Longitudinal effects of dysfunctional perfectionism and avoidant personality style on postpartum mental disorders: Pathways through antepartum depression and anxiety

Highlights

• Dysfunctional perfectionism & avoidant personality style predict postpartum disorders.

• Antepartum mental disorders mediate the effect of personality on postpartum disorders.

• Antepartum depression affects postpartum disorders stronger than antepartum anxiety.

• Personality influences bonding impairment through postpartum depression.

Abstract

Background

There is first evidence that some personality characteristics raise the risk of postpartum depression (PPD). The present longitudinal study investigates whether dysfunctional perfectionism and avoidant personality style predict PPD, postpartum anxiety (PPA) and bonding impairment (BI) directly or indirectly through antepartum anxiety (APA) and antepartum depression (APD).

Methods

Pregnant women were recruited in two obstetric departments in Germany. The assessment occurred at two measurement time points: In the third trimester of pregnancy (N=297) and twelve weeks postpartum (N=266). Six questionnaires were administered during pregnancy: perfectionism, personality styles, anxiety, and depression. Postpartum, data on PPA, PPD and BI were collected. We conducted two path analyses in order to examine direct and indirect effects of the two personality characteristics on postpartum disorders.

Results

Testing for direct effects of dysfunctional perfectionism and avoidant personality style on PPD, PPA, and BI did not yield significant results. Instead, significant indirect effects were found: PPD, PPA, and BI were influenced indirectly by dysfunctional perfectionism and avoidant personality style via APD and APA. This model explained high portions of the variance of PPD, PPA, and impaired bonding. Each of the two personality characteristics explained a unique part of the outcome measures. The influence on BI was mediated by PPD. APD affected PPD and PPA more strongly than APA.

Limitation

Path models with manifest (observed) variables may lead to measurement errors. Self-rating questionnaires may raise the problem of social desirability.

Conclusion

Dysfunctional perfectionism and avoidant personality style are significant risk factors for PPD, PPA, and BI. Screenings of both variables, as well as of APA and APD, which mediated the effect of personality traits on postpartum syndromes, are necessary.

Introduction

Postpartum depression (PPD) is the most common mental illness in the perinatal period. It is characterized by typical symptoms of major depression, and in addition, by ambiguous feelings towards the child, sometimes culminating in infanticidal thoughts, doubts about functioning as a mother, and is often combined with anxiety. The high international prevalence of between 13% and 19% (O’Hara and Swain, 1996, O’Hara and McCabe, 2013), with a peak in the first three months (Hübner-Liebermann et al., 2012, Gavin et al., 2005) and its negative consequences for the child and the mother–infant relationship (Goodman et al. (2011), see O’Hara and McCabe (2013)) may explain the increase in research on PPD risk factors during the last decade. Although some risk factors have been detected so far, there is a lot to learn until we can predict PPD in its complexity, with a sufficient explanatory model (O’Hara and McCabe, 2013). In particular, it is not clear whether risk factors influence PPD directly or indirectly (e.g. in combination with pre-existing mood disorders).

Well-known risk factors for PPD are depression and anxiety during pregnancy, as well as a previous history of mood disorders, mainly depression, and lack of social support (Beck, 2001, Martini et al., 2015, Milgrom et al., 2008, O’Hara and Swain, 1996, Robertson et al., 2004). The prevalence of 20% for antepartum depression (APD) appears to be higher than that of PPD (Giardinelli et al., 2011, Hübner-Liebermann et al., 2012, Milgrom et al., 2008).

The prevalence of antepartum anxiety disorders varies considerably between different studies (see Goodman et al. (2014) for a review) from 12% to 25% (Bennett et al., 2004, Ross and McLean, 2006) to 50%–60% (Faisal-Cury and Menezes, 2007, Lee et al., 2007). Yet, there is little data on postpartum anxiety (PPA) prevalence. In Germany, a prevalence of 11.1% has been reported (Reck et al., 2008). Interestingly, anxiety disorders before or during pregnancy have proven to be a stronger predictor of PPD (and for PPA) than depression (Della Vedova et al., 2011, Lee et al., 2007; Matthey et al., 2003). Ross et al. (2003) reported that nearly 50% of all depressed women in the peripartum period suffered from co-morbid anxiety. Women with an anxiety disorder during pregnancy (antepartum anxiety, APA), especially those with high anxiety scores, have a threefold risk of intense PPD symptoms (Austin et al., 2007, Sutter-Dallay et al., 2004). Coelho et al. (2011) found that an antepartum generalized anxiety disorder strongly predicted PPD at different time points. Nevertheless, peripartum anxiety is understood less than depression, because up to now, most screenings focus on depressive symptoms (Goodman et al., 2014, Matthey et al., 2003).

