Pharmacological interventions for antisocial personality disorder
Najat R Khalifa,Simon Gibbon,Birgit A Völlm,Natalie H-Y Cheung,Lucy McCarthy
DOI: https://doi.org/10.1002/14651858.CD007667.pub3
2020-09-03
Abstract:Background: Antisocial personality disorder (AsPD) is associated with rule-breaking, criminality, substance use, unemployment, relationship difficulties, and premature death. Certain types of medication (drugs) may help people with AsPD. This review updates a previous Cochrane review, published in 2010. Objectives: To assess the benefits and adverse effects of pharmacological interventions for adults with AsPD. Search methods: We searched CENTRAL, MEDLINE, Embase, 13 other databases and two trials registers up to 5 September 2019. We also checked reference lists and contacted study authors to identify studies. Selection criteria: Randomised controlled trials in which adults (age 18 years and over) with a diagnosis of AsPD or dissocial personality disorder were allocated to a pharmacological intervention or placebo control condition. Data collection and analysis: Four authors independently selected studies and extracted data. We assessed risk of bias and created 'Summary of findings tables' and assessed the certainty of the evidence using the GRADE framework. The primary outcomes were: aggression; reconviction; global state/global functioning; social functioning; and adverse events. Main results: We included 11 studies (three new to this update), involving 416 participants with AsPD. Most studies (10/11) were conducted in North America. Seven studies were conducted exclusively in an outpatient setting, one in an inpatient setting, and one in prison; two studies used multiple settings. The average age of participants ranged from 28.6 years to 45.1 years (overall mean age 39.6 years). Participants were predominantly (90%) male. Study duration ranged from 6 to 24 weeks, with no follow-up period. Data were available from only four studies involving 274 participants with AsPD. All the available data came from unreplicated, single reports, and did not allow independent statistical analysis to be conducted. Many review findings were limited to descriptive summaries based on analyses carried out and reported by the trial investigators. No study set out to recruit participants on the basis of having AsPD; many participants presented primarily with substance abuse problems. The studies reported on four primary outcomes and six secondary outcomes. Primary outcomes were aggression (six studies) global/state functioning (three studies), social functioning (one study), and adverse events (seven studies). Secondary outcomes were leaving the study early (eight studies), substance misuse (five studies), employment status (one study), impulsivity (one study), anger (three studies), and mental state (three studies). No study reported data on the primary outcome of reconviction or the secondary outcomes of quality of life, engagement with services, satisfaction with treatment, housing/accommodation status, economic outcomes or prison/service outcomes. Eleven different drugs were compared with placebo, but data for AsPD participants were only available for five comparisons. Three classes of drug were represented: antiepileptic; antidepressant; and dopamine agonist (anti-Parkinsonian) drugs. We considered selection bias to be unclear in 8/11 studies, attrition bias to be high in 7/11 studies, and performance bias to be low in 7/11 studies. Using GRADE, we rated the certainty of evidence for each outcome in this review as very low, meaning that we have very little confidence in the effect estimates reported. Phenytoin (antiepileptic) versus placebo One study (60 participants) reported very low-certainty evidence that phenytoin (300 mg/day), compared to placebo, may reduce the mean frequency of aggressive acts per week (phenytoin mean = 0.33, no standard deviation (SD) reported; placebo mean = 0.51, no SD reported) in male prisoners with aggression (skewed data) at endpoint (six weeks). The same study (60 participants) reported no evidence of difference between phenytoin and placebo in the number of participants reporting the adverse event of nausea during week one (odds ratio (OR) 1.00, 95% confidence interval (CI) 0.06 to 16.76; very low-certainty evidence). The study authors also reported that no important side effects were detectable via blood cell counts or liver enzyme tests (very low-certainty evidence). The study did not measure reconviction, global/state functioning or social functioning. Desipramine (antidepressant) versus placebo One study (29 participants) reported no evidence of a difference between desipramine (250 to 300 mg/day) and placebo on mean social functioning scores (desipramine = 0.19; placebo = 0.21), assessed with the family-social domain of the Addiction Severity Index (scores range from zero to one, with higher values indicating worse social functioning), at endpoint (12 weeks) (very low-certainty evidence). Neither of the studies included in this comparison measured the other primary outcomes: aggression; reconviction; global/state functioning; or adverse events. Nortriptyline (antidepressant) versus placebo One study (20 participants) reported no evidence of a difference between nortriptyline (25 to 75 mg/day) and placebo on mean global state/functioning scores (nortriptyline = 0.3; placebo = 0.7), assessed with the Symptom Check List-90 (SCL-90) Global Severity Index (GSI; mean of subscale scores, ranging from zero to four, with higher scores indicating greater severity of symptoms), at endpoint (six months) in men with alcohol dependency (very low-certainty evidence). The study measured side effects but did not report data on adverse events for the AsPD subgroup. The study did not measure aggression, reconviction or social functioning. Bromocriptine (dopamine agonist) versus placebo One study (18 participants) reported no evidence of difference between bromocriptine (15 mg/day) and placebo on mean global state/functioning scores (bromocriptine = 0.4; placebo = 0.7), measured with the GSI of the SCL-90 at endpoint (six months) (very low-certainty evidence). The study did not provide data on adverse effects, but reported that 12 patients randomised to the bromocriptine group experienced severe side effects, five of whom dropped out of the study in the first two days due to nausea and severe flu-like symptoms (very low-certainty evidence). The study did not measure aggression, reconviction and social functioning. Amantadine (dopamine agonist) versus placebo The study in this comparison did not measure any of the primary outcomes. Authors' conclusions: The evidence summarised in this review is insufficient to draw any conclusion about the use of pharmacological interventions in the treatment of antisocial personality disorder. The evidence comes from single, unreplicated studies of mostly older medications. The studies also have methodological issues that severely limit the confidence we can draw from their results. Future studies should recruit participants on the basis of having AsPD, and use relevant outcome measures, including reconviction.