Impact of Maternal Depression and Adult Attachment Style

Arch Womens Ment Health (2016) 19:927–935 DOI 10.1007/s00737-016-0648-y ORIGINAL ARTICLE Postpartum bonding: the impact of maternal depression and adult attachment style N. Nonnenmacher 1 & D. Noe 1 & J. C. Ehrenthal 2 & C. Reck 3 Received: 2 February 2016 / Accepted: 2 June 2016 / Published online: 14 June 2016 # Springer-Verlag Wien 2016 Abstract Maternal depression poses a risk for the developing mother-infant relationship. Similarly, maternal insecure attachment styles may limit the ability to adequately connect with the newborn during the postpartum period. The aim of this study was to investigate the effect of maternal depression and insecure attachment (insecure and dual/disorganized) on maternal bonding in a sample of n = 34 women with depression according to DSM-IV and n = 59 healthy women. Maternal depression was assessed 3 to 4 months postpartum with the Structured Clinical Interview for DSM-IV Axis I Disorders (SCIDI), bonding with the Postpartum Bonding Questionnaire, and maternal attachment style with the Attachment Style Interview.

Women with current and lifetime depression as well as women with dual/disorganized attachment style reported lower bonding. Explorative analysis revealed that d ep r es s i on p ar t i al l y m e d i a t e d t he l i n k be t w e e n dual/disorganized attachment style and bonding with a medium-sized mediation effect. The combination of maternal depression and dual/disorganized attachment style may pose a special risk constellation for the developing mother–infant bond that should be addressed in prevention and early intervention programs. * N. Nonnenmacher Nora.Nonnenmacher@med.uni-heidelberg.de 1 Department of General Psychiatry, Heidelberg University Hospital, Voßstraße 2, 69115 Heidelberg, Germany 2 Department of General Internal Medicine and Psychosomatics, Heidelberg University Hospital, Thibautstr. 1, 69115 Heidelberg, Germany 3 Department of Psychology, Ludwig-Maximilians University, Leopoldstraße 13, 80802 Munich, Germany Keywords Maternal depression . Maternal adult attachment style . Maternal dual/disorganized attachment style . Bonding . Mother–child relationship Introduction The postpartum period is a challenging time for mothers as the care for an infant requires constant attention to the infant’s needs. Bonding, described as the emotional tie of a mother to her newborn (Klaus et al. 1995) that develops gradually after birth, ensures protection and nurture of the newborn and is highly relevant for the establishing mother–infant relationship (Brockington 2004; Klaus and Kennell 1982). Expressions of bonding in maternal postpartum behavior include proximity seeking, touch, contact, gaze, baby talk, positive expression, cuddling, smiling, and adaptation to cues expressed by the infant (Reck et al. 2006). On the other hand, bonding is associated with cognitive-emotional variables, including positive feelings, emotional warmth, and affection (Klaus et al. 1995). In animal studies, this maternal behavior was related to diverse positive effects for the offspring, such as a longer period of brain development, engendering diverse morphology and physiology of the brain, more adaptive stress reactivity, and higher levels of social affiliation (Meaney 2001). Studies in humans resulted in similar findings. For example, adverse parental bonding was associated with higher autonomic stress response to distress stimuli in children (Dalsant et al. 2015). Furthermore, bonding is influenced by situational factors as well as by personality variables of the mother. A powerful situational predictor is maternal mental health. There is robust evidence that the probability for impaired bonding is higher in women suffering from depression (Dubber et al. 2015; Moehler et al. 2006; Reck et al. 2006; Tietz et al. 2014). Here, not only current maternal 928 psychopathology but also remitted and lifetime diagnoses have long-term influences on maternal variables such as self-confidence concerning the maternal role (Reck et al. 2012) and parenting behavior (Thomas et al. 2015). Bonding capacities are also related to developmentally shaped relationship expectancies, as for example described by attachment theory. Mental representations of early and repeated interactions with caregivers (Bowlby 1977) influence adult attachment. These representations are organized in Binternal working models^ and comprise expectations and beliefs about the self and attachment figures (Bowlby 1982), which are then internalized and generalized over the course of development. These internal working models are relatively stable across the lifespan (Grossmann and Grossmann 1991) and influence attitudes towards closeness, attachment behavior, quality of romantic relationships, and social support in adulthood (Bowlby 1977; Grossmann and Grossmann 2004; Main et al. 1985). Even more relevant for our study, attachment representations affect the way caregivers interact with and feel about their children (Bowlby 1982). While self-report questionnaires mainly assess attachment with regard to two continuous dimensions of insecurity—anxiety and avoidance (where anxious individuals are characterized as ambivalent, desiring proximity but concurrently fearing rejection and avoidant individuals as seeking independence and autonomy, e.g., Paetzold et al. 2015)—a particular strength of related interview measures is that they provide further information about disintegrated or contradictory internal working models of attachment. Focusing on adult representations of own childhood experiences with the family of origin, the Adult Attachment Interview (George et al. 1985; Main et al. 1985) provides a classification corresponding to child’s disorganization labeled Bunresolved^ for individuals with an unresolved state of mind and disorientation or disorganization when discussing potentially traumatic events, such as loss or abuse (Main and Goldwyn 1984). Due to the evaluation of narrative coherence, the AAI is particularly distinct from other attachment measures when assessing unconscious aspects of attachment and internal representation of primary caregivers, as well as unresolved states of minds. When investigating other aspects of attachment, as for example current relationships or contradictory behaviors and affects in adulthood, it might be suitable to make use of other attachment measures. The Attachment Style Interview (ASI) (Bifulco et al. 2002) is one kind of interview which focuses on attachment in current adult relationships, with regard to the quality of close, supportive relationships with the partner (if present) and other close confidants, and was therefore used in this study. We assume that the way a person feels about current close relationships to other adults might also reflect the way the relationship to the baby is formed and experienced. Impact of Maternal Depression and Adult Attachment Style

