REP 4

Contents lists available at ScienceDirect

Child Abuse & Neglect

journal homepage: www.elsevier.com/locate/chiabuneg

Full length article

How is childhood emotional abuse related to major depression in adulthood? The role of personality and emotion acceptance

Philipp Schulza,⁎, Thomas Bebloa, Hedda Ribbertb, Leona Katerc, Stephanie Spannhorstb, Martin Driessena, Kristina Hennig-Fastb,d,⁎⁎

a Evangelisches Klinikum Bethel (EvKB), Department of Psychiatry and Psychotherapy, Research Division, Remterweg 69-71, D-33617 Bielefeld, Germany b Evangelisches Klinikum Bethel (EvKB), Department of Psychiatry and Psychotherapy, Division of General Psychiatry, Bethesdaweg 12, D-33617 Bielefeld, Germany c Evangelisches Klinikum Bethel (EvKB), Department of Internal and Geriatric Medicine Johannesstift, Schildescher Straße 99, D-33611 Bielefeld, Germany d Faculty of Psychology, Department of Applied Psychology, Health, Development, Enhancement, and Intervention, University of Vienna, Liebiggasse 5, 1010 Vienna, Austria

A R T I C L E I N F O

Keywords: Child maltreatment Emotional abuse Depression Personality Emotion regulation Mediation

A B S T R A C T

Accumulated evidence provides support that childhood emotional abuse (CEA) is related to adult major depressive disorder (MDD) outcomes. However, the psychological mechanisms of this relation are still not well understood. Changes in personality and emotion regulation are in- dicated to play a mediating role what should be examined in this paper. A sample of 123 MDD inpatients was examined in a prospective observational study with two times of measurement. Patients provided data on childhood trauma history, personality disorder (PD) traits and emotion acceptance. Self- and expert-ratings of depressive symptoms were assessed at baseline and at the end of treatment. Treatment duration as an objective indicator of treatment outcome was ad- ditionally considered. Partial correlation analyses revealed associations between CEA and self- ratings of MDD symptom severity and symptom improvement independent of sexual and physical abuse. Expert-ratings of depression and treatment duration were not related to CEA. Mediation analyses revealed that particularly the factors borderline psychopathology as well as acceptance of pleasant emotions mediated the association of CEA and self-rated MDD symptoms. Passive- aggressive PD traits mediated the link between CEA and a lower self-rated symptom improve- ment. CEA affect specific personality traits and acceptance of emotions. This association may play a critical role for self-reported depressive symptoms with implications for prevention, psychoe- ducation, and treatment of MDD.

1. Introduction

Child maltreatment is known to be a risk factor in development of MDD (Driessen, Schroeder, Widmann, Schönfeld, & Schneider, 2006; Silverman, Reinherz, & Giaconia, 1996) and a predictor of a negative course of MDD treatment (Harkness, Bagby, & Kennedy,

http://dx.doi.org/10.1016/j.chiabu.2017.07.022 Received 12 January 2017; Received in revised form 27 July 2017; Accepted 28 July 2017

⁎ Corresponding author at: Evangelisches Klinikum Bethel (EvKB), Department of Psychiatry and Psychotherapy, Research Division, Remterweg 69-71, D-33617 Bielefeld, Germany. ⁎⁎ Corresponding author at: Evangelisches Klinikum Bethel (EvKB), Department of Psychiatry and Psychotherapy, Division of General Psychiatry, Bethesdaweg 12,

D-33617 Bielefeld, Germany. E-mail addresses: [email protected] (P. Schulz), [email protected] (K. Hennig-Fast).

Child Abuse & Neglect 72 (2017) 98–109

Available online 05 August 2017 0145-2134/ © 2017 Elsevier Ltd. All rights reserved.

T

2012; Nanni, Uher, & Danese, 2012; Tunnard et al., 2014). A growing body of evidence provides support that particularly childhood emotional abuse (CEA; including emotional neglect) is related to the symptoms and course of MDD (Chapman et al., 2004; McLaughlin et al., 2010; Suija, Aluoja, Kalda, & Maaroos, 2011; Widom, DuMont, & Czaja, 2007). However, the psychological me- chanisms of this relation are still not well understood possibly due to the fact that until recently, research focused on “more obvious” subtypes of maltreatment, predominantly sexual and physical abuse. Thereby, evidence indicates that the impact of CEA is at least comparable to childhood sexual and physical abuse with regard to the risk of onset of depression (Carr, Martins, Stingel, Lemgruber, & Juruena, 2013; Chapman et al., 2004; Infurna et al., 2016; Liu, Alloy, Abramson, Iacoviello, & Whitehouse, 2009) and symptom severity of MDD (Carvalho Fernando et al., 2012; Martins, Baes, Tofoli, & Juruena, 2014; Spertus, Yehuda, Wong, Halligan, & Seremetis, 2003). Remarkably, results of a recent systematic review and meta-analysis suggest that the risk of developing a depressive disorder across life-span is almost double that of survivors of CEA compared to survivors of childhood physical abuse (Norman et al., 2012). Given the relatively high prevalence of CEA in the population and the adverse effects in relation to MDD (Taillieu, Brownridge, Sareen, & Afifi, 2016), our study aims to shed more light on the psychological mechanisms beyond this re- lation.

Although the intermediate pathways between CEA and MDD outcomes still remain unclear, different psychological consequences of child maltreatment have been described, which may explain the link to MDD, e.g. altered emotion regulation strategies (Cicchetti & Toth, 2005), decrease of self-esteem (Finzi-Dottan & Karu, 2006), and a dysfunctional attributional style which can foster a negative cognitive style (Alloy, Abramson, Smith, Gibb, & Neeren, 2006; Gibb et al., 2001). Referring to hopelessness theory of depression by Abramson, Metalsky, & Alloy (1989) it is hypothesized that specifically CEA contribute to the development of negative cognitive styles. Thus, CEA may work as a crucial risk factor for psychological dysfunction across life-span.

1.1. The impact of personality traits/disorders to the association between CEA and MDD

It has been demonstrated that sexual and physical abuse (Brown & Anderson, 1991) but also CEA are associated with several personality disorder traits/PDs in non-clinical (Tyrka, Wyche, Kelly, Price, & Carpenter, 2009) and clinical samples (Cohen et al., 2014; Wingenfeld et al., 2010). Particularly, borderline personality disorder (BPD) was found to be strongly affected by experiences of childhood sexual abuse, physical abuse (Brown & Anderson, 1991) and by CEA (Zanarini et al., 1997). Several studies revealed associations between experiences of CEA and different areas of personality dysfunction: In a sample of psychiatric inpatients Lobbestael, Arntz, & Bernstein (2010) showed that emotional abuse was independently associated with paranoid, schizotypal, bor- derline, and cluster C PDs; emotional neglect was independently associated with histrionic and borderline PD. In accordance with Lobbestael’s findings Cohen et al., (2014) showed that emotional abuse was specifically associated with cluster C personality traits in a mixed sample of psychiatric patients. A recent review investigated the specific contributions of child abuse subtypes to adult psychiatric disorders (Carr et al., 2013). The authors reported that emotional abuse was associated with borderline, narcissistic, and passive-aggressive PDs. Nevertheless, the vast majority of latter presented literature captured associations between personality traits/ disorders and CEA within non-representative samples. In this regard, Taillieu et al. (2016) investigated a nationally representative sample from the United States (n = 34.653): The authors showed that CEA was associated with an increased risk of a lifetime diagnosis of several axis II mental disorders. In sum CEA seem to foster dysfunctions with regard to various personality domains, which suggests that it affects socio-emotional competence as a superordinate concept of all personality clusters (e.g., Saarni, 1999).

In addition, several epidemiological studies have demonstrated that the prevalence of comorbid PDs in MDD inpatients is much higher (20–50%) compared to the general population (6–10%; Corruble, Ginestet, & Guelfi, 1996) and that these depressed patients with comorbid PDs tend to show poorer treatment outcomes (Newton-Howes, Tyrer, & Johnson, 2006). More specific, cluster B and C PD traits appear to be critical for an unfavorable course of MDD (Iacoviello, Alloy, Abramson, Whitehouse, & Hogan, 2007).

