) 38.69 (15.44)

Test for heterogeneity: not applicable Test for overall effect z=2.60

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 03.12. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 12 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data)

hostility - general (SCL-90 subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.73 1.04 0.97 SD 0.85 Notes

hostile belligerence (IMPS subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 6.56 4.43 SD 6.39 5.72 Notes Nonsignificant difference

Analysis 03.13. Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 13 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)

Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 4.76 3.50 SD 4.82 4.46 Notes Nonsignificant difference

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

61

Analysis 03.14.
Review:

Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 14 Behaviour: 2b. Impulsiveness - traits by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO Outcome: 14 Behaviour: 2b. Impulsiveness - traits by 6 months Study N MAOI Mean(SD) N Placebo Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

01 BIS:self-control (higher scores = poor) Soloff 1993 Subtotal (95% CI) 34 34 p=1 237.66 (34.73) 28 28 237.74 (42.02) 100.0 100.0 -0.08 [ -19.54, 19.38 ] -0.08 [ -19.54, 19.38 ]

Test for heterogeneity: not applicable Test for overall effect z=0.01

02 STIC (higher scores = poor) Soloff 1993 Subtotal (95% CI) 34 34 p=0.3 30.12 (7.95) 28 28 28.04 (7.70) 100.0 100.0 2.08 [ -1.83, 5.99 ] 2.08 [ -1.83, 5.99 ]

Test for heterogeneity: not applicable Test for overall effect z=1.04

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 03.15.

Comparison 03 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs PLACEBO, Outcome 15 Appitite and weight (higher scores = poor, skewed data)

increased appetite score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.50 0.36 0.49 SD 0.58 Notes

weight gain score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Placebo N 34 28 Mean 0.50 0.41 0.76 SD 0.83 Notes

Analysis 04.01. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)

Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 18.30 24.03 SD 11.24 13.24 Notes Nonsignificant difference

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

62

Analysis 04.02. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) Study N Soloff 1993 Total (95% CI) 34 34 p=0.04 MAOI Mean(SD) 60.10 (10.70) N 30 30 Haloperidol Mean(SD) 54.95 (9.15) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI 5.15 [ 0.29, 10.01 ] 5.15 [ 0.29, 10.01 ]

Test for heterogeneity: not applicable Test for overall effect z=2.08

-10.0

-5.0

0

5.0

10.0

Favours control

Favours treatment

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

63

Analysis 04.03. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) Study N Soloff 1993 Total (95% CI) 34 34 p=0.4 MAOI Mean(SD) 72.56 (17.66) N 30 30 Haloperidol Mean(SD) 76.13 (15.34) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -3.57 [ -11.66, 4.52 ] -3.57 [ -11.66, 4.52 ]

Test for heterogeneity: not applicable Test for overall effect z=0.87

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 04.04. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)

Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.99 1.34 SD 0.64 0.67 Notes F=4.01, p<0.05 (P>H)

Analysis 04.05. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) Study MAOI n/N Soloff 1993 Total (95% CI) 27/38 38 Haloperidol n/N 31/36 36 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.83 [ 0.65, 1.05 ] 0.83 [ 0.65, 1.05 ]

Total events: 27 (MAOI), 31 (Haloperidol) Test for heterogeneity: not applicable Test for overall effect z=1.56 p=0.1

0.1 0.2

0.5

1

2

5

10

Favours control

Favours treatment

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

64

Analysis 04.06. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) Study N Soloff 1993 Total (95% CI) 34 34 p=0.04 MAOI Mean(SD) 16.85 (6.48) N 30 30 Haloperidol Mean(SD) 20.73 (8.13) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -3.88 [ -7.51, -0.25 ] -3.88 [ -7.51, -0.25 ]

Test for heterogeneity: not applicable Test for overall effect z=2.09

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 04.07. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 07 Mental state: 2b. Anxiety general - by 6 months (SCL-90, higher scores = poor, skewed data)

Mental state: 2b. Anxiety general - by 6 months (SCL-90, higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 1.05 1.57 SD 0.78 0.78 Notes F=6.17, p</=0.05 (P>H)

Analysis 04.08. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 08 Mental state: 3a. Depression scores - by 6 months (HAM-D-17, high score= poor))
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 08 Mental state: 3a. Depression scores - by 6 months (HAM-D-17, high score=poor)) Study N Soloff 1993 Total (95% CI) 34 34 p<0.00001 MAOI Mean(SD) 10.71 (5.32) N 30 30 Haloperidol Mean(SD) 18.57 (5.46) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -7.86 [ -10.51, -5.21 ] -7.86 [ -10.51, -5.21 ]