Like the AAI, the ASI also describes an attachment strategy which is incoherent with regard to associated behavioral aspects, such as N. Nonnenmacher et al. contradictory patterns of approach and avoidance or fear and anger of others. As it constitutes a combination of anxious and avoidant attachment styles, this category is labeled dual/disorganized. Dual/disorganized attachment is directly associated to psychopathology (Schimmenti and Bifulco 2008), interpersonal problems in adolescence (Schimmenti and Bifulco 2008), and discomfort in romantic relationships in adulthood (Paetzold et al. 2015) but also to maladaptive parenting (Busch et al. 2008; Goldberg et al. 2003). So far, organized insecure attachment strategies were closely linked to depression per se, its duration and severity (Conradi and de Jonge 2009; McMahon et al. 2008), and the occurrence of postpartum depression (Bifulco et al. 2004; Ikeda et al. 2014). Moreover, an association of insecure adult attachment with lower quality of relationship to the baby (Rholes et al. 1995; Wilkinson and Mulcahy 2010) and less adequate parenting behavior (Bifulco et al. 2009; Haltigan et al. 2014) was reported.

The role of dual/disorganized attachment for the maternal subjective relationship to the baby has not yet been examined. In sum, bonding, maternal depression, and maternal adult attachment style are all related, but the exact relationship between the three variables remains unclear. To the best of the authors’ knowledge, this is the first study to investigate the role of maternal depression and attachment style for the subjective bonding relationship mothers develop towards their newborn. In line with previous studies, we assume that mothers with depression report lower bonding than nondepressed mothers. Furthermore, we assume that depressed women show an insecure or dual/disorganized attachment style more often. Due to the link between adult attachment and quality of close relationships, we hypothesize that maternal insecure or dual/disorganized attachment style negatively influences the way a mother feels about the relationship to her child, operationalized via self-reported maternal bonding. As dual/disorganized attachment differs from insecure attachment in the inconsistency of attachment strategy and maladaptive overall functioning, we focused on this attachment style in explorative analyses. Impact of Maternal Depression and Adult Attachment Style