Taken together, empirical findings suggest that changes in personality functioning may mediate the link between child mal- treatment and adult MDD, which is in line with studies examining samples of adolescents (Gibb et al., 2001), a study of an epide- miological sample (Sato, Uehara, Narita, Sakado, & Fujii, 2000), and with studies in clinical MDD samples (Huh, Kim, Yu, & Chae, 2014; Kounou et al., 2013). However, there is still a lack of studies investigating the specific contribution of CEA to PD traits, which may mediate the link between CEA and MDD, beyond the effects of further child maltreatment subtypes.

1.2. The meaning of emotion regulation to the association between CEA and MDD

In a comprehensive review it was highlighted that emotion regulation strategies should be of interest to understand the link between child CEA and MDD (Alloy et al., 2006). Emotion regulation is defined as a psychological process including various skills such as acceptance of emotions, awareness of emotions, the ability to control impulsive behaviors, the ability to act in line with desired goals when experiencing negative emotions and to choose the appropriate emotion regulation strategy as required by an emotionally challenging situation (Gratz & Roemer, 2004). Several adverse effects of child maltreatment on emotion regulation in children are postulated by Cicchetti and Toth (2005) and Maughan and Cicchetti (2002), worth mentioning here are difficulties in discrimination of emotional expressions, a response bias to angry emotional expressions, and affective lability as well as socially inappropriate emotion expressions. Previous studies have proven evidence for the persistence of such emotion regulation deficits in a sample of children (Kim & Cicchetti, 2010) and in a sample of women with histories of child adversities (Cloitre, Miranda, Stovall, & Han, 2005). Regarding the specific impact of CEA on adult emotion regulation, Burns, Jackson, & Harding (2010) showed that emotional abuse was a stronger predictor of a range of emotion regulation difficulties compared to physical and sexual abuse. In agreement with this finding Carvalho Fernando et al. (2014) observed in a clinical sample that emotional abuse and neglect were

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

99

specifically related to difficulties in emotion regulation, independent from the influence of sexual abuse and physical abuse. A further study by Gratz et al. (2007) revealed that emotional non-acceptance was more strongly associated with emotional abuse than with sexual or physical abuse.

Etiological theories postulate that altered emotion regulation strategies play a critical role for the development of MDD (Barlow, Allen, & Choate, 2004; Gross & Muñoz, 1995). Positive and negative effects of different emotion regulation strategies on depressive disorders were investigated in a meta-analysis (Aldao, Nolen-Hoeksema, & Schweizer, 2010). The results show that acceptance, problem solving, and reappraisal were protective factors while avoidance, rumination, and suppression were risk factors for de- pressive disorders. Generally, it is well described that the suppression of negative emotion is a common dysfunctional emotion regulation strategy in MDD patients as confirmed in the latter mentioned meta-analysis by Aldao et al. (2010). However, evidence investigating the role of positive emotions in MDD patients is scarce. In this regard the results of a cross-sectional study showed that MDD inpatients reported a reduced acceptance of negative and positive emotions, and both, reduced acceptance of negative and positive emotions were accompanied by higher levels of depressive symptoms (Beblo et al., 2012).

Only few studies investigated if emotion regulation impairments, affected by child maltreatment, result MDD outcomes in later life. With respect to CEA O'Mahen, Karl, Moberly, & Fedock (2015) revealed that emotion regulation deficits (partially) mediated the link between emotional abuse and depression in a sample of pregnant women. Furthermore, in a large sample of low-income African Americans emotion dysregulation partially mediated the effect of emotional abuse on depressive symptoms (Crow, Cross, Powers, & Bradley, 2014). Recently, it was shown that emotion regulation partially mediated the relationship between child mal- treatment and depressive outcomes within a large sample of depressed inpatients; more specifically, acceptance, awareness, toler- ance, and willingness to confront as emotion regulation strategies partially mediated the impact of child maltreatment on depression severity (Hopfinger, Berking, Bockting, & Ebert, 2016). Nevertheless, there is a lack of evidence on the specific role of emotion acceptance as a mediator between CEA and MDD outcomes. Considering that CEA predicts emotional non-acceptance (Gratz, Bornovalova, Delany-Brumsey, Nick, & Lejuez, 2007) and emotional non-acceptance predicts depression (Aldao et al., 2010; Beblo et al., 2012; Hopfinger et al., 2016) it is relevant to know if there apply different effects of the acceptance of positive or negative emotions on the relation between CEA and depression what is also addressed in this paper.

1.3. Hypotheses

Hypothesis (1). CEA is associated with elevated levels of depressive symptoms, a diminished symptom improvement, and longer treatment duration, even when controlling for sexual and physical abuse.

Hypothesis (2). CEA is associated with a lower acceptance of emotions and elevated PD traits in all personality clusters (according to DSM-IV), even when controlling for sexual and physical abuse.

Hypothesis (3). Elevated Cluster B and C PD traits as well as altered emotion acceptance mediate the links between CEA and depression severity, symptom improvement, and treatment duration, while controlling for sexual and physical abuse.

An exploratory analysis should clarify if PD traits and emotion acceptance contribute to the expected link between CEA and MDD to a comparable extent, also when controlled for sexual and physical abuse.

2. Methods

2.1. Design and procedures

This prospective observational study enrolled patients treated on a depression ward in the Clinic of Psychiatry and Psychotherapy of the Evangelisches Klinikum Bethel (Germany) between February 2013 and November 2015. The multimodal routine therapy consisted of cognitive behavioral and psychodynamic therapy elements following German clinical guidelines. Treatment services consisted of psychotherapeutic single and group sessions, psychoeducation, exercise therapy, relaxation training, imagination, oc- cupational therapy, art therapy, music therapy, discussions with primary nurses etc. Each patient received his/her own schedule with individual frequencies and combinations of treatment services. However, each patient received at least one weekly therapeutic single and group session, a weekly psychoeducation session in groups, a weekly discussion with their primary nurse, and usually two other treatment services (e.g., exercise therapy and imagination) per day. Within the first seven days after hospitalization participants received questionnaires capturing depressive symptoms. During the weekly visit further questionnaires assessing childhood trauma history, PD traits, and emotion acceptance were handed out in the above order. Approximately five days prior to discharge parti- cipants received questionnaires capturing depressive symptoms again. All participants were adults and gave their written informed consent prior to study participation. All study procedures (questionnaires) were part of the standard diagnostic procedures in the clinic. Ethical standards were in accordance with the declaration of Helsinki.

2.2. Participants

A total of 212 inpatients were recruited for study participation according to the following inclusion criteria: Current MDD as primary diagnosis, 18–65 years of age and sufficient German skills. Exclusion criteria were current substance dependence (e.g., F10.2), bipolar and/or psychotic disorder, cognitive impairments due to neuropsychiatric disorders and treatment duration of less

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

100

than four weeks. A total number of 89 inpatients were excluded from analysis due to invalid data (n = 14), psychiatric reasons (n = 17), a short treatment duration (n = 16), and/or missing information on childhood trauma history (n = 66), resulting in an included sample size of n = 123. No significant differences were observed between included (n = 123) and excluded samples (n = 89) with respect to sex, educational level, employment, treatment experience, and self-rated as well as expert-rated depression severity at baseline and after six weeks of treatment.

2.3. Measures

2.3.1. Diagnoses Mental disorders were diagnosed according to ICD-10 by experienced and trained therapists (World Health Organization, 1992).

To ensure precise assessment of the current MDD episode, ICD-10 international diagnosis checklists (IDCL) completed the diagnosis (Hiller, Zaudig, & Mombour, 1995); for assessment of personality psychopathology see section personality disorder traits.