Test for heterogeneity: not applicable Test for overall effect z=5.82

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

65

Analysis 04.09. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 09 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data)

BDI
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 15.69 23.50 SD 9.46 13.16 Notes F=3.29, p<0.1 (P>H)

HAM-D-24
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 13.50 18.60 SD 7.71 7.44 Notes F=6.41, p<0.05 (P>H)

SCL-90 subscale
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 1.22 2.01 SD 0.83 0.95 Notes F=12.54, p<0.001 (P>H)

Analysis 04.10. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 10 Mental state: 4. Dysphoria - hysteroid dysphoria - by 6 months (HDQ, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 10 Mental state: 4. Dysphoria - hysteroid dysphoria - by 6 months (HDQ, higher scores = poor) Study N Soloff 1993 Total (95% CI) 34 34 p=0.3 MAOI Mean(SD) 11.61 (3.50) N 30 30 Haloperidol Mean(SD) 12.43 (3.40) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -0.82 [ -2.51, 0.87 ] -0.82 [ -2.51, 0.87 ]

Test for heterogeneity: not applicable Test for overall effect z=0.95

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 04.11. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 11 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)
Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd 66

paranoid ideation (SCL-90)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.92 1.06 0.96 SD 0.87 Notes

paranoid projection (IMPS)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 1.76 2.53 4.63 SD 3.56 Notes

psychoticism (SCL-90)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.60 0.98 SD 0.63 0.73 Notes F=3.04, p<0.1 (P>H)

schizotypal symptoms (SSI)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 6.71 8.90 SD 5.96 5.24 Notes F=3.81, p<0.05, (P>H)

Analysis 04.12. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 12 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed)

hypersomnia score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.57 0.87 0.84 SD 0.71 Notes

leaden paralysis score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.90 1.05 1.05 SD 0.86 Notes

obsessive-compulsive score (SCL-90 subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 1.02 1.52 SD 0.74 0.93 Notes F=5.29, p<0.05 (P>H)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

67

reactivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 2.37 2.30 1.43 SD 1.52 Notes

rejection sensitivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.90 0.57 0.70 SD 0.71 Notes

Analysis 04.13. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 13 Behaviour: 1a. Hostility - by 6 months (BDHI total, higher scores = poor)
Review: Pharmacological interventions for people with borderline personality disorder by 6 months (BDHI total, higher scores = poor) Haloperidol N 30 30 p=0.2 Mean(SD) 43.53 (13.53) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -4.84 [ -11.94, 2.26 ] -4.84 [ -11.94, 2.26 ] Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 13 Behaviour: 1a. Hostility Study N Soloff 1993 Total (95% CI) 34 34 MAOI Mean(SD) 38.69 (15.44)

Test for heterogeneity: not applicable Test for overall effect z=1.34

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 04.14. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 14 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data)

hostility - general (SCL-90 subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.73 0.79 0.66 SD 0.85 Notes

hostile belligerence (IMPS subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 6.56 4.17 SD 6.39 3.91 Notes F=3.83, p<0.05 (H>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

68

Analysis 04.15. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 15 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)

Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 4.76 3.07 SD 4.82 4.52 Notes F=7.25, p<0.05 (H>P)

Analysis 04.16. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 16 Behaviour: 2b. Impulsiveness - traits by 6 months
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC Outcome: 16 Behaviour: 2b. Impulsiveness - traits by 6 months Study N MAOI Mean(SD) N Haloperidol Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

01 BIS:self-control (higher scores = poor) Soloff 1993 Subtotal (95% CI) 34 34 p=0.7 237.66 (34.73) 30 30 240.95 (37.60) 100.0 100.0 -3.29 [ -21.10, 14.52 ] -3.29 [ -21.10, 14.52 ]

Test for heterogeneity: not applicable Test for overall effect z=0.36

02 STIC (higher scores = poor) Soloff 1993 Subtotal (95% CI) 34 34 p=0.8 30.12 (7.95) 30 30 29.49 (8.26) 100.0 100.0 0.63 [ -3.35, 4.61 ] 0.63 [ -3.35, 4.61 ]

