Systematic review and meta-analysis of the role of personality disorder in randomised controlled trials of pharmacological interventions for adults with mood disorders

Highlights

• Personality disorder (PD) frequently co-occurs with mood disorders.

• A systematic review and meta-analysis was undertaken to examine the role of PD in randomised controlled trials (RCTs) of pharmacological interventions for mood disorders.

• PD does not appear to clearly affect pharmacological treatment or remission outcomes of co-occurring mood disorders.

• There is a lack of studies on bipolar disorder which examine the role of PD in relation to treatment outcomes.

ABSTRACT

Background

Personality disorder (PD) may affect the efficacy of pharmacological interventions for mood disorders, but the extent to which this occurs is uncertain. We aimed to examine the available published evidence concerning the role of PD in pharmacological treatment outcomes of randomised controlled trials (RCTs) for adults with mood disorders (i.e. depressive and bipolar spectrum disorders).

Methods

A systematic search of Cochrane Central Register of Controlled Clinical Trials, PubMed, EMBASE, PsycINFO, CINAHL Complete, and Google Scholar databases was undertaken to identify studies of interest. Data were independently extracted by two reviewers. The Cochrane Risk of Bias tool was used to assess methodological quality and risk of bias. A random effects model was utilised and statistical heterogeneity was assessed using the I² statistic. This systematic review was registered with PROSPERO (CRD42018089279) and the protocol is published.

Results

The search yielded 11,640 studies. Subsequent to removing duplicates, 9657 studies were screened at title and abstract stage and 1456 were assessed at full-text stage. Eighteen studies met criteria for inclusion in this review. Meta-analysis did not reveal a significant difference between groups for treatment outcome (standardised mean difference 0.22 [-0.09, 0.54]; I²: 69%, p=0.17) and remission (risk ratio 0.84 [0.64, 1.11]; I²: 51%, p=0.22).

Limitations

This review was limited by lack of studies on bipolar disorder.

Conclusion

PD comorbidity does not appear to affect treatment efficacy of pharmacological interventions for adults with mood disorders.

Introduction

Mood disorders, including depressive and bipolar spectrum disorders, are among the leading cause of disability worldwide (James et al., 2018). Despite the pharmacological interventions available for the treatment of mood disorders, many people still experience adverse outcomes and do not reach symptomatic remission. Personality disorder (PD) frequently co-occurs with mood disorders (Quirk et al., 2017) and might be a contributing factor to these negative outcomes. PD occurs when an individual's personality structure prevents them from attaining adaptive solutions to life's general tasks (Livesley, 2011). These tasks include having stable and cohesive representations of the self and others; developing the aptitude for intimacy, affiliation, and attachment; and being able to respond flexibly in social situations via prosocial behaviour and/or cooperative relationships (Livesley, 2011). Independently, PD is associated with poor physical health (El-Gabalawy et al., 2010; Quirk et al., 2016) and increased risk of suicide (Nordentoft et al., 2013). Comorbid PD has been shown to accelerate time to relapse and to slow time to remission from major depressive disorder (MDD), compared with individuals with MDD only; these outcomes have been shown to be most severe for those with obsessive-compulsive and borderline PDs (Grilo et al., 2010).

Previous reviews examining the effects of PD on treatment outcomes for depressive disorders have yielded mixed results across treatment modalities and have largely been limited by methodological discrepancies. Earlier reviews (Reich and Green, 1991; Reich and Vasile, 1993; Shea et al., 1992) suggest that PD has a negative effect on the treatment outcome of depressive disorders. Mulder (2002) found that the most rigorous of studies (i.e., PD assessment via structured clinical interview and treatment-controlled) showed no difference in treatment outcome between those with PD and depression and those with depression only. Reich (2003), in a narrative review, reported that maladaptive personality traits were associated with poorer treatment outcomes in most studies, but that no effect was present in studies where treatment length was longer (six to 18 months). Kool et al. (2005) examined randomised controlled trials (RCTs) of pharmacotherapies for depressed outpatients with/out comorbid PD (diagnosed by Diagnostic and Statistical Manual of Mental Disorders [DSM] criteria) and also found no significant difference between groups. Most recently, Young (2020) conducted a narrative review and concluded that the current literature supports the notion that PD can negatively affect treatment outcomes for individuals with MDD and that such patients should be treated with psychotherapeutic interventions. In contrast, Newton-Howes et al. (2006) and Newton-Howes et al. (2014) reviewed treatment response in studies of any treatment modality for depression. These reviews found that the presence of PD more than doubled the risk of poor outcome in cohort studies, while in RCTs this risk was one and a half times increased.

The research in relation to bipolar disorders is less robust. To the authors’ knowledge, only one review has examined the effect of PD on bipolar disorder treatment outcomes. Bieling et al. (2007) reviewed twelve studies that included patients with bipolar disorder and found that PD was associated with worse treatment outcome. However, there was a large amount of variability in the methodology of studies reviewed. Most studies were naturalistic and only one clinical trial was eligible for inclusion. In the included clinical trial, Preston et al. (2004) retrospectively diagnosed borderline PD in two samples of individuals with bipolar disorder trailing lamotrigine, and found no significant difference in outcome between individuals with and without borderline PD.

Given the substantial heterogeneity in studies of co-occurring PD and depressive disorder and the limited literature on PD and bipolar disorder treatment outcomes, an updated review on the topic is timely. The aim of this systematic review, therefore, was to determine if the presence of PD is associated with a differential treatment outcome in RCTs examining the effects of a pharmacological intervention for the treatment of a mood disorder. Specifically, the principal objective was to examine whether treatment outcomes at the end of the RCT treatment phase differed between individuals with or without PD.