2.3.2. Childhood trauma To assess self-reported childhood emotional, physical, and sexual abuse as well as emotional and physical neglect, participants

completed the German version of the Childhood Trauma Questionnaire (CTQ; Bernstein et al., 2003; Wingenfeld et al., 2010). Each CTQ scale consists of five items with scores ranging from 5 (none or minimal) to 25 (severe to extreme). The CTQ is a reliable and widely used questionnaire to assess self-reported maltreatment experiences in adults retrospectively. Good to excellent psychometric properties have been demonstrated in several samples of psychiatric inpatients (Carvalho Fernando et al., 2012; Carvalho Fernando et al., 2014), apart from the subscale physical neglect (Wingenfeld et al., 2010). Thus, we excluded this subscale from the analysis. To assess childhood emotional abuse (CEA) we summed up the scores of the emotional abuse and emotional neglect subscales, which were highly correlated (r = 0.75, p < 0.001). Total scores of CEA, sexual and physical abuse were used for statistical analyses in this paper.

2.3.3. Depressive symptoms and treatment The self-rated depression severity was assessed by the German version of the Beck depression inventory-second edition (BDI-II),

which is a reliable and valid self-rating instrument containing 21 items (Beck, Steer, & Brown, 1996; Hautzinger, Keller, & Kühner, 2006). The expert-rated depression severity was captured by attending therapists by the German version of the Montgomery-Åsberg Depression Rating Scale (MADRS) (Montgomery & Asberg, 1979; Schmidtke, Fleckenstein, Moises, & Beckmann, 1988). The sum score of the BDI was used to assess self-rated depression severity at baseline and the sum score of the MADRS was used to assess expert- rated depression severity at baseline. The difference between the sum score at baseline and on average after six weeks of treatment served as a measure for symptom improvement for both, BDI (self-rated symptom improvement) and MADRS (expert-rated symptom improvement). Self- or expert-rated treatment response was calculated, if there was at least a 50% decrease in the sum scores of BDI or MADRS from baseline until end of treatment. As an objective indicator of treatment outcome duration of treatment was assessed (days). Shorter treatment duration should reflect a better treatment outcome.

2.3.4. Personality disorder traits To assess personality psychopathology we used the Personality Style and Disorder Inventory (PSDI, German version; Kuhl and

Kazén, 2009), a widely used self-report measurement. The PSDI allows assessing DSM-IV and ICD-10 personality traits as a di- mensional approach: Characteristics of specific personality styles vary between adaptive to clinical levels, which may reflect a PD. It includes 140 items, each rated on a four point scale ranging from 0 (does not apply at all) to 3 (fully applies). In our study the following 11 PD subscales were assessed: paranoid, schizoid, schizotypal, antisocial, borderline, histrionic, narcissistic, avoidant, dependent, ana- ncastic and negativistic (passive-aggressive). An example for an item of the borderline scale is: “My feelings often change abruptly and impulsively”. The authors of the PSDI demonstrated sufficient to good internal consistencies for the different personality subscales (α = 0.73 to α = 0.85; Kuhl & Kazén, 2009), the validity has been demonstrated in clinical samples (e.g., Baumann, Kaschel, & Kuhl, 2007). T values, adjusted for sex and education were used for the current analyses.

2.3.5. Emotion acceptance Emotion acceptance was captured by means of the “Emotion Acceptance Questionnaire” (EAQ, in German FrAGe; Beblo et al.,

2011). Participants were asked to indicate whether items applied to themselves, on a scale ranked from 1 (does not apply at all) to 6 (applies completely). The 32 items of the EAQ cover the following four subscales by an equal number of items: (1) acceptance of negative emotions (e.g., “I usually allow myself to accept unpleasant feelings”), (2) suppression of negative emotions, (e.g., “I try to push aside unpleasant feelings”), (3) acceptance of positive emotions (e.g., “I can easily let in pleasant feelings”), and (4) suppression of positive emotions (e.g., “I block out pleasant feelings”). The total score for the acceptance of unpleasant emotions was built by subscales 1 and 2. The total score for the acceptance of pleasant emotions was built by subscales 3 and 4. All four subscales are summed up to create the total score (acceptance of emotions). Item scores of subscales 2 and 4 are recoded (1 = 6, 2 = 5, 3 = 4, 4 = 3, 5 = 2, 6 = 1) to build the scores of the main scales. The initial validation study revealed very good reliabilities for the total score (α = 0.91, split half = 0.89) and the subscales for negative (α = 0.89, split half = 0.89) and positive (α = 0.90, split half = 0.83) emotions (Beblo et al., 2011). The EAQ has already been administered in psychiatric samples confirming its good psychometric properties including correlations with comparable questionnaires (Beblo et al., 2013; Beblo et al., 2012). Raw scores of the total scores for the scales

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

101

acceptance of emotions, acceptance of pleasant emotions and acceptance of unpleasant emotions were used in the current work. Means of EAQ scales in the current work were comparable to means of EAQ scales in MDD inpatients previously reported (Beblo et al., 2012).

2.4. Data analysis

In order to identify confounding variables pre-analyses were run in terms of correlation analyses between predictor/criterion variables (maltreatment subtypes/depression measures) and demographics. All child maltreatment subtypes were significantly in- terrelated and, on a zero-order level, significantly related to some criterion variables. Sex and employment were significantly related to CEA. Thus, maltreatment subtypes and sex as well as employment were included as covariates in all statistical models used in this paper.

To test the hypothesis that CEA is specifically related to depression severity, symptom improvement and treatment duration, partial correlations were run while controlling for sex, employment, and other maltreatment subtypes. Baseline depression severity was controlled for analyses with symptom improvement and treatment duration as outcome measure.

To identify mediators for the link between CEA and MDD, we computed partial correlations between CEA and PD traits as well as emotion acceptance scales. PD traits and emotion acceptance scales, significantly associated with CEA, were included in subsequent simple mediation analyses.

For mediation analyses CEA served as the predictor, personality traits and emotion acceptance scales as mediator and depression measures as outcome variables. In order to reveal the specific contribution of the different mediators on the association between CEA and depression severity, we conducted an exploratory, comparative mediation analysis. Thus, we took into account the significant mediators simultaneously in one model. Mediators were defined as significant, if an indirect-only effect occurred (significant indirect and insignificant direct effect; Zhao, Lynch, & Chen, 2010).

To test the significance of indirect effects we calculated Sobel-tests and the bootstrap enhanced confidence intervals (CIs). The CI for the indirect effect was a bias corrected bootstrapped CI based on 10.000 bootstrap samples. We set the critical α at 0.05 for each analysis using two-tailed testing. For statistical analyses SPSS 20.0 and the PROCESS-tool 2.13 were used (Preacher & Hayes, 2004).

2.4.1. Missing data To account for missing data we imputed missing values for mediator and outcome variables by the multiple imputation procedure

(Graham, 2009). The proportion of missing data on variables ranged from 1% (BDI baseline) to 21% (EAQ scale) and Little’s missing completely at random-test revealed that missing data were completely at random, χ2(438) = 401.34, p = 0.90. Given the relatively small sample size and high proportion of missing data for some variables, we imputed 20 data sets as recommended (Graham, 2009). Due to skewed distributions of some imputed variables, we used the predictive mean matching algorithm which has shown to provide plausible imputations, despite non-normal distributed variables (Vink, Frank, Pannekoek, & van Buuren, 2014). For all correlation analyses the pooled effect sizes were reported. Since PROCESS is not compatible with pooling procedure in multiple imputation data sets using SPSS, we chose the first imputed data set to conduct the mediation analyses (Graham, 2009). All analyses were additionally run with a complete-case data set (n = 72) leading to comparable results.

3. Results

3.1. Sample characteristics and clinical data

Sample characteristics are displayed in Table 1. In 36 out of 123 (29.3%) participants axis I comorbid disorders were ascertained. Anxiety disorders in 16 patients (five patients with PTSD), current and previous substance misuse (e.g., F10.1) in 13 patients, dysthymia in six patients, eating disorders in three patients. Four patients had two psychiatric comorbid disorders. Six patients had a clinically ascertained PD. Exploratory analyses revealed that these six patients were not outliers with respect to all relevant variables (i.e., depression, child maltreatment history etc.). Five participants did not have any school degree (4.1%), 58 a secondary school degree (47.2%), and 60 a high school degree (48.8%).