Test for heterogeneity: not applicable Test for overall effect z=0.31

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 04.17. Comparison 04 ANTIDEPRESSENT (MONOAMINE OXIDADE INHIBITER) vs ANTIPSYCHOTIC, Outcome 17 Appetite and weight (higher scores = poor, skewed data)

increased appetite score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.50 0.45 0.69 SD 0.58 Notes

weight gain score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Phenelzine Haloperidol N 34 30 Mean 0.50 0.28 0.66
69

SD 0.83

Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

weight gain score (ADI subscale)
Study

(Continued )
N Mean SD Notes

Intervention

Analysis 05.01.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 01 Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)

Global state: 1. Borderline symptomatology - by 6 months (BSI, higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 24.03 20.08 SD 13.24 12.44 Notes No group differences.

Analysis 05.02.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better)

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 02 Global state: 2. Global functioning score - by 6 months (GAS, higher scores = better) Study N Soloff 1989 Soloff 1993 Total (95% CI) 28 30 58 p=0.4 Haloperidol Mean(SD) 55.35 (12.36) 54.95 (9.15) N 28 28 56 Placebo Mean(SD) 48.16 (9.95) 58.43 (12.80) Weighted Mean Difference (Fixed) 95% CI Weight (%) 49.0 51.0 100.0 Weighted Mean Difference (Fixed) 95% CI 7.19 [ 1.31, 13.07 ] -3.48 [ -9.24, 2.28 ] 1.75 [ -2.37, 5.86 ]

Test for heterogeneity chi-square=6.46 df=1 p=0.01 I² =84.5% Test for overall effect z=0.83

-10.0

-5.0

0

5.0

10.0

Favours control

Favours treatment

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

70

Analysis 05.03.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor)

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 03 Global state: 3a. Mean symptom score - by 6 months (IMPS, higher score = poor) Study N Soloff 1989 Soloff 1993 Total (95% CI) 28 30 58 p=0.7 Haloperidol Mean(SD) 76.86 (34.37) 76.13 (15.34) N 28 28 56 Placebo Mean(SD) 97.79 (38.00) 72.46 (23.26) Weighted Mean Difference (Fixed) 95% CI Weight (%) 22.5 77.5 100.0 Weighted Mean Difference (Fixed) 95% CI -20.93 [ -39.91, -1.95 ] 3.67 [ -6.55, 13.89 ] -1.86 [ -10.85, 7.14 ]

Test for heterogeneity chi-square=5.00 df=1 p=0.03 I² =80.0% Test for overall effect z=0.40

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 05.04.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 04 Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)

Global state: 3b. Mean symptom score - by 6 months (SCL-90, higher score = poor, skewed data)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 1.02 1.35 1.34 1.16 SD 0.74 0.87 0.67 0.84 Nonsignificant difference. Notes Nonsignificant difference.

Analysis 05.05. Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement)
Review: Pharmacological interventions for people with borderline personality disorder Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 05 Mental state: 1. Not ’responders’ - by 6 months (not </=7 on HAM-D-17 + min. 50% improvement) Study Haloperidol n/N Soloff 1993 Total (95% CI) 9/36 36 Placebo n/N 11/34 34 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.77 [ 0.37, 1.63 ] 0.77 [ 0.37, 1.63 ]

Total events: 9 (Haloperidol), 11 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.68 p=0.5

0.1 0.2

0.5

1

2

5

10

Favours control

Favours treatment

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

71

Analysis 05.06.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor)

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 06 Mental state: 2a. Anxiety score - intropunitiveness by 6 months (IMPS, higher scores = poor) Study N Soloff 1989 Soloff 1993 Total (95% CI) 28 30 58 p=0.8 Haloperidol Mean(SD) 20.35 (7.18) 20.73 (8.13) N 28 28 56 Placebo Mean(SD) 23.06 (8.11) 18.36 (8.63) Weighted Mean Difference (Fixed) 95% CI Weight (%) 53.7 46.3 100.0 Weighted Mean Difference (Fixed) 95% CI -2.71 [ -6.72, 1.30 ] 2.37 [ -1.95, 6.69 ] -0.36 [ -3.30, 2.58 ]

Test for heterogeneity chi-square=2.85 df=1 p=0.09 I² =64.9% Test for overall effect z=0.24

