Research Proposal

ORIGINAL ARTICLE

Postpartum bonding: the role of perinatal depression, anxiety and maternal–fetal bonding during pregnancy

S. Dubber & C. Reck & M. Müller & S. Gawlik

Received: 13 February 2014 /Accepted: 20 July 2014 /Published online: 5 August 2014 # Springer-Verlag Wien 2014

Abstract Adverse effects of perinatal depression on the mother–child interaction are well documented; however, the influence of maternal–fetal bonding during pregnancy on postpartum bonding has not been clearly identified. The sub- ject of this study was to investigate prospectively the influence of maternal–fetal bonding and perinatal symptoms of anxiety and depression on postpartum mother–infant bonding. Data from 80 women were analyzed for associations of symptoms of depression and anxiety as well as maternal bonding during pregnancy to maternal bonding in the postpartum period using the Edinburgh Postnatal Depression Scale (EPDS), the State– Trait Anxiety Inventory (STAI), the Pregnancy Related Anxiety Questionnaire (PRAQ-R), the Maternal–Fetal Attachment Scale (MFAS) and the Postpartum Bonding Questionnaire (PBQ-16). Maternal education, MFAS, PRAQ-R, EPDS and STAI-T significantly correlated with the PBQ-16. In the final regression model, MFAS and EPDS postpartum remained significant predictors of postpartum bonding and explained 20.8 % of the variance. The results support the hypothesized negative relationship between ma- ternal–fetal bonding and postpartum maternal bonding impair- ment as well as the role of postpartum depressive symptoms. Early identification of bonding impairment during pregnancy and postpartum depression in mothers plays an important role

for the prevention of potential bonding impairment in the early postpartum period.

Keywords Postpartum depression . Anxiety . Pregnancy .

Maternal–fetal attachment . Postpartum bonding

Introduction

Anxiety and depression in women during pregnancy and the postpartum period have been matters of great scientific inter- est (Bennett et al. 2004; Figueiredo and Costa 2009; Matthey et al. 2003; Tronick and Reck 2009). Pregnancy and puerpe- rium are times of particular vulnerability, therefore many women exhibit symptoms of emotional distress during this period. While maternal prenatal and postnatal depression and its effects on both the mother’s health and her child’s devel- opment is a well-recognized health issue (Grigoriadis et al. 2013), there are limited studies on the prospective value of anxiety and depressive symptoms and their link to the emo- tional involvement with the fetus during pregnancy (Figueiredo and Costa 2009).

According to a systematic literature review, antenatal de- pression affects approximately 12 % of women with the highest prevalence in the second and third trimesters of preg- nancy (Bennett et al. 2004). There is an overwhelming amount of data suggesting an association between untreated depres- sion or anxiety during pregnancy and unfavorable outcomes for both mother and fetus such as lower birth weight, preterm birth as well as behavioral problems of the child postpartum (Bonari et al. 2004; Grigoriadis et al. 2013; Grote et al. 2010; Marcus 2009). Postpartum Depression, according to DSM-IV criteria, is reported to be prevalent in 6.1 % of women in Germany (Reck et al. 2008) and can negatively influence mother–child interactions (Moehler et al. 2007; Reck et al. 2004).

S. Dubber (*) Center for Psychosocial Medicine, Heidelberg University Hospital, Vossstr. 2, 69115 Heidelberg, Germany e-mail: [email protected]

C. Reck: M. Müller Department of Psychology, Ludwig-Maximilians-Universität München, Leopoldstr. 13, 80802 Munich, Germany

S. Gawlik Department of Obstetrics and Gynecology, Heidelberg University Hospital, INF 440, 69120 Heidelberg, Germany

Arch Womens Ment Health (2015) 18:187–195 DOI 10.1007/s00737-014-0445-4

Regarding anxiety, empirical studies have reported a high prevalence of anxiety symptoms in more than 25 % of preg- nant women (Britton 2011; Ross and McLean 2006). Anxiety levels seemed to be higher during pregnancy when compared to the postpartum period (Andersson et al. 2006; Buist et al. 2003; Figueiredo and Costa 2009). Reck et al. (2008) reported a prevalence rate of 11.1 % for postpartum anxiety disorders which is higher than the rate of 6.1 % found for postpartum depressive disorders.

Maternal bonding1 was described by Cranley (1981) as a qualitative change in the relationship of a mother to her infant, which already emerges during pregnancy. This change takes place due to the physically developing fetus and psychological adjustments accompanying the upcoming mothership. Cranley (1981) defined maternal–fetal bonding as “the extent to which women engage in behaviors that represent an affiliation and interaction with their unborn child,” and developed the Maternal–Fetal Attachment Scale (MFAS) to measure the construct. Brockington (2004) argues that the development of the relationship between a caregiver and an infant is the most significant process after birth. Biologically, maternal bonding has the function of securing the nurturing and protection, and thus, the survival of the child (Carter and Keverne 2002).

The development of maternal–fetal bonding is crucial because it can positively influence maternal health prac- tices during pregnancy and thus, neonatal outcome (Alhusen et al. 2012). Furthermore, women with higher maternal–fetal bonding are reported to show more secure postpartum attachment styles and their children showed better early development compared to women with lower maternal–fetal bonding and less secure attachment styles (Alhusen et al. 2013). Moreover, for the child, the attach- ment bonds developed in early childhood may constitute the foundation of the individual attachment style well into adulthood (Waters et al. 2003).