CEA was significantly associated with female sex (r = 0.27, p < 0.01), unemployment (r = −0.25, p < 0.01), physical abuse (r = 0.64, p < 0.001) and sexual abuse (r = 0.30, p < 0.001). According to the cut-off-scores proposed by Bernstein and Fink (1998) at least slight to moderate abusive experiences were reported by 75.6% for emotional neglect, 62.6% for emotional, 34.4% for physical and 26.0% for sexual abuse. A previous study showed comparable levels of abusive experiences in a large sample of MDD patients (Wingenfeld et al., 2010). The mean score of the BDI was 30.6 (SD = 11.2) indicating currently severe depressive symptoms (Beck et al., 1996), the mean score of the MADRS was 27.3 (SD = 6.3) indicating currently moderate depressive symptoms (Neumann & Schulte, 1989). The rate of treatment responder was 36.1% respecting BDI and 40.3% respecting MADRS (in the complete-case data set). Correlations of self-rated and expert-rated depression severity were significant at baseline and at the end of treatment (r = 0.25, p < 0.01; r = 0.45, p < 0.001), but self- and expert-rated symptom improvement did not significantly cor- relate (r = 0.14, p = 0.16).

3.2. Association between CEA and depression

When controlling for sex, employment, and the other maltreatment subtypes solely CEA was found to be significantly and po- sitively associated with self-rated baseline depression severity and negatively associated with symptom improvement (r = 0.28,

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

102

p < 0.01; r = −0.19, p < 0.05; see Table 2). Neither the associations between CEA and expert-rated baseline depression severity and symptom improvement nor between CEA and treatment duration were significant.

3.3. Association between CEA and personality disorder traits/emotion acceptance

Descriptive statistics of the PSDI and EAQ subscales as well as partial correlations with CEA are presented in Table 3. Clinical expressions of PD traits were most often observed for anancastic (8.2%), passive-aggressive (7.9%), and avoidant (7.0%) PDs with T- Values two standard deviations above the mean. Higher expressions of CEA were significantly associated with paranoid, passive- aggressive, borderline and schizoid personality disorder traits, when controlling for sex, employment, sexual and physical abuse. Higher expressions of CEA were also correlated to lower emotion acceptance, particularly to lower acceptance of pleasant emotions. As the total score of emotion acceptance is composed of the two EAQ subscales, we included solely the subscale acceptance of pleasant emotions in subsequent mediation analyses.

Table 1 Sample characteristics of N = 123 inpatients.

Baseline and Clinical Data N M SD Range

baseline data age 40.3 11.8 18–64 sex (male/female) (n) 73/50 years of education 11.2 1.6 8–14 employment (unemployed/employed) (n) 49/74 marital status (no partnership/partnership) (n) 74/49

clinical data previous treatment experiences yes/no (n) 89/34 duration of current completed treatment (days) 47.7 10.1 28–87 recurrent depressive disorder yes/no (n) 50/73 severity of current episode (moderate/severe) (n) 37/86 axis I comorbidity yes/no (n) 36/87 AD yes/no (n)a 110/13

SSRI 44/110 Tri- and Tetracyclic AD 25/110 MAO-inhibitor 1/110 Other AD 57/110

self-rated baseline depression severity (BDI) 30.6 11.2 0–58 self-rated end of treatment depression severity (BDI) 20.0 13.2 0–58 self-rated symptom improvement (BDI) 10.6 9.3 −13 to 28 expert-rated baseline depression severity (MADRS) 27.3 6.3 4–44 expert-rated end of treatment depression severity (MADRS) 15.3 5.3 3–30 expert-rated symptom improvement (MADRS) 12.0 7.9 −3 to 39 childhood emotional abuse and neglect (CTQ) 26.2 10.7 10–50 childhood physical abuse (CTQ) 8.1 4.8 5–25 childhood sexual abuse (CTQ) 6.5 3.8 5–25

AD = Antidepressant, SSRI = Selective serotonin reuptake inhibitor, MAO = Monoamine oxidase, BDI = Beck Depression Inventory, MADRS = Montgomery-Åsberg Depression Rating Scale, CTQ = Childhood Trauma Questionnaire, M = Mean, SD = Standard deviation.

a n = 17 patients received two antidepressants simultaneously.

Table 2 Partial correlations between maltreatment subtypes and self-rated as well as expert-rated depression severity, symptom improvement and treatment duration. Correlations between the respective maltreatment subtypes (predictors) and the five criterion variables were controlled for the other types of child maltreatment respectively.

Correlations with Self-rated Depression Correlations with Expert-rated Depression Correlations with Indicator of Treatment Outcome

Maltreatment subtypes Baseline depression severity

Symptom improvement

Baseline depression severity

Symptom improvement

Treatment duration (days)a

Childhood emotional abuseb

0.28** −0.19* 0.07 −0.08 0.00

Childhood physical abuse −0.11 −0.12 0.02 0.00 0.08 Childhood sexual abuse 0.15 0.17 0.09 −0.04 −0.05

Correlations were controlled for sex, employment, and the other maltreatment subtypes respectively. Correlations with symptom improvement and treatment duration were additionally controlled for depression severity at baseline. Not significant = p > 0.05,* p < 0.05, ** p < 0.01. Significant effects are shown in bold.

a Positive correlations indicate that more experiences of child abuse are associated with longer treatment duration. b Childhood emotional abuse = Sum score of the CTQ scales emotional abuse and neglect.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

103

3.4. Mediation analyses

Results of the correlation analyses confirmed significant and specific associations between CEA and self-rated depression out- comes, PD traits and acceptance of pleasant emotions. Thus, we conducted subsequent mediation analyses to assess the mediating role of PD traits and emotion acceptance. The results of simple mediation analyses are shown in Table 4. There were significant indirect effects of CEA on the relation of depression severity and passive-aggressive PD traits as well as BPD traits and for acceptance of pleasant emotions. We found solely a significant indirect effect of CEA on the relation of symptom improvement and passive- aggressive PD traits.

Table 3 Descriptive statistics of personality disorder traits and emotion acceptance as well as partial correlations with childhood emotional abuse (CEA), controlled for other types of child maltreatment.

Data on PSDI and EAQ scales M SD Range % T-values above 70 Partial r with CEAa

personality disorder traits (PSDI T-values) cluster A paranoid 56.1 8.6 29 – 77 3.5% 0.43*** schizoid 59.8 8.8 39 – 80 6.1% 0.30** schizotypal 46.1 8.9 25 – 66 0% 0.04

cluster B antisocial 44.7 10.4 31 – 72 0.9% 0.12 borderline 56.5 9.1 31 – 80 6.1% 0.39*** histrionic 44.0 9.7 21 – 65 0% 0.01 narcisstic 42.9 10.5 21 – 70 0.9% 0.07

cluster C avoidant 58.6 9.5 33 – 80 7.0% 0.12 dependent 55.0 9.2 33 – 80 2.6% 0.13 anancastic 57.7 9.8 35 – 80 8.2% 0.05

appendix of DSM-IV passive-aggressive 58.1 9.5 35 – 79 7.9% 0.42***

emotion acceptanceb (raw scores) acceptance of emotions 3.5 0.8 1.8 – 5.6 Na −0.32*** acceptance of unpleasant emotions 2.9 1.0 1.2 – 5.3 Na −0.11 acceptance of pleasant emotions 4.1 1.1 2.1 – 6.0 Na −0.38***

CEA = Childhood emotional abuse, M = Mean, SD = Standard deviation; Correlations were controlled for sex, employment, physical, and sexual abuse. Not sig- nificant = p > 0.05, ** p < 0.01, *** p < 0.001. Significant effects are shown in bold.

a Sum score of the CTQ scales emotional abuse and neglect. b Values for raw scores could range from 1 to 6, higher values reflect a higher acceptance of emotions.

Table 4 Simple mediation analyses. In each model childhood emotional abuse served as predictor variable (X).