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 05.07.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 07 Mental state: 2b. Anxiety scores - by 6 months (higher scores = poor, skewed data)

anxiety - general (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 1.21 1.54 1.57 1.18 SD 0.98 1.07 0.78 1.00 F=3.31, <0.1 (P>H) Notes Nonsignificant difference

anxiety - phobic (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 0.83 1.00 SD 0.99 0.97 Notes Nonsignificant difference.

interpersonal sensitivity (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 1.21 1.74 SD 0.92 1.06 Notes F=9.62, p</=0.01 (H>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

72

Analysis 05.08.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 08 Mental state: 3a. Depression scores - by 6 months (HAM-D-17 higher scores = poor)

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 08 Mental state: 3a. Depression scores - by 6 months (HAM-D-17 higher scores = poor) Study N Soloff 1989 Total (95% CI) 28 28 p=0.3 Haloperidol Mean(SD) 12.42 (5.70) N 28 28 Placebo Mean(SD) 13.86 (5.65) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -1.44 [ -4.41, 1.53 ] -1.44 [ -4.41, 1.53 ]

Test for heterogeneity: not applicable Test for overall effect z=0.95

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 05.09.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 09 Mental state: 3b. Depression scores - by 6 months (higher scores = poor, skewed data)

BDI
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 16.22 23.04 23.50 19.54 SD 12.32 14.88 13.16 13.04 Nonsignificant difference. Notes F=8.42, p</=0.01 (H>P)

HAM-D-17
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 18.57 11.18 SD 5.46 6.43 Notes Nonsignificant difference.

HAM-D-24
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 16.12 20.96 18.60 15.04 SD 8.34 9.05 7.44 9.02 F=2.70, p =0.1 (P>H) Notes F=5.06, p</=0.05 (H>P)

SCL-90 subscale
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 1.50 1.99 SD 1.07 1.11
73

Notes F=6.18, p</=0.05 (H>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

SCL-90 subscale (Continued )
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 2.01 1.58 SD 0.95 1.08 Notes F=4.20, p<0.05 (P>H)

Analysis 05.10.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 10 Mental state: 5. Schizotypal symptoms - by 6 months (data skewed, higher scores = poor)

paranoid ideation (SCL-90)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 0.75 1.44 1.06 1.18 SD 0.73 1.00 0.96 0.98 Notes F=13.60, p</=0.001 (H>P)

paranoid projection (IMPS)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 0.57 2.62 2.53 2.18 SD 2.10 3.78 4.63 4.53 Notes F=4.03, p</=0.05 (H>P)

perceptual distortions (IMPS)
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 0.93 2.62 SD 2.32 3.71 Notes F=3.36, p</=0.1 (H>P)

psychoticism (SCL-90)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 0.66 1.05 0.98 1.00 SD 0.55 0.73 0.73 0.88
74

Notes F=6.03, p<0.05 (H>P)

Nonsignificant difference.

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

schizotypal symptoms (SSI)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 4.15 5.46 8.90 7.46 SD 2.48 2.38 5.24 7.06 Nonsignificant dfifference. Notes F=4.69, p</=0.05 (H>P)

Analysis 05.11.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 11 Mental state: 6. Other symptoms - by 6 months (high scores = poor, data skewed)

hypersomnia score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 0.87 0.71 0.85 SD 0.84 Notes

leaden paralysis score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 1.05 0.66 0.72 SD 1.05 Notes

obsessive-compulsive score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 1.12 1.50 1.52 1.19 SD 0.80 1.06 0.93 0.86 F=3.14, p<0.1 (P>H) Notes F=6.80, p</=0.05 (H>P)

reactivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 2.30 2.46 1.39 SD 1.43 Notes

rejection sensitivity score (ADI subscale)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo N 30 28 Mean 0.57 0.73 0.69 SD 0.70 Notes

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

75

somatization score (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 0.74 0.91 SD 0.68 0.78 Notes Nonsignificant difference.

unspecified additional items (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Intervention Haloperidol Placebo N 28 28 Mean 1.21 1.53 SD 0.84 0.86 Notes Nonsignificant difference.