In recent years, personality characteristics have been included in the research on risk factors for PPD. Some evidence exists that certain personality characteristics increase the vulnerability to PPD. Higher levels of neuroticism have been found to be a stable predictor for PPD (Dudek et al., 2014, Jones et al., 2010, Podolska et al., 2010), but neuroticism is generally a well-known risk factor for depression in other life periods, so that it might not predict PPD specifically. Some data also report harm-avoidance and self-directedness, traits from Cloninger's Temperament and Character Inventory (Josefsson et al., 2007), as correlated with PPD. Moreover, perfectionism has proven to be a predictor of different mood disorders and especially of major depression. Recently, it has been the subject of controversy in the context of PPD (Gelabert et al., 2012, Macedo et al., 2009, Vliegen et al., 2006). Perfectionism is a personality disposition defined as the setting of excessively high performance standards and overly critical evaluations of one's own behavior (cf. Frost et al. (1990)). Perfectionism contains both negative (maladaptive or unhealthy) and positive (adaptive or sound) facets (Hawkins et al., 2006, Stumpf and Parker, 2000, Terry-Short et al., 1995) and can be divided into functional perfectionism (personal standards, organization) and dysfunctional perfectionism (concern over mistakes, doubts about actions, parental expectations and parental criticism). Two studies reported associations between high (maladaptive/dysfunctional) perfectionism, mainly excessive concern over mistakes, and PPD (Gelabert et al., 2012, Mazzeo et al., 2006). In contrast, two other studies did not find a higher risk for PPD in perfectionist women (Maia et al., 2012, Sweeney and Fingerhut, 2013). Cloninger et al. (2012) showed that highly persistent (i.e., perfectionistic) people generally have a higher risk of suffering from anxiety disorders, but to the best of our knowledge, there has been no study analyzing the effect of perfectionism on PPA. The only personality characteristics that have recently been described as enhanced in peripartum anxiety disorders are low self-esteem and low self-efficacy (Martini et al., 2015).

Going beyond the discussion on the role of personality characteristics as predictors of PPD, personality disorders (PDs) have been analyzed increasingly in the context of PPD. There is preliminary evidence of some PDs being correlated with PPD, particularly Cluster C PDs: dependent, avoidant, and obsessive–compulsive (Aceti et al., 2012, Akman et al., 2007, Enfoux et al., 2013). There is little data on the causal relationship between PPD and Cluster C PDs. Akman et al. (2007) found that avoidant, dependent, and obsessive–compulsive PDs independently predict PPD. Especially high levels of avoidant PD symptoms seem to be a significant predictor of poor treatment outcome. Joyce et al. (2007) found that the major determinant of psychotherapy outcome was avoidant symptoms. Whether an avoidant PD not only predicts PPD, but also PPA, has not yet been investigated. Besides Cluster C PDs, there is also some evidence that Cluster B PDs, mainly borderline PD, is predominant in PPD mothers or often combined with avoidant/dependent PDs (Aceti et al., 2012, Apter et al., 2012, Enfoux et al., 2013). Since personality styles – in the sense of not fully pronounced PDs – are risk factors of mental illnesses (Caspi et al., 1996, Skodol et al., 2005) our aim was to analyze the effect of personality styles on PPD, PPA, and BI. The avoidant PD seems to be the most influential PD for PPD, based on the few existing studies and will be examined further in the present study, also in the context of PPA.

A recent study observed a relationship between PDs as risk factors of PPD and impaired maternal bonding (Schwarze et al., 2014). Impaired bonding is a disturbance of the mother's emotional relationship towards the infant, associated with a lack of maternal feeling, unresponsiveness, feelings of rejection, and sometimes aggressive impulses. In contrast, impaired attachment refers to a disturbed relationship of the infant with its caregiver (see Klaus et al. (1995) for the concept of bonding, and Ainsworth et al. (1978) and Bowlby (1997) for attachment theory). Mothers suffering from PPD often display an impaired relationship with the child (Dubber et al., 2015; Edhborgh et al., 2003; Moehler et al., 2006; Reck et al., 2004). According to Brockington et al. (2001), approximately 30% of mothers with PPD have a bonding impairment. In a German study, up to 75% of PPD mothers were found to be afflicted with impaired bonding (Hornstein et al., 2009). Yet, subclinical depressive symptoms also influence a mother's bonding (Moehler et al., 2006, Tietz et al., 2014). The effect of anxiety on bonding is less clear so far. There are only a few studies dealing with bonding impairment and anxiety (Dubber et al., 2015, Edborgh et al., 2011, Tietz et al., 2014) and the results are heterogeneous. It seems that BI correlates with PPA, but is influenced mainly by PPD (Dubber et al., 2015, Tietz et al., 2014).

The objective of our study was to investigate the relationships between personality characteristics, peripartum anxiety and depression, as well as BI, in a longitudinal study of women from the third trimester of pregnancy until twelve weeks postpartum. The present prospective study combines, for the first time, the analysis of distinct personality characteristics – perfectionism and avoidant personality style – as risk factors for PPD and PPA, as well as their relationships with BI. We integrated the well-known risk factors APD and APA in our model and tested the extent to which dysfunctional perfectionism and avoidant personality style directly influence both PPD and PPA, or whether there is an indirect pathway through antepartum illness (APA and/or APD). We assumed that there might be differential effects of the two personality characteristics on the two postpartum disorder entities and that the direct effects may be controlled by antepartum mental health. Because of the scarcity of data on the effect of personality styles on BI, we followed an explorative approach to examine the relationships between personality styles, peripartum disorders, and BI. Given that some studies reported correlations between PPD and BI (Brockington et al., 2001, Dubber et al., 2015; Edhborgh et al., 2003; Hornstein et al., 2009; Moehler et al., 2006; Reck et al., 2004; Tietz et al., 2014), we hypothesized that the effects of the personality characteristics on BI are mediated mainly by PPD.