Considering mental health during pregnancy Alhusen (2008) reports in her review that both depression and anxiety, as well as substance abuse during pregnancy have a negative association with maternal–fetal bonding. In the postpartum, Reck et al. (2004) states in an overview that it is well known that affective disorders like depression can negatively influ- ence mother–infant interactions. Depressed mothers are often described as being passive, withdrawn, unresponsive or intru- sive (Field 1998; Field 2010). They also express more nega- tive feelings towards their children than non-depressed mothers (Reck et al. 2004). Current literature suggests that children of mothers with mental disorders are at risk of later

psychopathology and poor functioning in a range of develop- mental domains (Goodman and Gotlib 1999; Moehler et al. 2007; Wan and Green 2009). Moreover, the meta-analysis conducted by Atkinson et al. (2000) showed a significant impact of maternal depression on later child attachment security.

The adverse effects of anxiety on the mother’s emotional involvement with the infant are not as obvious in the literature as the impact of depression (Figueiredo and Costa 2009). Britton (2011) also documented in a non-clinical sample of 296 women 1 month after delivery that ratings of overall temperamental difficulty were independently associated with symptoms of anxiety and depression in the early postpartum period. Feldman et al. (1997) stated that increased anxiety pre- and postnatally seem to interfere with the mother’s ability to bond and interact sensitively with the child.

However, little is known about the effects of anxiety and depression on the mother's emotional involvement with the fetus during pregnancy and its predictive value for postpartum bonding. The bond of a mother to her fetus developing during pregnancy is assumed to be an important precursor of post- partum bonding. In her work, Raphael-Leff describes that even before birth, parents begin to ascribe characteristics to their baby, partly based on discernible fetal rhythms and responses, partly based on their own fantasies and imagina- tions (Raphael-Leff 2001). Early identifications or unresolved conflicts with the women's own parents may surface at this time of transition from being the child of her mother to being the mother of her own child. As a result, a damaging internal image may hereby color the experience of the growing fetus as it may influence early parenthood. Raphael-Leff (2001) states that although the sex and the personality of each baby inevi- tably influences caregiving, antenatal conceptualization of the baby generally primes the parent's orientation.

Only few quantitative studies so far have investigated this presumption. The study by Müller (1996) reported a moderate positive correlation between prenatal, assessed with the Prenatal Attachment Inventory (PAI; Müller 1993) and post- natal bonding, assessed with the Maternal Attachment Inventory (MAI; Müller 1994) in a US sample consisting of 128 women recruited from childbirth education classes. Van Bussel et al. (2010) drew the same conclusion by finding a comparable correlation in a Dutch sample with 263 partici- pating women from a routine clinic sample who completed three different quantitative measures for mother-to-infant bonding. Van Bussel et al. (2010) used the Maternal Antenatal Attachment Scale (MAAS; Condon 1993) to mea- sure prenatal maternal bonding; to assess postpartum maternal bonding the Maternal Postpartum Attachment Scale (MPAS) by Condon and Corkindale (1998), the Postpartum Bonding Questionnaire (PBQ; Brockington et al. 2001) and the Mother- to-Infant Bonding Scale (MIBS; Taylor et al. 2005) were admin- istered. Furthermore, Siddiqui and Hägglöf (2000) investigated

1 As “attachment” and “bonding” are often used synonymously we find it necessary to distinguish these terms. In this study, the term bonding refers to the feelings the mother has towards her fetus/infant, whereas the term attachment stands for the relationship a child has developed to its mother.

188 S. Dubber et al.

in a sample of 100 randomly recruited women whether maternal prenatal attachment assessed with the PAI was associated with the mother–infant relationship postpartum. They revealed that maternal prenatal bonding towards the fetus functions as a strong predictor of early mother–infant relationship. However, this study did not consider potential confounders as maternal stress and anxiety pre- and postnatally.

Therefore, in the current study we aimed at identifying the influence of maternal pre- and postnatal symptoms of depres- sion and anxiety and maternal–fetal bonding during pregnan- cy on postpartum bonding. We hypothesized that maternal– fetal bonding during pregnancy predicts postpartum mother– infant bonding. Furthermore, we included an investigation on the influence of maternal symptoms of depression and anxiety on the postpartum mother–infant relationship.

Methods

Study design and procedure

Originally, 433 pregnant women were recruited at the Heidelberg University Women’s Hospital as part of the Heidelberg Peripartum Study between January 2007 and January 2010. The study is a within-group longitudinal design using a questionnaire survey at two time points. After giving their informed consent, the participants received a prenatal questionnaire set during late pregnancy (M=32 weeks of gestation) with the request to fill them out at home and return them. The questionnaires collected socio-demographic data as well as self-reported information about maternal–fetal bond- ing, symptoms of depression and anxiety. Three months post- partum (M=12 weeks), a second set of questionnaires was sent to the participants’ homes. The set of questionnaires collected the same information as the prenatal set just the questionnaire regarding maternal–fetal bonding was replaced by one regarding maternal postpartum bonding.

For the current assessment, women younger than 18 years, bearing multiplets, with serious medical con- ditions as well as the ones speaking inadequate German were excluded from the study. Out of the final cohort sample of 334, two overlapping random subsamples were recruited for assessment of maternal–fetal bonding and assessment of postpartum bonding impairment due to economic reasons (see Fig. 1). In order to exclude effects of drop-outs, missing values and random sub- sampling on the study results, Little’s MCAR test (miss- ing-completely-at-random-condition) was run with the final study sample (N=334) and all relevant variables. This test ensures the representativeness of subsamples for the final study sample.

The Health Service’s Ethics Committee approved all com- ponents of the project.

Sample

Due to limited overlapping of the random subsamples for assessment of maternal–fetal and postpartum bonding (N= 30), Little’s MCAR test was run on the total data set (N=334) to ensure the representativeness of subsamples for the whole study sample and thereby to exclude effects of drop-outs and missing values. This test was revealed non-significant (p= 0.051) meaning that the MCAR condition is fulfilled. Regarding the fact that the critical p value is just missed, we considered confounders for our analyses (as described below) to increase the MAR plausibility (missing at random) in order to obtain consistent estimations of our subsamples.