Analysis 05.12.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 12 Behaviour: 1a. Hostility - by 6 months (higher scores = poor)

Pharmacological interventions for people with borderline personality disorder by 6 months (higher scores = poor) Placebo N Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 12 Behaviour: 1a. Hostility Study N 01 BDHI total Soloff 1993 Subtotal (95% CI) 30 30 p=0.2 43.53 (13.53) 28 28 47.88 (12.41) 100.0 100.0 -4.35 [ -11.03, 2.33 ] -4.35 [ -11.03, 2.33 ] Haloperidol Mean(SD)

Test for heterogeneity: not applicable Test for overall effect z=1.28

02 indirect hostility (BDHI subscore) Soloff 1989 Subtotal (95% CI) 28 28 p=0.3 6.07 (1.90) 28 28 6.65 (2.00) 100.0 100.0 -0.58 [ -1.60, 0.44 ] -0.58 [ -1.60, 0.44 ]

Test for heterogeneity: not applicable Test for overall effect z=1.11

03 verbal hostility (BDHI subscore) Soloff 1989 Subtotal (95% CI) 28 28 p=0.9 8.34 (3.07) 28 28 8.22 (3.11) 100.0 100.0 0.12 [ -1.50, 1.74 ] 0.12 [ -1.50, 1.74 ]

Test for heterogeneity: not applicable Test for overall effect z=0.15

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 05.13.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 13 Behaviour: 1b. Hostility scores - by 6 months (higher scores = poor, skewed data)
76

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

hostility - general (SCL-90 subscale)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 0.78 1.39 0.79 1.04 SD 0.82 1.05 0.66 0.97 Notes F=10.36, p</=0.01 (H>P)

hostile belligerence (IMPS subscale)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 5.02 9.33 4.17 4.43 SD 5.55 9.71 3.91 5.72 Nonsignificant difference. Notes F=7.31, p</=0.01 (H>P)

Analysis 05.14.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 14 Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)

Behaviour: 2a. Impulsiveness - behaviour by 6 months (WSIAP, higher scores = poor, data skewed)
Study Soloff 1989 Soloff 1989 Soloff 1993 Soloff 1993 Intervention Haloperidol Placebo Haloperidol Placebo N 28 28 30 28 Mean 4.46 8.28 3.07 3.50 SD 3.86 9.45 4.52 4.46 Nonsignificant difference. Notes F=6.32, p</=0.05 (H>P)

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

77

Analysis 05.15.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 15 Behaviour: 2b. Impulsiveness - traits by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 15 Behaviour: 2b. Impulsiveness - traits by 6 months Study N Haloperidol Mean(SD) N Placebo Mean(SD) Weighted Mean Difference (Fixed) 95% CI Weight (%) Weighted Mean Difference (Fixed) 95% CI

01 BIS:self-control (higher scores = poor) Soloff 1989 Soloff 1993 Subtotal (95% CI) 28 30 58 p=0.8 104.02 (15.71) 240.95 (37.60) 28 28 56 103.06 (21.48) 237.74 (42.02) 81.3 18.7 100.0 0.96 [ -8.90, 10.82 ] 3.21 [ -17.36, 23.78 ] 1.38 [ -7.51, 10.27 ]

Test for heterogeneity chi-square=0.04 df=1 p=0.85 I² =0.0% Test for overall effect z=0.30

02 STIC (higher scores = poor) Soloff 1989 Soloff 1993 Subtotal (95% CI) 28 30 58 p=0.3 10.23 (4.08) 29.49 (8.26) 28 28 56 9.20 (4.36) 28.04 (7.70) 77.5 22.5 100.0 1.03 [ -1.18, 3.24 ] 1.45 [ -2.66, 5.56 ] 1.12 [ -0.82, 3.07 ]

Test for heterogeneity chi-square=0.03 df=1 p=0.86 I² =0.0% Test for overall effect z=1.13

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 05.16.
Review:

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 16 Leaving the study early

Pharmacological interventions for people with borderline personality disorder

Comparison: 05 ANTIPSYCHOTICS vs PLACEBO Outcome: 16 Leaving the study early Study Antipsychotics n/N Goldberg 1986 Soloff 1989 Total (95% CI) 13/24 3/31 55 Placebo n/N 11/26 1/29 55 Relative Risk (Fixed) 95% CI Weight (%) 91.1 8.9 100.0 Relative Risk (Fixed) 95% CI 1.28 [ 0.72, 2.29 ] 2.81 [ 0.31, 25.48 ] 1.42 [ 0.80, 2.51 ]

Total events: 16 (Antipsychotics), 12 (Placebo) Test for heterogeneity chi-square=0.49 df=1 p=0.49 I² =0.0% Test for overall effect z=1.19 p=0.2

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 05.17.

