Efficacy of pharmacotherapy in depressed patients with and without personality disorders: A systematic review and meta-analysis

Abstract

Background

During the past decades personality pathology was considered to have a negative influence on the outcome of pharmacotherapy of depressive disorders. Recently, there has been a shift towards a less negative opinion. Still, the evidence in the literature remains inconclusive. This may be explained by methodological differences between published studies.

Objective

To present a meta-analysis of the results of Randomised Controlled Trials with pharmacotherapy in the treatment of depression with comorbid personality disorders.

Method

Systematic literature search for RCTs in adult ambulatory patients with major depressive disorder and comorbid PDs; pooling of data and meta-analysis according to strict methodological criteria.

Results

The difference in remission rates between the groups with and without personality disorders in high quality studies was 3%; this difference was neither statistically significant nor clinically relevant.

Limitations

Due to the specific and sensitive methods of the search only six studies could be included in the meta-analysis. Due to lack of data, analyses of drop-out rates could not be made.

Conclusion

When only data from high quality RCTs are included, comorbidity of personality disorder and major depression does not have a negative effect on the treatment outcome of pharmacotherapy for major depression.

Introduction

For a long time, personality disorders have been considered to affect negatively the outcome of treatment of depressive disorders. These considerations sprout from historical backgrounds: both Kraepelin, Freud and Abraham already speculated that comorbid personality pathology might be related not only to the etiology and manifestation of depression, but also to the responsiveness to treatment (Abraham, 1960, Freud, 1959, Ilardi and Craighead, 1995, Kraepelin, 1921). Given the impressive co-occurrence of personality disorders (PDs) and depression (Shea et al., 1990), it is highly relevant to study the consequences of this type of comorbidity. Earlier research mainly seems to underline the negative influence of comorbid personality disorders (PD) on treatment outcome in depression. In 1972, Paykel reported that patients with a past history of neurotic behavior, the “anxious depressives”, responded less well to treatment with amitryptiline (Paykel, 1972). A review on the results of tricyclic antidepressants also reported a poorer outcome for depressed patients with neurotic, hypochondriacal or hysterical personality traits (Bielski and Friedel, 1976). In the National Institute of Mental Health study, PDs were found to predict worse outcome of therapy and earlier recurrence of depression (Frank et al., 1987, Pilkonis and Frank, 1988). Also more recent review articles, covering the literature since 1990, have concluded that comorbidity was associated with significantly worse outcome in social functioning, higher levels of residual symptoms (Shea et al., 1990) and poorer treatment response (Reich and Green, 1991, Shea et al., 1992, Ilardi and Craighead, 1995). In 2002 however, a review of over 50 studies relating PD to treatment outcome concluded that the best-designed studies reported the least effect of personality pathology on treatment outcome of depression (Mulder, 2002). Most of the interventions in these studies were pharmacotherapy. Although overall high neuroticism scores predicted worse outcome, properly randomised and controlled studies of antidepressants appeared to show equal effectiveness in patients with or without comorbid PDs. According to Mulder, differences in findings on the association between PD and treatment outcome of depression were influenced by the quality of personality assessment, differences in treatment (especially in uncontrolled studies), differences in depression characteristics such as severity and chronicity of depression, and differences in treatment duration. It was concluded that comorbid personality pathology should not necessarily be seen as an impediment to good treatment response, although pooling of the then available higher quality studies was considered impossible due to substantial heterogeneity among these studies.

Recently a number of publications have appeared suggesting that indeed the outcome of depression treatment with comorbid PD may not be as negative as previously expected (Joyce et al., 2003, Russell et al., 2003). Bearing this in mind, we sought to review the currently available literature on this issue, and determine whether a meta-analysis can now be performed on this issue using only well-designed studies of pharmacotherapy for depression with or without PD of good methodological quality and limiting clinical heterogeneity. The central question was to determine whether there was a significant difference in the efficacy of pharmacotherapy for depressed patients with PDs compared to those without. Our hypothesis was that there would not be a major negative influence.