The final sample (N=80) was composed of participants having no missing values regarding the primary outcome (postpartum bonding). Women were aged between 24 and 44 years (M=32.8, SD=4.4) and by median were pregnant for the second time. Overall, 80.4 % of the participants were married, 16.1 % in a relationship and 3.5 % divorced. Almost half of the participants (47.0 %) held a university degree, 15.2 % had a university entrance qualification, 34.8 % of the women left school with a high secondary qualification and only 3.0 % had a low secondary qualification. Infants were born between the 38th and 42nd weeks of gestation (M=39.7, SD=1.2). 52.6 % of the women had an induction of labour and 59.7 % received an epidural during delivery. Birth dura- tion was M=6.0 h (SD=5.8) on average. Half of the mothers (50.6 %) delivered spontaneously, while 17.7 % had a primary (planned) caesarean section. 24.1 % of the infants were born by secondary (unplanned) caesarean section and 7.6 % by vacuum extraction. About half (51.9 %) of the infants were female. The average birth weight was M=3,332.2 g (SD= 432.1). 84.8 % of the infants were breast fed postpartum. Descriptive statistics of study variables are demonstrated in Tables 1 and 2.

Instruments

Edinburgh Postnatal Depression Scale (EPDS)

To assess the participants’ severity of depression, both pre- and postnatally, the German version of the Edinburgh Postpartum Depression Scale (Cox et al. 1987) by Bergant et al. (1998) was chosen. It is a ten-item self-rating scale, scored from 0 to 3, that has been validated for the detection of prenatal and postnatal depression in numerous studies (Matthey et al. 2006). A higher sum score indicates greater depression. Originally developed as a screening instrument for the postnatal period, the EPDS is a feasible questionnaire during pregnancy (Cox et al. 1996). The scale is sensitive to changes in severity of depression with a sensitivity and specificity of 91 %, respective 95 % in detecting depressive disorders in mothers (Matthey et al. 2001). In our German sample, data reveal a Cronbach’s α of 0.832

Postpartum bonding 189

(prepartum) and 0.825 (postpartum). Matthey et al. (2006) recommend a cut-off of 15 or more for assessment in the prepartum and of 13 or more in the postpartum.

State–Trait Anxiety Inventory

Anxiety was assessed with the German version of the State– Trait Anxiety Inventory (STAI; Spielberger et al. 1970; Laux

et al. 1970). The state scale (STAI-S) evaluates feelings of apprehension, tension, nervousness and worry as anxiety of a temporary condition while the STAI trait scale (STAI-T) refers to anxiety as a personality feature. Both subscales comprise 20 items. Participants rated on a 4-point scale from 1 (almost never/not at all) to 4 (almost always/very much). Both scales are analyzed separately, and each sum score ranges between 20 and 80 points, with a higher score indicating greater anxiety. The subscales were administered during pregnan- cy and after giving birth. We found the STAI suitable for our study as the instrument does not contain any somatic symptoms, minimizing possible bias due to pregnancy-related symptoms. Grant et al. (2008) validat- ed the STAI to DSM-IV criteria for the prenatal period. Several studies have demonstrated that the STAI has adequate concurrent validity and internal consistency (r=0.83). In our German sample data have an internal consistency of α=0.903 (state) and α=0.895 (trait) in the prepartum and α=0.917 (state) and α=0.900 (trait) in the postpartum. Grant et al. (2008) argues and uses a cut-off score of 40 or more.

Assessed for eligibility (N = 433)

Excluded (n = 99): < 37 weeks of gestation (n = 26) Fetal congenital abnormalities (n = 61) Multiple gestations (n = 12)

Random subsample Maternal-fetal bonding

MFAS (n = 82)

R an

do m

P B

Q -1

6 sa

m pl

e Final cohort sample (N = 334)

R ep

re se

nt at

iv e

sa m

pl e

Little’s MCAR-test (missing- completely-at-random-condition) MAR-plausibility (missing at random)

E nr

ol lm

en t

Random subsample Postpartum Bonding

PBQ-16 (n = 80)

Subjects with complete MFAS and PBQ-16 data (n = 30)

Fig. 1 Flowchart depicting sample recruitment, screening and procedures

Table 1 Descriptive statistics of study variables

N Minimum Maximum Mean SE SD

MFAS 30 68.00 147.00 117.72 3.76 20.60

EPDS pre 54 0.00 19.00 5.15 0.59 4.37

STAI-S pre 54 20.00 59.00 35.57 1.23 9.07

STAI-T pre 55 21.00 54.00 33.93 1.12 8.30

PRAQ-R 55 11.00 40.00 20.60 0.87 6.44

EPDS post 77 0.00 21.00 4.36 0.44 3.86

STAI-S post 79 22.00 67.00 32.72 0.88 7.80

STAI-T post 77 20.00 58.00 31.90 0.88 7.74

190 S. Dubber et al.

Pregnancy Related Anxiety Questionnaire

The Pregnancy Related Anxiety Questionnaire (PRAQ) was developed by Van den Bergh (1990). We used the abbreviated version of the PRAQ-R by Huizink et al. (2004) with ten items to examine pregnancy-related anxiety closely. Due to a lack of a German version we translated it into German. The quality of this translation was validated via a retranslation into English carried out by a native English speaking psychologist. The abbreviated PRAQ-R consists of the following three scales: 'fear of giving birth', 'fear of bearing a physically or mentally handicapped child' and 'concern about one’s appearance'. The following internal consistencies for the different pregnancy trimesters were demon- strated: scale 1 (birth) Cronbach’s α between 0.79 and 0.83; scale 2 (health of the child) between 0.87 and 0.88; scale 3 (own appearance) between 0.76 and 0.83. The mothers’ answers are scored from 1 to 5 with the total sum ranging between 10 and 50. As the items are coded positively, a value of 10 represents the lowest level of anxiety. In our German sample the internal consistency reaches α=0.816.