Mediators (M) Outcomes (Y) b of direct effect path X → Y b of path X → M b of path M → Y β of indirect patha

paranoid personality depression severity (BDI) 0.28* 0.43*** 0.24 0.07 [−0.01, 0.20] symptom improvement (BDI) −0.17 0.39*** −0.11 −0.04 [−0.12, 0.04]

schizoid personality depression severity (BDI) 0.31* 0.32*** 0.21 0.05 [−0.01, 0.15] symptom improvement (BDI) −0.21* 0.28** −0.01 −0.00 [−0.05, 0.05]

passive-aggressive personality depression severity (BDI) 0.25 0.49*** 0.27* 0.10* [0.01, 0.22] symptom improvement (BDI) −0.12 0.43*** −0.23* −0.09* [−0.18, −0.02]

borderline personality depression severity (BDI) 0.14 0.43*** 0.58*** 0.18*** [0.10, 0.29] symptom improvement (BDI) −0.20* 0.30** −0.05 −0.01 [−0.09, 0.05]

accpetance of pleasant emotions depression severity (BDI) 0.14 −0.05*** −5.00*** 0.18*** [0.09, 0.29] symptom improvement (BDI) −0.19 −0.04*** 0.70 −0.02 [-0.12, 0.04]

The models with depression severity at baseline as outcome variable were controlled for sex, employment, physical, and sexual abuse. The models with symptom improvement as outcome variable were additionally controlled for baseline depression severity. b = unstandardized regression coefficients; β = standardized re- gression coefficients; Total effect model with BDI at baseline as outcome variable: R2 = 0.12, F(5, 117) = 3.22, p < 0.01; Total effect model with symptom im- provement as outcome variable: R2 = 0.16, F(7, 115) = 3.02, p < 0.01. Significant indirect effects are shown in bold. 95% BCa CI are reported in brackets.

a An asterisk indicates a significant result of the Sobel-test. * p < 0.05. ** p < 0.01. *** p < 0.001.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

104

Since we identified several significant mediators of the association between CEA and depression severity, we additionally per- formed an exploratory mediation analysis to determine the specific contributions of the three significant mediators (BPD traits, passive-aggressive PD traits and acceptance of pleasant emotions). Fig. 1 displays the results of the comparative mediation analysis. The indirect effects of CEA on depression severity remained significant with regard to BPD traits and acceptance of pleasant emotions when taking into account all mediators simultaneously (b = 0.19, p = 0.008, BCa CI [0.08, 0,34]; b = 0.18, p = 0.005, BCa CI [0.07, 0.33]). There was no indirect effect for passive-aggressive personality traits. The standardized indirect effects provide support for strong mediating effects of BPD traits and acceptance of pleasant emotions (see Fig. 1).

4. Discussion

To the best of our knowledge, this is the first study investigating the mediating role of PD traits and emotion acceptance between CEA and MDD symptoms as well as treatment outcome in a sample of MDD inpatients. Our results yield that particularly elevated BPD traits and a lower acceptance of pleasant emotions mediated the link between CEA and self-rated depressive symptoms. Furthermore, elevated passive-aggressive PD traits mediated the association between CEA and self-rated symptom improvement. Other significant relationships between CEA and cluster A PDs (schizoid, schizotypal) were observed, but these PD traits showed no mediating effects on CEA and MDD associations. However, there were no associations between CEA and expert-ratings of depressive symptoms and between CEA and treatment duration.

Our results provide new insights into the association between CEA and depression. That is, a history of CEA may affect self-rated depressive symptoms. Changes in personality functioning and emotion acceptance may play a critical role in self-report of depressive symptoms in MDD patients with a history of CEA.

4.1. Association between CEA and depression

The finding that CEA is specifically related to self-ratings of depressive symptoms is generally in line with considerations of the hopelessness theory of depression (Abramson et al., 1989). Rose and Abramson (1992) assumed that victims of emotional mal- treatment are faced with negative beliefs more directly (e.g., “You are so stupid!”) as compared to victims of physical and sexual

Fig. 1. Comparative mediation analysis with CEA as predictor and self-rated depression severity at baseline as outcome variable. The model was controlled for sex, employment, physical, and sexual abuse. Unstandardized path coefficients are displayed. 95% BCa CI are reported in brackets. An asterisk indicates a significant result of the Sobel-test, ** p < 0.01. Total effect model: R2 = 0.12, F(5, 117) = 3.22, p < 0.01.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

105

abuse, who have to draw conclusions about the reasons of the abusive behavior by the perpetrator sometimes (cited by Alloy et al., 2006, p. 56). In line with previous findings our results confirm that CEA is more strongly associated with depressive symptoms than physical and sexual abuse (Norman et al., 2012; Spertus et al., 2003). The significant association between CEA and a low self-rated symptom improvement in our study is partially consistent with the recent finding by an international randomized trial within a large population of MDD outpatients (Williams, Debattista, Duchemin, Schatzberg, & Nemeroff, 2016): The authors demonstrated a lack of self-rated and expert-rated treatment response in survivors of child maltreatment; this observation is also corresponding to results of a recent meta-analysis (Nanni et al., 2012).

Our finding that CEA is not correlated to expert-ratings of depression, expert-rated treatment outcome, and treatment duration is generally in line with previous studies (Johnstone et al., 2009; Tunnard et al., 2014). In agreement with our finding that elevated BPD traits mediated the link between CEA and self-rated depressive symptoms, it is known that personality disorder traits influence self reports of other psychological constructs (Mattila‐Evenden, Svanborg, Gustavsson, & Åsberg, 1996). Thus, it can be speculated that correlations between self-rated CEA and self-rated depression might be in part, related to a response style influenced by BPD traits of the participants, e.g. a detailed description of psychopathological symptoms (Mattila-Evenden et al., 1996). Alternatively, expert- ratings could be biased, too. As it is known that patients with Cluster B personality dysfunction dramatically act out their symptoms, the experts may have underestimated the severity of the depressive symptoms especially in patients with BPD traits (who reported increased CEA) (Mattila-Evenden et al., 1996). Thus, it can also be speculated that missing correlations between self-rated CEA and expert-rated depression might be related to a rater-bias of the experts. Furthermore, the missing correlation between CEA and treatment duration can be explained by methodological reasons: In some cases treatment duration may also depend on the post- discharge treatment plan and in these cases short treatment duration should not necessarily reflect an overall treatment success.

Nevertheless, other studies mostly revealed a significant link between child maltreatment and a diminished treatment success by the use of expert-ratings (Nanni et al., 2012). Notably, the vast majority of studies in this field did not control for different mal- treatment subtypes and up to now specific relations between CEA and self- and expert-rated treatment outcomes are still under- studied.

4.2. Association between CEA and personality disorder traits as well as emotion acceptance

In this study we could confirm that CEA is related to cluster A PD traits, BPD traits, passive-aggressive PD traits (Carr et al., 2013; Lobbestael et al., 2010; Zanarini et al., 1997) as well as to emotion regulation deficits (Burns et al., 2010; Carvalho Fernando et al., 2014; Gratz et al., 2007). Particularly, BPD traits − in which emotion regulation impairments are a core feature – were found to be robustly associated with CEA in several previous studies (e.g., Zanarini et al., 1997) providing strong support for the validity of our results.

However, we cannot confirm the associations between CEA and cluster C PD traits on the one hand (Cohen et al., 2014) or cluster C PDs on the other (Lobbestael et al., 2010). Notably, the samples of those studies included nonpatients (Lobbestael’s sample) as well as inpatients with mixed diagnoses (i.e., anxiety disorders, substance abuse). Furthermore, the authors of these studies analyzed the effects of emotional abuse and neglect separately instead of analyzing a combined factor of emotional abuse and neglect. Both aspects may explain the divergent findings.