Comparison 05 ANTIPSYCHOTICS vs PLACEBO, Outcome 17 Weight gain score (ADI subscale, higher scores = poor, skewed data)
78

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Weight gain score (ADI subscale, higher scores = poor, skewed data)
Study Soloff 1993 Soloff 1993 Intervention Haloperidol placebo N 30 28 Mean 0.28 0.41 0.76 SD 0.66 Notes

Analysis 06.01.
Review:

Comparison 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE, Outcome 01 Leaving the study early - before 3 weeks of treatment

Pharmacological interventions for people with borderline personality disorder

Comparison: 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE Outcome: 01 Leaving the study early - before 3 weeks of treatment Study Loxapine n/N Leone 1982 Total (95% CI) 6/40 40 5/40 40 Chlorpromazine n/N Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 1.20 [ 0.40, 3.62 ] 1.20 [ 0.40, 3.62 ]

Total events: 6 (Loxapine), 5 (Chlorpromazine) Test for heterogeneity: not applicable Test for overall effect z=0.32 p=0.7

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 06.02.
Review:

Comparison 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE, Outcome 02 Adverse effects: One or more adverse effects reported

Pharmacological interventions for people with borderline personality disorder

Comparison: 06 ANTIPSYCHOTIC COMPARISONS: 1. LOXAPINE vs CHLORPROMAZINE Outcome: 02 Adverse effects: One or more adverse effects reported Study Loxapine n/N Leone 1982 Total (95% CI) 14/40 40 11/40 40 Chlorpromazine n/N Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 1.27 [ 0.66, 2.45 ] 1.27 [ 0.66, 2.45 ]

Total events: 14 (Loxapine), 11 (Chlorpromazine) Test for heterogeneity: not applicable Test for overall effect z=0.72 p=0.5

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

79

Analysis 07.01.
Review:

Comparison 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL, Outcome 01 Leaving the study early - by 6 weeks

Pharmacological interventions for people with borderline personality disorder

Comparison: 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL Outcome: 01 Leaving the study early - by 6 weeks Study Thiothixine n/N Serban 1984 Total (95% CI) 2/26 26 Haloperidol n/N 4/26 26 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.50 [ 0.10, 2.50 ] 0.50 [ 0.10, 2.50 ]

Total events: 2 (Thiothixine), 4 (Haloperidol) Test for heterogeneity: not applicable Test for overall effect z=0.84 p=0.4

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 07.02.
Review:

Comparison 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL, Outcome 02 Adverse effects: Adverse reactions to medication by 6 months

Pharmacological interventions for people with borderline personality disorder

Comparison: 07 ANTIPSYCHOTIC COMPARISONS: 2. THIOTHIXENE vs HALOPERIDOL Outcome: 02 Adverse effects: Adverse reactions to medication by 6 months Study Thiothixine n/N Serban 1984 Total (95% CI) 21/26 26 Haloperidol n/N 22/26 26 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.95 [ 0.74, 1.22 ] 0.95 [ 0.74, 1.22 ]

Total events: 21 (Thiothixine), 22 (Haloperidol) Test for heterogeneity: not applicable Test for overall effect z=0.37 p=0.7

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

80

Analysis 08.01.
Review:

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 01 Leaving the study early

Pharmacological interventions for people with borderline personality disorder

Comparison: 08 MOOD STABILISERS vs PLACEBO Outcome: 01 Leaving the study early Study Mood stabiliser n/N 01 carbemazepine De la Fuente 1994 Subtotal (95% CI) 2/10 10 0/10 10 100.0 100.0 5.00 [ 0.27, 92.62 ] 5.00 [ 0.27, 92.62 ] Placebo n/N Relative Risk (Fixed) 95% CI Weight (%) Relative Risk (Fixed) 95% CI

Total events: 2 (Mood stabiliser), 0 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=1.08 02 divalporex sodium Hollander 2001 Subtotal (95% CI) 6/12 12 4/4 4 100.0 100.0 0.50 [ 0.28, 0.88 ] 0.50 [ 0.28, 0.88 ] p=0.3

Total events: 6 (Mood stabiliser), 4 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=2.40 p=0.02

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 08.02.
Review:

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 02 Mental state: 1. Not responsive - by 6 months (> 2 on CGI-I)

Pharmacological interventions for people with borderline personality disorder

Comparison: 08 MOOD STABILISERS vs PLACEBO Outcome: 02 Mental state: 1. Not responsive - by 6 months (> 2 on CGI-I) Study Divalproex sodium n/N Hollander 2001 Total (95% CI) 7/12 12 Placebo n/N 4/4 4 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.58 [ 0.36, 0.94 ] 0.58 [ 0.36, 0.94 ]

Total events: 7 (Divalproex sodium), 4 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=2.21 p=0.03

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 08.03.