Maternal–Fetal Attachment Scale

In order to assess maternal-fetal bonding we used the revised MFAS by Van den Bergh (1989), which was originally devel- oped by Cranley (1981). Due to a lack of a German version we

translated the MFAS into German. The quality of this transla- tion was validated via a retranslation into English carried out by a native English speaking psychologist. The 24-item scale consists of four subscales: 'anticipation of interaction with the baby', 'giving of self', 'name for the baby', 'interaction with fetus'. The items are scored on a 7-point Likert-scale (1 for ‘definitely no’ to 7 for ‘definitely yes’). The sum scores range between 24 and 168. A higher sum score represents greater bonding to the fetus. In our German sample an α of 0.792 is reached.

Postpartum Bonding Questionnaire

Postpartum bonding of the mother to her child was assessed by the abridged German version of the Postpartum Bonding Questionnaire (PBQ-16; Reck et al. 2006). Originally, the questionnaire was developed by Brockington et al. (2001) as a screening instrument in order to diagnose bonding disorders consisting of 25 items yielding scores on four factors. The abridged German scale consists of 16 items. As the four factor structure of the original English version of the PBQ could not be confirmed by Reck et al. (2006) in a German validation study (N=862), nine items were eliminated for a one factor structure solution. The response categories range from ‘al- ways’ to ‘never’ on a 6-point Likert scale with a higher sum score indicating impaired bonding. The sum scores range between 0 and 80 points. Reck et al. (2006) report a Cronbach’s α of 0.85. In our German sample the internal consistency is α=0.831.

Statistical analysis

For all analyses conducted in this study, we used the Statistical Package for Social Sciences (IBM SPSS v. 21.0). Power analyses were run by G*Power 3.1.9.2 (Faul et al. 2009, 2007). For the maximum sample, large- and medium-sized effects can be ruled out for correlation coefficients and regres- sion steps. For the small sample, only large effects can be excluded sufficiently in case of non-significant results (Table 3). Because of scale-specific amounts of missing values the valid number of cases n varies dependent on the data subsets used for the particular test (MCAR condition ful- filled). Kolmogorov–Smirnov test and Shapiro–Wilk test were run to analyze deviation from normal distribution of study variables. Since a deviation from normal distribution was revealed for almost every variable, we decided to run non- parametric analyses regarding robust estimations.

First, study variables and potential covariates were corre- lated with postpartum bonding (PBQ-16). Second, regression analyses (backwards) containing the significant variables out of step one were performed on the PBQ-16 score (robust estimation with Wald statistics). The independent variables were removed if they failed to explain a significant amount of

Table 2 Frequencies and percentages of participants below and above cut-off values

Scale Cut-off f % Valid % Cumulative %

EPDS prea <15 51 63.75 94.44 94.44

≥15 3 3.75 5.56 100.00 Total 54 67.50 100.00

EPDS postb <13 74 92.50 96.10 96.10

≥13 3 3.75 3.90 100.00 Total 77 96.25 100.00

STAI-T preb <40 42 52.50 76.36 76.36

≥40 13 16.25 23.64 100.00 Total 55 68.75 100.00

STAI-S preb <40 38 47.50 70.37 70.37

≥40 16 20.00 29.63 100.00 Total 54 67.50 100.00

STAI-T postb <40 66 82.50 85.71 85.71

≥40 11 13.75 14.29 100.00 Total 77 96.25 100.00

STAI-S postb <40 71 88.75 89.87 89.87

≥40 8 10.00 10.13 100.00 Total 79 98.75 100.00

a According to the recommendations of Matthey et al. (2006) b According to Grant et al. (2008)

Postpartum bonding 191

variance of the PBQ-16. Third, the regression weights of the final model were bootstrapped (N=1,000 samples, Mersenne– Twister; initial value 13, 95 % percentile confidence interval) to ensure their explanatory value despite skewed distribution.

Results

Study variables and covariates

The results of the correlation of potential independent covar- iates with postpartum bonding revealed only maternal educa- tion to be significantly correlated with postpartum bonding (ρ=0.251, p<0.05). The higher mothers were educated, the more bonding impairment was reported. Regarding the study variables, prenatal pregnancy related anxiety (PRAQ-R), post- partum symptoms of depression (EPDS) and postpartum trait anxiety (STAI-T) were significantly positively correlated with postpartum bonding. The correlations and p values of the study variables with the PBQ-16 are demonstrated in Table 4.

Maternal–fetal bonding during pregnancy (MFAS) was significantly correlated with postpartum bonding: the higher mothers were attached to their unborn child during pregnancy, the less bonding impairment was reported.

Regression analysis

Finally, we performed a stepwise regression analysis (backwards) on postpartum bonding with the significant var- iables out of the former analyses (maternal education, mater- nal–fetal bonding, pregnancy-related anxiety and postpartum

symptoms of depression) by the use of generalized linear modeling with robust estimators and Wald statistics (Table 5).

The final model contained maternal–fetal bonding (ΔR2adj=9.84 %; β=−0.42, c=0.63, estimated by linear re- gression) and postpartum symptoms of depression (ΔR2adj= 10.99 %; β=0.51, c=0.72, estimated by linear regression) as independent variables. The regression weights were bootstrapped, revealing that maternal–fetal bonding (B= −0.076, SE=0.026, p<0.01) as well as postpartum depressive symptoms (B=0.529, SE=0.183, p=0.01) remain significant in explaining postpartum bonding. These last two steps re- vealed the results to be invariant of statistical method and confirmed maternal depressive symptoms as the stronger as- sociated variable.