More specifically, we were interested if maternal depression mediates the expected relationship between dual/disorganized attachment style and bonding. Method Procedure and participants Participating mothers (N = 107) gave birth between May 2009 and December 2012 and were recruited in local maternity hospitals (n = 104) and the mother–infant unit (n = 3) of the Department of General Psychiatry, Heidelberg University Hospital. Eligible mothers were screened for depressive syndromes according to DSM-IV and subsequently interviewed Postpartum bonding: maternal depression and adult attachment with the Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) (Wittchen et al. 1997) within the first 3 m o n t h s p o s t p a r t u m . T h e P o s tp a r t u m B o n d i n g Questionnaire-16 (PBQ-16) (German version, Reck et al. 2006) and the ASI (Bifulco et al. 2002) were completed at an assessment point 3 to 4 months postpartum. The interviews were videotaped and subsequently rated by trained psychologists or master students, which were blind to the mental health status of the mothers. The initial sample size of N = 107 mothers was reduced to n = 93 mothers due to missing interviews (n = 3) and the exclusion of women without depression but with other psychiatric diagnoses (n = 11). According to their mental health status, mothers were assigned to the healthy (n = 59) or depressed (n = 34) group. The healthy group consisted of mothers without current or lifetime psychiatric disorders or psychotherapeutic experience. Mothers in the depressed group suffered from major (n = 29) or minor (n = 5) depression with 35.3 % comorbid anxiety disorders (n = 12), 14.7 % eating disorders (n = 5), and 5.9 % posttraumatic stress disorders (n = 2). A total of n = 15 mothers (44.1 %) were in an acute depressive episode, whereas n = 19 mothers (55.9 %) had a lifetime diagnosis of depression. All participating mothers needed to be at least 18 years old and have adequate knowledge of the German language. The study protocol was approved by the Ethics Committee of the Medical Faculty, Heidelberg University Hospital. Written informed consent was obtained after the study procedures had been fully explained. Sample characteristics Maternal age ranged from 24 to 44 years with a mean of M = 33.95 years (SD = 4.52) within the overall sample (n = 93). On average, women had M = 1.53 children (SD = 0.58) ranging from 1 to 4.

Infant’s age ranged from 12.7 to 17.9 weeks with a mean age of M = 14.59 (SD = 1.41) weeks. With regard to the total sample, n = 49 (52.7 %) children were female and n = 44 (47.3 %) were male. The distribution of maternal education level in the overall sample was as follows: n = 22 (23.7 %) completed intermediate secondary education, n = 12 (12.9 %) qualified for university admission, and n = 59 (63.4 %) held a university degree. Detailed sociodemographic information can be found in Table 1. Measures Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) Maternal mental health was assessed with the German version of the SCID-I (Wittchen et al. 1997) by trained, experienced psychologists. The SCID-I is a widely used 929 semistructured interview for the diagnosis of selected axis I disorders. Impact of Maternal Depression and Adult Attachment Style

The authors of the German version evaluated the psychometric properties regarding reliability, applicability, and efficacy as satisfying. Postpartum Bonding Questionnaire-16 (PBQ-16) The PBQ was originally developed as a self-report measure for the screening of bonding disorders (Brockington et al. 2001). We used the abridged 16-item German version (PBQ16, Reck et al. 2006). Items such as BI feel close to my baby^ or BMy baby irritates me^ are rated on a 6-point Likert scale from 0 (Bnever^) to 5 (Balways^). High values on the PBQ-16 indicate lower levels of bonding. In our sample, internal consistency was good (Cronbach’s α = 0.89), and the sum scores ranged between 0 and 40 (M = 9.16, SD = 7.94). Attachment Style Interview (ASI) This semistructured interview (Bifulco et al. 2002) focuses on current behavior and attitudes to assess adult attachment, including secure, anxious (enmeshed or fearful), and avoidant (angry-dismissive or withdrawn) attachment styles. In the first part of the interview, participants describe the quality of their close, supportive relationship with the partner (if present) and other close confidants. Scales assessed are Bconfiding, active emotional support, positive and negative interaction, and felt attachment^ for each of the three support figures. The quality of each relationship can be coded and added to an overall Bability to make and maintain relationships.^ The second part of the interview assesses personal attitudes regarding attachment and can be assigned to either an anxious (e.g., fear of separation, desire for company, anger) or avoidant type (e.g., mistrust, constraints on closeness, high self-reliance, fear of rejection).