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 03 Mental state: 2. Depression scores - by 6 months (BDI, higher scores = poor, skewed data)
81

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

Mental state: 2. Depression scores - by 6 months (BDI, higher scores = poor, skewed data)
Study Hollander 2001 Hollander 2001 Intervention Divalproex Sodium Placebo N 12 4 Mean 8.2 18.0 SD 9.1 7.0 Notes F=2.69, df=1,12, p=.135.

Analysis 08.04.
Review:

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 04 Behaviour: 1. Acting out by 6 months (AOS behavioural dyscontrol)

Pharmacological interventions for people with borderline personality disorder

Comparison: 08 MOOD STABILISERS vs PLACEBO Outcome: 04 Behaviour: 1. Acting out by 6 months (AOS behavioural dyscontrol) Study Carbamazepine n/N De la Fuente 1994 Total (95% CI) 8/10 10 Placebo n/N 6/10 10 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 1.33 [ 0.74, 2.41 ] 1.33 [ 0.74, 2.41 ]

Total events: 8 (Carbamazepine), 6 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.95 p=0.3

0.1 0.2

0.5

1

2

5

10

Favours treatment

Favours control

Analysis 08.05.
Review:

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 05 Behaviour: 2a. Aggression score - by 6 months (AQ, higher scores = poor)

Pharmacological interventions for people with borderline personality disorder

Comparison: 08 MOOD STABILISERS vs PLACEBO Outcome: 05 Behaviour: 2a. Aggression score - by 6 months (AQ, higher scores = poor) Study Divalproex sodium N Hollander 2001 Total (95% CI) 12 12 p=0.2 Mean(SD) 76.10 (17.20) N 4 4 Placebo Mean(SD) 85.30 (12.00) Weighted Mean Difference (Fixed) 95% CI Weight (%) 100.0 100.0 Weighted Mean Difference (Fixed) 95% CI -9.20 [ -24.46, 6.06 ] -9.20 [ -24.46, 6.06 ]

Test for heterogeneity: not applicable Test for overall effect z=1.18

-10.0

-5.0

0

5.0

10.0

Favours treatment

Favours control

Analysis 08.06.

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 06 Behaviour: 2b. Aggression scores - by 6 months (higher scores = poor, skewed data).
82

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

OAS-M aggression subscale
Study Hollander 2001 Hollander 2001 Intervention Divalproex Sodium Placebo N 12 4 Mean 1.8 6.0 SD 2.9 4.4 Notes p=.68

OAS-M irritability subscale
Study Hollander 2001 Hollander 2001 Intervention Divalproex Sodium Placebo N 12 4 Mean 2.4 5.7 SD 1.9 0.6 Notes p = .53

OAS-M suicidality subscale
Study Hollander 2001 Hollander 2001 Intervention Divalproex Sodium Placebo N 12 4 Mean 0.9 0.3 SD 1.2 0.6 Notes p = .38

Analysis 08.07.
Review:

Comparison 08 MOOD STABILISERS vs PLACEBO, Outcome 07 Behaviour: 3. Not improved behavioural dyscontrol (AOS)

Pharmacological interventions for people with borderline personality disorder

Comparison: 08 MOOD STABILISERS vs PLACEBO Outcome: 07 Behaviour: 3. Not improved behavioural dyscontrol (AOS) Study Carbamazepine n/N De la Fuente 1994 Total (95% CI) 7/10 10 Placebo n/N 8/10 10 Relative Risk (Fixed) 95% CI Weight (%) 100.0 100.0 Relative Risk (Fixed) 95% CI 0.88 [ 0.53, 1.46 ] 0.88 [ 0.53, 1.46 ]

Total events: 7 (Carbamazepine), 8 (Placebo) Test for heterogeneity: not applicable Test for overall effect z=0.51 p=0.6

0.1 0.2

0.5

1

2

5

10

Favours control

Favours treatment

Pharmacological interventions for people with borderline personality disorder (Review) Copyright © 2007 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd

83


				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:17
posted:2/26/2009
language:English
pages:89