In a last robust regression model we only included prepartum variables significantly correlated to maternal post- partum bonding. This aimed at effects of prepartum anxiety to be covered by the effect of postpartum depression. In this model, pregnancy-related anxiety and maternal–fetal bonding were included. Maternal–fetal bonding remains, thereby, the only significant variable (B=−0.078, χ2=7.697, p<0.01).

Table 3 Power analysis for maximum and minimum sample size

N Small effecta Medium effectb Large effectc

rd 80 0.144 0.793 0.999

30 0.083 0.384 0.863

R step 1e,f 80 0.154 0.819 0.996

30 0.082 0.347 0.710

R step 2e,g 80 0.182 0.869 0.998

30 0.092 0.415 0.785

a α=0.05 b r=0.1, f2 =0.02 c r=0.3, f2 =0.15 c r=0.5, f2 =0.35 d Pearson correlation as approximation for Spearman correlation e Linear regression as approximation for generalized linear model f Backward step eliminating one non-significant variable among four out of four g Backward step eliminating one non-significant variable among three out of four

Table 4 Correlations between study variables and the PBQ-16

a Spearman's ρ b Two-tailed c Pairwise case exclusion

Parameter ρa pb Nc

MFAS −0.401 0.028 30 EPDS pre 0.262 0.056 54

STAI-S pre 0.199 0.149 54

STAI-T pre 0.247 0.069 55

PRAQ-R 0.271 0.045 55

EPDS post 0.417 0.000 77

STAI-S post 0.197 0.081 79

STAI-T post 0.315 0.005 77

Table 5 Stepwise backward regression (robust estimation) on PBQ-16

Parameter B SE 95 % Wald CI Inferential statistics

Lower Upper Wald χ2 df pa

(Intercept) 24.78 5.25 14.48 35.08 22.24 1 0.000

Education 0.78 0.62 −0.43 1.99 1.59 1 0.207 MFAS −0.07 0.03 −0.13 −0.02 7.25 1 0.007 PRAQ-R 0.00 0.06 −0.11 0.11 0.00 1 0.959 EPDS post 0.50 0.15 0.21 0.79 11.05 1 0.001

(Intercept) 24.67 4.98 14.90 34.44 24.49 1 0.000

Education 0.79 0.64 −0.46 2.03 1.53 1 0.216 MFAS −0.07 0.03 −0.13 −0.02 7.22 1 0.007 EPDS post 0.50 0.15 0.20 0.80 10.73 1 0.001

(Intercept) 28.32 2.84 22.76 33.89 99.50 1 0.000

MFAS −0.08 0.02 −0.12 −0.03 9.84 1 0.002 EPDS post 0.53 0.16 0.22 0.84 10.99 1 0.001

a Two-tailed

192 S. Dubber et al.

Pregnancy-related anxiety remains non-significant (B=0.020, χ2=0.068, p=0.794).

Discussion

The data presented above indicate an important association between maternal–fetal bonding during pregnancy as well as maternal postpartum depressive symptoms on postpartum bonding. In this study, they both affect the mother’s emotional bond to the infant postpartum and explain 20.8 % of the variance, which is, according to Cohen (1988), a medium- sized effect.

These results add to the current body of literature and provide further insight into the understanding of the antenatal roots of the mother–infant relationship, as has been demonstrat- ed in only a few previous studies to our knowledge (Figueiredo and Costa 2009; Müller 1996; Siddiqui and Hägglöf 2000; van Bussel et al. 2010). Our results confirm the findings by van Bussel et al. (2010) and Müller (1996). Both reported a signif- icant correlation between pre- and postnatal maternal bonding comparable to our results. Although Müller (1996) did not control for depression or anxiety, she demonstrated that 17 % of postnatal bonding could be explained by prenatal bonding.

Furthermore, the presumption that the maternal prenatal relationship to the fetus is predictive of postpartum bonding is supported by Carneiro et al. (2006). Carneiro et al. (2006) demonstrated that prenatal parental role play with a baby doll is associated with their later interaction with their infant. Gloger-Tippelt (2005) stated that over the progression of a pregnancy the expectant mothers attribute characteristic traits to their unborn child based on information on and experiences with them. Thereby, mothers create a mental image of their child which in turn strengthens their bond with them.

Although the study findings support a moderate relation- ship between prenatal and postnatal bonding in mothers, maternal depressive symptoms also remained significant in explaining the variance in postpartum bonding. Several stud- ies support the assumption that maternal mental health in the postpartum has an influence on mother–child bonding. Even subclinical maternal depressive symptoms are reported to be associated with a lower quality of maternal bonding (Moehler et al. 2006; Reck et al. 2006). Edhborg et al. (2011) demonstrated that mothers with postnatal depressive symptoms measured with the EPDS showed lower emotional bonding with their infants 2– 3 months postpartum. Brockington et al. (2001) found a rate of 29 % of mothers diagnosed with postpartum depression reporting bonding impairment. Moehler et al. (2007) reported that even maternal depressive symptoms not reaching the level of clinical diagnosis and treatment have an impact on child behavioral development; thus, emphasizing the need to further establish screening for depressive symptoms in clinical routine.

Furthermore, in the current study, high levels of maternal education, pregnancy-related fear and postnatal anxiety were independently and negatively associated with the mother’s emotional bonding in the postpartum period. However, these associations did not hold their statistical significance within the final regression model. Even in an exploratory model with just significant prepartum variables integrated (i.e., pregnancy-related anxiety and maternal–fetal bonding), only maternal–fetal bonding explained a significant amount of variance in postpartum bonding.