4.3. Personality and emotion acceptance as mediators for the association between CEA and depression

The result that passive-aggressive and particularly BPD traits mediated the link between CEA and self-rated depressive symptoms is in line with existing evidence. Findings in a sample of Korean outpatients diagnosed with depressive and anxiety disorders provided support that different child adversities were associated with specific interpersonal problems (Huh et al., 2014). However, very few studies investigated the link between child maltreatment and MDD as well as between PD and MDD simultaneously. Recently, Kounou et al. (2013) studied the mediating role of PDs in a sample of MDD patients in Togo and found emotional and sexual abuse to be independently associated with PD criteria and with the presence of MDD. Furthermore, Kounou et al. (2013) showed that PD symptoms partially mediated the relationship between child maltreatment and MDD, in accordance with our results.

Our finding that lower acceptance of pleasant emotions mediate the link between CEA and depressive symptoms is consistent with a previous study investigating a large sample of German MDD inpatients (Hopfinger et al., 2016). The authors demonstrated that emotion regulation deficits in general and emotion acceptance in particular mediate the impact of child maltreatment on depression severity. However, Hopfinger et al. (2016) reported that a range of other emotion regulation strategies (awareness, tolerance, willingness to confront) also mediate the child maltreatment depression link. Furthermore, they predicted depressive symptoms by child maltreatment in general without examining the specific contributions of maltreatment subtypes. Our result – that solely the acceptance of pleasant emotions is a mediator of the CEA to depression link – exceeds those previous findings. The result, that solely the acceptance of pleasant emotions is negatively associated with CEA while the acceptance of unpleasant emotions is unrelated to CEA is new and should be tested in future studies. This result suggests that therapeutic approaches helping to accept positive emotions may be particular beneficial for MDD patients with a history of CEA. Our results also indicate that elevated passive- aggressive PD traits mediate the link between CEA and a lower self-rated symptom improvement. This result appears to be consistent with the theoretical concept of the passive-aggressive PD. That is, patients with this PD show a passive resistance towards social and occupational performance requirements and they often feel unjustly treated or obliged (Kuhl & Kazén, 2009). These PD characteristics may lead to the subjective awareness of a lower symptom improvement as reported.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

106

4.4. Limitations

Despite several plausible results presented here some limitations should be considered. Although we assessed depressive symp- toms prospectively, the assessment of child maltreatment and PD traits as wells as emotion acceptance was cross-sectional. Thus, causality of the effects remains unclear and future studies should follow a cohort of MDD patients for a longer period of time. Furthermore, our study assessed child maltreatment retrospectively, hence self-reports can be biased by memory effects (Roy & Perry, 2004). Future studies may focus on memory bias in reports of child maltreatment in a sample of MDD patients. Although we used reliable and well validated questionnaires, the quality of our study results could be improved by the additional use of interview-based instruments. Particularly, we assessed solely dimensional PD traits using a questionnaire. Future studies could examine the mediating role of ascertained PDs in a sample of depressed patients. Furthermore, we specifically captured emotion acceptance as an emotion regulation strategy but not further emotion regulation strategies.

4.5. Conclusions and practical implications

Our study demonstrates that CEA is associated to self-rated depression severity and symptom improvement, but no association to expert-ratings or treatment duration could be shown. Elevated BPD traits and lower capacities to accept pleasant emotions appear to mediate the association between CEA and depressive symptoms. Passive-aggressive PD traits mediate the association between CEA and symptom improvement. These results suggest key conclusions relevant for primary caregivers and attending therapists as follows:

First, among child maltreatment subtypes CEA seem to be a crucial risk factor for MDD syndrome severity with important implications regarding primary prevention of MDD and psychoeducation for patients. Second, subclinical BPD traits (e.g., impair- ments in interpersonal relationships, emotion regulation deficits, and identity problems) and emotion acceptance might be starting points for attending therapists to reduce the risk of relapse in survivors of CEA. Third, to ensure a therapy success of MDD in CEA survivors attending therapists should carefully check for signs of passive resistance in their patients. Fourth, the multimodal MDD treatment following German guidelines seems to be effective for CEA survivors since the most unfavorable PD traits affected by CEA do not influence the therapy success.

Conflict of interest

None.

Role of the funding source

This study was funded by the Evangelisches Klinikum Bethel, Germany. The funding body was not involved in the study design, or in the collection, analysis, and interpretation of data. In addition, the funding body was not involved in the writing of the manuscript and in the decision to submit the manuscript for publication. There are no financial or other relationships that could be interpreted as a conflict of interest affecting this manuscript.

Acknowledgements

We thank Prof. Dr. Malte Schwinger, who kindly provided support for the multiple imputation analysis.

References

Abramson, L. Y., Metalsky, G. I., & Alloy, L. B. (1989). Hopelessness depression: A theory-based subtype of depression. Psychological Review, 96, 358–372. http://dx. doi.org/10.1037/0033-295X.96.2.358.

Aldao, A., Nolen-Hoeksema, S., & Schweizer, S. (2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical Psychology Review, 30, 217–237. http://dx.doi.org/10.1016/j.cpr.2009.11.004.

Alloy, L. B., Abramson, L. Y., Smith, J. M., Gibb, B. E., & Neeren, A. M. (2006). Role of parenting and maltreatment histories in unipolar and bipolar mood disorders: Mediation by cognitive vulnerability to depression. Clinical Child and Family Psychology Review, 9, 23–64. http://dx.doi.org/10.1007/s10567-006-0002-4.

Barlow, D. H., Allen, L. B., & Choate, M. L. (2004). Toward a unified treatment for emotional disorders. Behavior Therapy, 35, 205–230. http://dx.doi.org/10.1016/ S0005-7894(04)80036-4.

Baumann, N., Kaschel, R., & Kuhl, J. (2007). Affect sensitivity and affect regulation in dealing with positive and negative affect. Journal of Research in Personality, 41, 239–248. http://dx.doi.org/10.1016/j.jrp.2006.05.002.

Beblo, T., Scheulen, C., Fernando, S. C., Griepenstroh, J., Aschenbrenner, S., Rodewald, K., & Driessen, M. (2011). Psychometrische Analyse eines neuen Fragebogens zur Erfassung der Akzeptanz von unangenehmen und angenehmen Gefühlen (FrAGe). [Psychometric analysis of a new questionnaire assessing the acceptance of unpleasant and pleasant emotions (FrAGe)]. Zeitschrift für Psychiatrie, Psychologie und Psychotherapie, 59, 133–144. http://dx.doi.org/10.1024/1661-4747/ a000063.

Beblo, T., Fernando, S., Klocke, S., Griepenstroh, J., Aschenbrenner, S., & Driessen, M. (2012). Increased suppression of negative and positive emotions in major depression. Journal of Affective Disorders, 141, 474–479. http://dx.doi.org/10.1016/j.jad.2012.03.019.

Beblo, T., Fernando, S., Kamper, P., Griepenstroh, J., Aschenbrenner, S., Pastuszak, A., ... Driessen, M. (2013). Increased attempts to suppress negative and positive emotions in Borderline Personality Disorder. Psychiatry Research, 210, 505–509. http://dx.doi.org/10.1016/j.psychres.2013.06.036.

Beck, A. T., Steer, R. A., & Brown, G. K. (1996). BDI-II (2nd ed.). San Antonio, Tex: Pearson. Bernstein, D. P., & Fink, L. (1998). Childhood trauma questionnaire: A retrospective self-report: manualSan Antonio, TX: The Psychological Corporation. Bernstein, D. P., Stein, J. A., Newcomb, M. D., Walker, E., Pogge, D., Ahluvalia, T., ... Zule, W. (2003). Development and validation of a brief screening version of the

Childhood Trauma Questionnaire. Child Abuse & Neglect, 27, 169–190. http://dx.doi.org/10.1016/S0145-2134(02)00541-0. Brown, G. R., & Anderson, B. (1991). Psychiatric morbidity in adult inpatients with childhood histories of sexual and physical abuse. American Journal of Psychiatry,

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

107

148, 55–61. Burns, E. E., Jackson, J. L., & Harding, H. G. (2010). Child maltreatment, emotion regulation, and posttraumatic stress: The impact of emotional abuse. Journal of

Aggression, Maltreatment & Trauma, 19, 801–819. http://dx.doi.org/10.1080/10926771.2010.522947. Carr, C. P., Martins, C. M. S., Stingel, A. M., Lemgruber, V. B., & Juruena, M. F. (2013). The role of early life stress in adult psychiatric disorders: A systematic review

according to childhood trauma subtypes. The Journal of Nervous and Mental Disease, 201, 1007–1020. http://dx.doi.org/10.1097/NMD.0000000000000049. Carvalho Fernando, S., Beblo, T., Schlosser, N., Terfehr, K., Otte, C., Lowe, B., ... Wingenfeld, K. (2012). Associations of childhood trauma with hypothalamic-pituitary-

adrenal function in borderline personality disorder and major depression. Psychoneuroendocrinology, 37, 1659–1668. http://dx.doi.org/10.1016/j.psyneuen.2012. 02.012.