The ratings on these scales are used to categorize the type of attachment style (secure, high on anxiety: enmeshed, fearful or high on avoidance: angry-dismissive, withdrawn) and the degree of insecurity (markedly, moderately, or mildly insecure). Dual/ disorganized attachment style is characterized by a Bdouble^ classification of style, occurring when no clear attachment pattern can be recognized. In this case, two insecure attachment styles were evident across relationships (Bifulco et al. 2002). This attachment style is described to be contradictory in the approach and avoidance of others as well as in fear and anger to others (Bifulco and Thomas 2012). Furthermore, this classification is associated to traumatic experiences (Schimmenti and Bifulco 2008). In order to assure the quality of the ratings, n = 14 (14 %) randomly selected tapes were double coded by two independent study coders. Interrater reliability was computed using Cohen’s k (Cohen 1960) and was k = 0.82 for the double coded tapes. 930 N. Nonnenmacher et al. Table 1 Sample characteristics Group Healthy (n = 59) Maternal age (years) Number of children Child age (weeks) Infant gender male (%) Education Intermediate secondary education University admission University degree a t test b χ2 test Depressed (n = 34) M SD Range M SD Range P 33.75 1.59 14.66 4.52 0.83 1.51 24–44 1–4 12.9– 17.9 34.32 1.41 14.45 4.56 0.61 1.22 27–43 1–3 12.7– 17.4 0.56a 0.23a 0.53a Absolute frequency (%) 29 (49.2 %) 15 (44.1 %) 11 (18.7 %) 11 (32.3 %) 8 (13.6 %) 40 (67.8 %) 4 (11.8 %) 19 (55.9 %) Statistical analyses We used the Statistical Package for Social Sciences (IBMTM SPSS® v. 22) for all analyses. Impact of Maternal Depression and Adult Attachment Style

Maternal attachment was operationalized by the use of a variable with three levels: securely attached, insecurely attached (anxious and avoidant styles), and dual/disorganized attachment style. Due to small subsample sizes, women with anxious and avoidant insecure attachment styles were allocated to the category Binsecure attachment.^ To ensure comparability between the depressed and healthy group, differences concerning maternal age, child age, number of children, maternal education, and infant gender were explored (using t tests and χ2 tests) before carrying out the main analyses (see Table 1). For further determination of specific characteristics of maternal depression and to include possible situational influences on maternal bonding capacity, we operationalized a new variable called Bmaternal depression^ with regard to three levels of severity: without history of SCID-I diagnoses (never), lifetime diagnoses (lifetime) of depression, and current diagnoses during the postpartum period (current). The assumption that maternal depression is rank scaled regarding the impact on the mother–child relationship is supported by recent findings, demonstrating that maternal bonding was differently affected by the course of depression: While the lowest bonding was reported during an acute episode, it significantly increased when symptoms of depression remitted but still differed significantly from the reported bonding of healthy mothers (Reck et al. 2015).

Subsequently, we used Kruskal-Wallis test to analyze the distribution of maternal depression concerning P 0.40b 0.47b attachment style. Two Mann-Whitney U tests were performed as post hoc tests to examine subgroup differences of maternal depression. Assumed differences between the depressed and healthy group in their reported bonding were tested by ANOVA with three levels (never, lifetime, current). Tukey post hoc tests were computed to analyze subgroup differences of maternal depression (never, lifetime, current). A second ANOVA analyzed the influence of maternal attachment style (secure, insecure, dual/disorganized) on reported bonding. Tukey post hoc tests were computed to test for subgroup differences of maternal attachment style. In explorative analyses, maternal depression was tested as a mediator of the relationship between dual/disorganized attachment style (independent variable) and bonding (outcome variable). We carried out the standard procedure according to Preacher and Hayes (2004) using the SPSS®-Macro BPROCESS^ (v. 2.15) (Hayes 2013). The model estimates the direct effect of maternal dual/disorganized attachment style on maternal depression, the indirect effect of maternal dual/disorganized attachment style on bonding mediated by maternal depression as well as the direct effect of maternal dual/disorganized attachment style on bonding (see Fig. 1). The standard errors and confidence intervals of the indirect (mediated) effects are bootstrapped and bias-corrected (N = 1000 samples). Impact of Maternal Depression and Adult Attachment Style

Empirical p values are two-tailed. Before carrying out the mediation analysis, Pearson’s correlations were conducted to screen for associations between dual/disorganized attachment style, maternal depression, and bonding. Postpartum bonding: maternal depression and adult attachment Results Analyses regarding comparability of sociodemographic variables between depressed and healthy group revealed no significant differences (see Table 1). 931 p < 0.001). Differences in self-reported bonding between women with current and women with lifetime diagnoses failed to reach a conventional significance level (ΔM = 4.60, SE = 2.19, p = 0.095). Maternal attachment style and bonding Distribution of attachment styles The majority of healthy women were coded as having a secure attachment style (n = 40, 67.8 %). A total of n = 15 (25.4 %) women were insecurely attached, and n = 4 (6.8 %) showed dual/disorganized attachment style. Of the insecure attachment styles, anxious attachment was the most common (n = 12, 20.3 %). Only … Impact of Maternal Depression and Adult Attachment Style