These findings are consistent with recent literature. Although some studies demonstrate the significance of mater- nal anxiety on the mother–infant relation, this is not an as well researched area as maternal depression. Britton (2011) found that ratings of overall temperamental difficulty were indepen- dently associated with both symptoms of anxiety and depres- sion in the early postpartum period. He also found that anxious mothers were less sensitive, less responsive and showed less competence in parenting (Britton 2005). In contrast, Edhborg et al. (2011) did not find negative associations between perina- tal general anxiety and postpartum bonding in a community sample of 672 women. However, this mostly resembles our results. Despite the significant correlations between pregnancy- related fear and general anxiety with postpartum bonding, they did not hold their statistical significance in the final results. Thus, the different impact emphasizes the importance of differ- entiating between symptoms of anxiety or depression.

Limitations

Several points limiting our findings should be considered. First of all, the restrictive sample of consistent data regarding the MFAS and the PBQ-16 limits the generalizability of the found association to other populations. Our study demonstrat- ed correlates of depressive symptoms in women, but the findings need to be interpreted with caution as the EPDS only measures depressive symptoms; whether or not clinical levels of depression were experienced by the women cannot be revealed. We only used maternal self-report measures and no additional maternal diagnosis according to DSM-IV criteria. However, as reported above, the EPDS has a high sensitivity and specificity for depressive disorder and research showed that even maternal depressive symptoms not reaching the level of clinical diagnosis have an impact on child behavioral development (Moehler et al. 2007). The possibility remains that some of the associations described were the result of a confounding factor; for example it is known that depressed mothers often rate themselves as worse parents than they really are due to symptoms of low self-esteem and negative thoughts (Frankel and Harmon 1996; Hornstein et al. 2006).

While interesting correlations have emerged from the study, the design does not permit directional or causal conclu- sions to be drawn. Maternal depressive symptoms were

Postpartum bonding 193

assessed twice, during pregnancy and after delivery. As de- pressive symptoms might fluctuate over time (Matthey et al. 2012), a repeated assessment as in the current study is of benefit because it also reduces retrospective reporting bias.

As a specific characteristic, this sample had a relatively high level of education which should be considered when generalizing these findings on less educated populations. Furthermore, selective bias and social desirability have to be taken into account as potential confounding factors; therefore, our findings should be regarded as preliminary.

Conclusions

This study demonstrates that maternal–fetal bonding and post- partum depressive symptoms are significantly associated with the postpartum mother–child relationship and can serve as an important diagnostic aid in early identification of women for whom the mother–child interaction is likely to be suboptimal. By considering the two relevant symptoms dimensions, anxiety and depression, in the prenatal and postpartum period separate- ly, our results give a more comprehensive picture of the dimen- sions involved in postpartum bonding impairment, providing specific cues for intervention. The findings of this study are useful in the prevention of emotional disorders in childhood which are potentially associated with maternal bonding impair- ment. Despite the growing body of literature on maternal de- pression and bonding impairment and its effects on the child, preventive actions regarding psychiatric counseling and/or psycho-therapy are still lacking awareness in clinical routine.

Acknowledgments We would like to thank the women who were willing to participate in this study. Furthermore, we would like to thank Professor Bea Van den Bergh for providing us with her revised version of the MFAS.

Conflict of interest The authors declare that they have no conflict of interest.

References

Alhusen JL (2008) A literature update on maternal–fetal attachment. J Obstet Gynecol Neonatal Nurs 37(3):315–328. doi:10.1111/j.1552- 6909.2008.00241.x

Alhusen JL, Gross D, Hayat MJ, Woods AB, Sharps PW (2012) The influence of maternal-fetal attachment and health practices on neo- natal outcomes in low-income, urban women. Res Nurs Health 35(2):112–120. doi:10.1002/nur.21464

Alhusen JL, Hayat MJ, Gross D (2013) A longitudinal study of maternal attachment and infant developmental outcomes. Arch Women Ment Health 16(6):521–529. doi:10.1007/s00737-013-0357-8

Andersson L, Sundstrom-Poromaa I, Wulff M, Astrom M, Bixo M (2006) Depression and anxiety during pregnancy and six months postpar- tum: a follow-up study. Acta Obstet Gynecol Scand 85(8):937–944. doi:10.1080/00016340600697652

Atkinson L, Paglia A, Coolbear J, Niccols A, Parker KC, Guger S (2000) Attachment security: a meta-analysis of maternal mental health correlates. Clin Psychol Rev 20(8):1019–1040

Bennett HA, Einarson A, Taddio A, Koren G, Einarson TR (2004) Prevalence of depression during pregnancy: systematic review. Obstet Gynecol 103(4):698–709

Bergant AM, Nguyen T, Heim K, Ulmer H, Dapunt O (1998) Deutschsprachige Fassung und Validierung der "Edinburgh postna- tal depression scale". Dtsch Med Wochenzeitschr 123:35–40

Bonari L, Pinto N, Ahn E, Einarson A, Steiner M, Koren G (2004) Perinatal risks of untreated depression during pregnancy. Can J Psychiatry 49(11):726–735

Britton JR (2005) Pre-discharge anxiety among mothers of well new- borns: prevalence and correlates. Acta Paediatr 94:1771–1776. doi: 10.1080/08035250500192664

Britton JR (2011) Infant temperament and maternal anxiety and depressed mood in the early postpartum period. Womens Health 51(1):55–71. doi:10.1080/03630242.2011.540741

Brockington I (2004) Postpartum psychiatric disorders. The Lancet, 363(9405). doi:10.1016/S0140-6736(03)15390-1