Carvalho Fernando, S., Beblo, T., Schlosser, N., Terfehr, K., Otte, C., Lowe, B., ... Wingenfeld, K. (2014). The impact of self-reported childhood trauma on emotion regulation in borderline personality disorder and major depression. Journal of Trauma & Dissociation, 15, 384–401. http://dx.doi.org/10.1080/15299732.2013. 863262.

Chapman, D. P., Whitfield, C. L., Felitti, V. J., Dube, S. R., Edwards, V. J., & Anda, R. F. (2004). Adverse childhood experiences and the risk of depressive disorders in adulthood. Journal of Affective Disorders, 82, 217–225. http://dx.doi.org/10.1016/j.jad.2003.12.013.

Cicchetti, D., & Toth, S. L. (2005). Child maltreatment. Annual Review of Clinical Psychology, 1, 409–438. http://dx.doi.org/10.1146/annurev.clinpsy.1.102803. 144029.

Cloitre, M., Miranda, R., Stovall, M., & Han, H. (2005). Beyond PTSD: Emotion regulation and interpersonal problems as predictors of functional impairment in survivors of childhood abuse. Behavior Therapy, 36, 119–124. http://dx.doi.org/10.1016/S0005-7894(05)80060-7.

Cohen, L. J., Tanis, T., Bhattacharjee, R., Nesci, C., Halmi, W., & Galynker, I. (2014). Are there differential relationships between different types of childhood maltreatment and different types of adult personality pathology? Psychiatry Research, 215, 192–201. http://dx.doi.org/10.1016/j.psychres.2013.10.036.

Corruble, E., Ginestet, D., & Guelfi, J. D. (1996). Comorbidity of personality disorders and unipolar major depression: A review. Journal of Affective Disorders, 37, 157–170.

Crow, T., Cross, D., Powers, A., & Bradley, B. (2014). Emotion dysregulation as a mediator between childhood emotional abuse and current depression in a low-income African-American sample. Child Abuse & Neglect, 38, 1590–1598. http://dx.doi.org/10.1016/j.chiabu.2014.05.015.

Driessen, M., Schroeder, T., Widmann, B., Schönfeld, C.-E., & Schneider, F. (2006). Childhood trauma, psychiatric disorders: and criminal behavior in prisoners in Germany: A comparative study in incarcerated women and men. Journal of Clinical Psychiatry, 67, 1486–1492.

Finzi-Dottan, R., & Karu, T. (2006). From emotional abuse in childhood to psychopathology in adulthood: A path mediated by immature defense mechanisms and self- esteem. The Journal of Nervous and Mental Disease, 194, 616–621. http://dx.doi.org/10.1097/01.nmd.0000230654.49933.23.

Gibb, B. E., Alloy, L. B., Abramson, L. Y., Rose, D. T., Whitehouse, W. G., Donovan, P., ... Tierney, S. (2001). History of childhood maltreatment, negative cognitive styles, and episodes of depression in adulthood. Cognitive Therapy and Research, 25, 425–446. http://dx.doi.org/10.1023/A:1005586519986.

Graham, J. W. (2009). Missing data analysis: making it work in the real world. Annual Review of Psychology, 60, 549–576. http://dx.doi.org/10.1146/annurev.psych. 58.110405.085530.

Gratz, K. L., & Roemer, L. (2004). Multidimensional assessment of emotion regulation and dysregulation: Development, factor structure, and initial validation of the difficulties in emotion regulation scale. Journal of Psychopathology and Behavioral Assessment, 26, 41–54.

Gratz, K. L., Bornovalova, M. A., Delany-Brumsey, A., Nick, B., & Lejuez, C. W. (2007). A laboratory-based study of the relationship between childhood abuse and experiential avoidance among inner-city substance users: The role of emotional nonacceptance. Behavior Therapy, 38, 256–268.

Gross, J. J., & Muñoz, R. F. (1995). Emotion regulation and mental health. Clinical Psychology: Science and Practice, 2, 151–164. http://dx.doi.org/10.1111/j.1468- 2850.1995.tb00036.x.

Harkness, K. L., Bagby, R. M., & Kennedy, S. H. (2012). Childhood maltreatment and differential treatment response and recurrence in adult major depressive disorder. Journal of Consulting and Clinical Psychology, 80, 342–353. http://dx.doi.org/10.1037/a0027665.

Hautzinger, M., Keller, F., & Kühner, C. (2006). Beck depressions-inventar (BDI-II). Frankfurt: Harcourt Test Services. Hiller, W., Zaudig, M., & Mombour, W. (1995). Internationale Diagnosen Checklisten für ICD-10 und DSM-IV. Manual, Bern: Hans Huber. Hopfinger, L., Berking, M., Bockting, C. L. H., & Ebert, D. D. (2016). Emotion regulation mediates the effect of childhood trauma on depression. Journal of Affective

Disorders, 198, 189–197. http://dx.doi.org/10.1016/j.jad.2016.03.050. Huh, H. J., Kim, S.-Y., Yu, J. J., & Chae, J.-H. (2014). Childhood trauma and adult interpersonal relationship problems in patients with depression and anxiety

disorders. Annals of General Psychiyatry, 13(26), 1–13. http://dx.doi.org/10.1186/s12991-014-0026-y. Iacoviello, B. M., Alloy, L. B., Abramson, L. Y., Whitehouse, W. G., & Hogan, M. E. (2007). The role of cluster B and C personality disturbance in the course of

depression: A prospective study. Journal of Personality Disorders, 21, 371–383. http://dx.doi.org/10.1521/pedi.2007.21.4.371. Infurna, M. R., Reichl, C., Parzer, P., Schimmenti, A., Bifulco, A., & Kaess, M. (2016). Associations between depression and specific childhood experiences of abuse and

neglect: A meta-analysis. Journal of Affective Disorders, 190, 47–55. http://dx.doi.org/10.1016/j.jad.2015.09.006. Johnstone, J. M., Luty, S. E., Carter, J. D., Mulder, R. T., Frampton, C. M. A., & Joyce, P. R. (2009). Childhood neglect and abuse as predictors of antidepressant

response in adult depression. Depression and Anxiety, 26, 711–717. http://dx.doi.org/10.1002/da.20590. Kim, J., & Cicchetti, D. (2010). Longitudinal pathways linking child maltreatment, emotion regulation, peer relations, and psychopathology. Journal of Child Psychology

and Psychiatry, and Allied Disciplines, 51, 706–716. http://dx.doi.org/10.1111/j.1469-7610.2009.02202.x. Kounou, K. B., Bui, E., Dassa, K. S., Hinton, D., Fischer, L., Djassoa, G., ... Schmitt, L. (2013). Childhood trauma, personality disorders symptoms and current major

depressive disorder in Togo. Social Psychiatry and Psychiatric Epidemiology, 48, 1095–1103. http://dx.doi.org/10.1007/s00127-012-0634-2. Kuhl, J., & Kazén, M. (2009). Persönlichkeits-Stil- und Störungs-Inventar (PSSI) (2., überarbeitete und neunormierte Auflage) [Personality Styles and Disorder Inventory