Brockington IF, Oates J, George S, Turner D, Vostanis P, Sullivan M, Loh C, Murdoch C (2001) A screening questionnaire for mother–infant bonding disorders. Arch Womens Ment Health 3(4):133–140. doi: 10.1007/s007370170010

Buist A, Morse CA, Durkin S (2003) Men's adjustment to fatherhood: implications for obstetric health care. J Obstet Gynecol Neonatal Nurs 32(2):172–180

Carneiro C, Corboz-Warnery A, Fivaz-Depeursinge E (2006) The pre- natal Lausanne Trilogue Play: a new observational assessment tool of the prenatal co-parenting alliance. Infant Ment Health J 27(2): 207–228

Carter CS, Keverne EB (2002) The neurobiology of social affiliation and pair bonding. In: Pfaff DW, Arnold AP, Etgen AM, Fahrbach SE, Rubin RT (eds) Hormones, brain and behaviour, Vol. 1, Academic Press, San Diego, pp 299–337

Cohen J (1988) Statistical power analysis for the behavioral sciences. Erlbaum, Hillsdale

Condon JT (1993) The assessment of antenatal emotional attachment: development of a questionnaire instrument. Br J Med Psychol 66(Pt 2):167–183

Condon JT, Corkindale C (1998) The assessment of parent-to-infant attachment: development of a self-report questionnaire instrument. J Reprod Infant Psychol 16(1):57–77

Cox JL, Holden JM, Sagovsky R (1987) Detection of postnatal depres- sion. Development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry 150:782–786. doi:10.1192/bjp.150.6.782

Cox JL, Chapman G, Murray D, Jones P (1996) Validation of the Edinburgh postnatal depression scale (EPDS) in non-postnatal wom- en. J Affect Disord 39:185–189. doi:10.1016/0165-0327(96)00008-0

Cranley MS (1981) Development of a tool for the measurement of maternal attachment during pregnancy. Nurs Res 30(5):281–284

Edhborg M, Nasreen H-E, Kabir ZN (2011) Impact of postpartum de- pressive and anxiety symptoms on mothers’ emotional tie to their infants 2–3 months postpartum: a population-based study from rural Bangladesh. Arch Womens Ment Health 14(4):307–316. doi:10. 1007/s00737-011-0221-7

Faul F, Erdfelder E, Lang A-G, Buchner A (2007) GPower 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 39(2):175–191

Faul F, Erdfelder E, Buchner A, Lang A-G (2009) Statistical power analyses using G*Power 3.1: tests for correlation and regression analyses. Behav Res Methods 41(4):1149–1160

Feldman R, Greenbaum CW, Mayes LC, Erlich SH (1997) Change in mother–infant interactive behavior: relations to change in the moth- er, the infant, and the social context. Infant Behav Dev 20(2):151– 163. doi:10.1016/s0163-6383(97)90018-7

194 S. Dubber et al.

Feldman R, Weller A, Zagoory-Sharon O, Levine A (2007) Evidence for a neuroendocrinological foundation of human affiliation: plasma oxytocin levels across pregnancy and the postpartum period predict mother–infant bonding. Psychol Sci 18(11):965–970. doi:10.1111/j. 1467-9280.2007.02010.x

Field T (1998) Maternal depression effects on infants and early interven- tions. Prev Med 27(2):200–203. doi:10.1006/pmed.1998.0293

Field T (2010) Postpartum depression effects on early interactions, par- enting, and safety practices: a review. Infant Behav Dev 33(1):1–6. doi:10.1016/j.infbeh.2009.10.005

Figueiredo B, Costa R (2009) Mother’s stress, mood and emotional involvement with the infant: 3 months before and 3 months after childbirth. Arch Womens Ment Health 12(3):143–153. doi:10.1007/ s00737-009-0059-4

Frankel KA, Harmon RJ (1996) Depressed mothers: they don't always look as bad as they feel. J Am Acad Child Adolesc Psychiatry 35(3): 289–298. doi:10.1097/00004583-199603000-00009

Gloger-Tippelt G (2005) Psychologischer Übergang zur Elternschaft. In: Thun-Hohenstein L (ed) Übergänge – Wendepunkte und Zäsuren in der kindlichen Entwicklung. Vandenhoeck & Ruprecht, Göttingen, pp 55–74

Goodman SH, Gotlib IH (1999) Risk for psychopathology in the children of depressed mothers: a developmental model for understanding mechanisms of transmission. Psychol Rev 106(3):458–490

Grant KA, McMahon C, Austin MP (2008) Maternal anxiety during the transition to parenthood: a prospective study. J Affect Disord 108: 101–111. doi:10.1016/j.jad.2007.10.002

Grigoriadis S, Vonder Porten EH, Mamisashvili L, Tomlinson G, Dennis C- L, Koren G et al (2013) The impact of maternal depression during pregnancy on perinatal outcomes: a systematic review and meta-analy- sis. J Clin Psychiatry 74(4):e321–e341. doi:10.4088/JCP.12r07968

Grote NK, Bridge JA, Gavin AR, Melville JL, Iyengar S, Katon WJ (2010) A meta-analysis of depression during pregnancy and the risk of preterm birth, low birth weight, and intrauterine growth restric- tion. Arch Gen Psychiatry 67(10):1012–1024. doi:10.1001/ archgenpsychiatry.2010.111

Hornstein C, Schenk S, Wortmann-Fleischer S, Downing G, Schwarz M (2006) Videotherapie bei postpartalen Störungen. Ein interaktionales Behandlungskonzept bei Müttern mit. Depressionen und Psychosen Psychotherapeutics 51(5):363–368. doi:10.1007/s00278-006-0484-z