(PSDI)]. Göttingen: Hogrefe. Liu, R. T., Alloy, L. B., Abramson, L. Y., Iacoviello, B. M., & Whitehouse, W. G. (2009). Emotional maltreatment and depression: Prospective prediction of depressive

episodes. Depression and Anxiety, 26, 174–181. http://dx.doi.org/10.1002/da.20545. Lobbestael, J., Arntz, A., & Bernstein, D. P. (2010). Disentangling the relationship between different types of childhood maltreatment and personality disorders. Journal

of Personality Disorders, 24, 285–295. http://dx.doi.org/10.1521/pedi.2010.24.3.285. Martins, C. M. S., Baes, C. V. W., Tofoli, S. M.d. C., & Juruena, M. F. (2014). Emotional abuse in childhood is a differential factor for the development of depression in

adults. Journal of Nervous and Mental Disease, 202, 774–782. http://dx.doi.org/10.1097/NMD.0000000000000202. Mattila‐Evenden, M., Svanborg, P., Gustavsson, P., & Åsberg, M. (1996). Determinants of self‐rating and expert rating concordance in psychiatric out-patients, using

the affective subscales of the CPRS. Acta Psychiatrica Scandinavica, 94, 386–396. Maughan, A., & Cicchetti, D. (2002). Impact of child maltreatment and interadult violence on childrenśs emotion regulation abilities and socioemotional adjustment.

Child Development, 73, 1525–1542. McLaughlin, K. A., Green, J. G., Gruber, M. J., Sampson, N. A., Zaslavsky, A. M., & Kessler, R. C. (2010). Childhood adversities and adult psychiatric disorders in the

national comorbidity survey replication II: Associations with persistence of DSM-IV disorders. Archives of General Psychiatry, 67, 124–132. Montgomery, S. A., & Asberg, M. (1979). A new depression scale designed to be sensitive to change. The British Journal of Psychiatry, 134, 382–389. http://dx.doi.org/

10.1192/bjp.134.4.382. Nanni, V., Uher, R., & Danese, A. (2012). Childhood maltreatment predicts unfavorable course of illness and treatment outcome in depression: a meta-analysis.

American Journal of Psychiatry, 169, 141–151. http://dx.doi.org/10.1176/appi.ajp.2011.11020335. Neumann, N., & Schulte, R. (1989). MADR-Skala zur psychometrischen Beurteilung depressiver Symptome (MADRS). Perimed: Erlangen. Newton-Howes, G., Tyrer, P., & Johnson, T. (2006). Personality disorder and the outcome of depression: Meta-analysis of published studies. The British Journal of

Psychiatry: The Journal of Mental Science, 188, 13–20. http://dx.doi.org/10.1192/bjp.188.1.13. Norman, R. E., Byambaa, M., De, R., Butchart, A., Scott, J., Vos, T., & Tomlinson, M. (2012). The long-term health consequences of child physical abuse, emotional

abuse, and neglect: A systematic review and meta-analysis. PLoS Medicine, 9, e1001349. http://dx.doi.org/10.1371/journal.pmed.1001349. O'Mahen, H. A., Karl, A., Moberly, N., & Fedock, G. (2015). The association between childhood maltreatment and emotion regulation: Two different mechanisms

contributing to depression? Journal of Affective Disorders, 174, 287–295. http://dx.doi.org/10.1016/j.jad.2014.11.028.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

108

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments & Computers, 36, 717–731. http://dx.doi.org/10.3758/BF03206553.

Roy, C. A., & Perry, J. C. (2004). Instruments for the assessment of childhood trauma in adults. The Journal of Nervous and Mental Disease, 192, 343–351. http://dx.doi. org/10.1097/01.nmd.0000126701.23121.fa.

Saarni, C. (1999). The development of emotional competence. New York: Guilford Press. Sato, T., Uehara, T., Narita, T., Sakado, K., & Fujii, Y. (2000). Parental bonding and personality in relation to a lifetime history of depression. Psychiatry and Clinical

Neurosciences, 54, 121–130. Schmidtke, A., Fleckenstein, P., Moises, W., & Beckmann, H. (1988). Studies of the reliability and validity of the German version of the Montgomery-Asberg Depression

Rating Scale (MADRS). Schweizer Archiv fur Neurologie und Psychiatrie (Zurich, Switzerland: 1985), 139, 51–65. Silverman, A. B., Reinherz, H. Z., & Giaconia, R. M. (1996). The long-term sequelae of child and adolescent abuse: A longitudinal community study. Child

Abuse & Neglect, 20, 709–723. Spertus, I. L., Yehuda, R., Wong, C. M., Halligan, S., & Seremetis, S. V. (2003). Childhood emotional abuse and neglect as predictors of psychological and physical

symptoms in women presenting to a primary care practice. Child Abuse & Neglect, 27, 1247–1258. http://dx.doi.org/10.1016/j.chiabu.2003.05.001. Suija, K., Aluoja, A., Kalda, R., & Maaroos, H.-I. (2011). Factors associated with recurrent depression: a prospective study in family practice. Family Practice, 28, 22–28.

http://dx.doi.org/10.1093/fampra/cmq076. Taillieu, T. L., Brownridge, D. A., Sareen, J., & Afifi, T. O. (2016). Childhood emotional maltreatment and mental disorders: Results from a nationally representative

adult sample from the United States. Child Abuse & Neglect, 59, 1–12. Tunnard, C., Rane, L. J., Wooderson, S. C., Markopoulou, K., Poon, L., Fekadu, A., ... Cleare, A. J. (2014). The impact of childhood adversity on suicidality and clinical

course in treatment-resistant depression. Journal of Affective Disorders, 152–154, 122–130. http://dx.doi.org/10.1016/j.jad.2013.06.037. Tyrka, A. R., Wyche, M. C., Kelly, M. M., Price, L. H., & Carpenter, L. L. (2009). Childhood maltreatment and adult personality disorder symptoms: Influence of

maltreatment type. Psychiatry Research, 165, 281–287. http://dx.doi.org/10.1016/j.psychres.2007.10.017. Vink, G., Frank, L. E., Pannekoek, J., & van Buuren, S. (2014). Predictive mean matching imputation of semicontinuous variables. Statistica Neerlandica, 68, 61–90.

http://dx.doi.org/10.1111/stan.12023. Widom, C. S., DuMont, K., & Czaja, S. J. (2007). A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grown up.

Archives of General Psychiatry, 64, 49–56. http://dx.doi.org/10.1001/archpsyc.64.1.49. Williams, L. M., Debattista, C., Duchemin, A. M., Schatzberg, A. F., & Nemeroff, C. B. (2016). Childhood trauma predicts antidepressant response in adults with major

depression: data from the randomized international study to predict optimized treatment for depression. Translational Psychiatry, 6, e799. http://dx.doi.org/10. 1038/tp.2016.61.

Wingenfeld, K., Spitzer, C., Mensebach, C., Grabe, H. J., Hill, A., Gast, U., et al. (2010). Die deutsche Version des Childhood Trauma Questionnaire (CTQ): Erste Befunde zu den psychometrischen Kennwerten. [The German version of the Childhood Trauma Questionnaire (CTQ): Preliminary psychometric properties]. Psychotherapie, Psychosomatik, medizinische Psychologie, 60, 442–450. http://dx.doi.org/10.1055/s-0030-1247564.

World Health Organization (1992). The ICD-10 classification of mental and behavioural disorders: Clinical descriptions and diagnostic guidelines. Geneva: World Health Organization.

Zanarini, M. C., Williams, A. A., Lewis, R. E., Reich, R. B., Vera, S. C., Marino, M. F., ... Frankenburg, F. R. (1997). Reported pathological childhood experiences associated with the development of borderline personality disorder. The American Journal of Psychiatry, 154, 1101–1106. http://dx.doi.org/10.1176/ajp.154.8. 1101.

Zhao, X., Lynch, J. G., & Chen, Q. (2010). Reconsidering Baron and Kenny: Myths and truths about mediation analysis. Journal of Consumer Research, 37, 197–206. http://dx.doi.org/10.1086/651257.

P. Schulz et al. Child Abuse & Neglect 72 (2017) 98–109

109