Huizink AC, Mulder EJ, Robles de Medina PG, Visser GH, Buitelaar JK (2004) Is pregnancy anxiety a distinctive syndrome? Early Hum Dev 79(2):81–91. doi:10.1016/j.earlhumdev.2004.04.014

Laux L, Glanzmann P, Schaffner P, Spielberger CD (1970) Das State– Trait-Angstinventar. Beltz Test, Weinheim

Marcus SM (2009) Depression during pregnancy: rates, risks and conse- quences—Motherisk Update 2008. Can J Clin Pharmacol 16(1): e15–e22

Matthey S, Barnett B, Kavanagh DJ, Howie P (2001) Validation of the Edinburgh Postnatal Depression Scale for men, and comparison of item endorsement with their partners. J Affect Disord 64(2–3):175– 184. doi:10.1016/S0165-0327(00)00236-6

Matthey S, Barnett B, Howie P, Kavanagh DJ (2003) Diagnosing post- partum depression in mothers and fathers: whatever happened to anxiety? J Affect Disord 74(2):139–147

Matthey S, Henshaw C, Elliott S, Barnett B (2006) Variability in use of cut- off scores and formats on the Edinburgh Postnatal Depression Scale— implications for clinical and research practice. Arch Womens Ment Health 9(6):309–315. doi:10.1007/s00737-006-0152-x

Matthey S, Ross-Hamid C (2012) Repeat testing on the Edinburgh Depression Scale and the HADS-A in pregnancy: Differentiating between transient and enduring distress. J Affect Disord 141(2–3): 213–221. doi:10.1016/j.jad.2012.02.037

Moehler E, Brunner R, Wiebel A, Reck C, Resch F (2006) Maternal depressive symptoms in the postnatal period are associated with long-term impairment of mother–child bonding. Arch Womens Ment Health 9(5):273–278. doi:10.1007/s00737-006-0149-5

Moehler E, Kagan J, Parzer P, Brunner R, Reck C, Wiebel A, Poustka L, Resch F (2007) Childhood behavioral inhibition and maternal symp- toms of depression. Psychopathology 40(6):446–452. doi:10.1159/ 000107429

Müller ME (1993) The development of the prenatal attachment inventory. Western J Nurs Res 15:199–211

Müller ME (1994) A questionnaire to measure mother-to-infant attach- ment. J Nurs Meas 2:129–141

Müller M (1996) Prenatal and postnatal attachment: a modest correlation. J Obstet Gynecol Neonatal Nurs 25(2):161–166

Raphael-Leff J (2001) Pregnancy: the inside story. Karnac Books, London

Reck C, Hunt A, Fuchs T, Weiss R, Noon A, Moehler E, Downing G, Tronick EZ, Mundt C (2004) Interactive regu- lation of affect in postpartum depressed mothers and their infants: an overview. Psychopathology 37(6):272–280. doi:10. 1159/000081983

Reck C, Klier CM, Pabst K, Stehle E, Steffenelli U, Struben K, Backenstrass M (2006) The German version of the Postpartum Bonding Instrument: psychometric properties and association with postpartum depression. Arch Womens Ment Health 9(5):265–271. doi:10.1007/s00737-006-0144-x

Reck C, Struben K, Backenstrass M, Stefenelli U, Reinig K, Fuchs T, Sohn C, Mundt C (2008) Prevalence, onset and comorbidity of postpartum anxiety and depressive disorders. Acta Psychiatr Scand 118:459–468. doi:10.1111/j.1600-0447.2008.01264.x

Ross LE, McLean LM (2006) Anxiety disorders during pregnancy and the postpartum period: a systematic review. J Clin Psychiatry 67(8): 1285–1298

Siddiqui A, Hägglöf B (2000) Does maternal prenatal attachment predict postnatal mother–infant interaction? Early Hum Dev 59(1):13–25. doi:10.1016/s0378-3782(00)00076-1

Spielberger CD, Gorsuch RL, Lushene RE (1970) State–Trait Anxiety Inventory, manual for the State–Trait Anxiety Inventory. Consulting Psychologist Press, Palo Alto

Taylor A, Atkins R, Kumar R, Adams D, Glover V (2005) A new Mother-to-Infant Bonding Scale: links with early maternal mood. Arch Womens Ment Health 8(1):45–51

Tronick E, Reck C (2009) Infants of depressed mothers. Harvard Rev Psychiatry 17(2):147–156. doi:10.1080/10673220902899714

van Bussel JCH, Spitz B, Demyttenaere K (2010) Three self-report questionnaires of the early mother-to-infant bond: reliability and validity of the Dutch version of the MPAS, PBQ and MIBS. Arch Womens Ment Health 13(5):373–384. doi:10.1007/s00737-009- 0140-z

Van den Bergh B (1989) De emotionele toestand van de (zwangere) vrouw, obstetrische complicaties en het gedrag en de ontwikkeling van de foetus en van het kind tot de leeftijd van zeven maanden. Unpublished doctoral thesis, Catholic University of Leuven

Van den Bergh BR (1990) The influence of maternal emotions during pregnancy on fetal and neonatal behavior. J Prenat Perinat Psychol Health 5(2):119–130

Wan MW, Green J (2009) The impact of maternal psychopathology on child–mother attachment. Arch Womens Ment Health 12(3):123– 134. doi:10.1007/s00737-009-0066-5

Waters E, Merrick S, Treboux D, Crowell J, Albersheim L (2003) Attachment security in infancy and early adulthood: a twenty-year longitudinal study. In: Hertzig ME, Farber EA (eds) Annual progress in child psychiatry and child development: 2000–2001. Brunner- Routledge, New York, pp 63–72

Postpartum bonding 195

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.