Does body image influence the relationship between body weight and breastfeeding maintenance in new mothers?

: Objectives. Obese women have lower breastfeeding initiation and maintenance than healthy weight women. Research generally focuses on biomedical explanations for these differences. The impact of psychosocial factors, including women's post-childbirth well-being and body image cognitions in relation to breastfeeding are less well understood. We aimed to investigate women's body image after childbirth in hospital, and 6-8 weeks later, studying the impact of body image and psychological distress on breastfeeding maintenance at 6-8 weeks, comparing obese and healthy weight women. Design. Longitudinal semi-structured questionnaire survey. Methods. Demographic and biomedical factors were measured around childbirth. Body image and psychological distress were assessed within 72 hours of birth and by postal questionnaire at 6-8 weeks, for 70 obese and 70 healthy weight women initiating either exclusive (breastmilk only) breastfeeding or mixed feeding (with formula milk) in hospital. Breastfeeding status was re-assessed at 6-8 weeks. Results. Obese women were less likely to exclusively breastfeed in hospital and maintain breastfeeding to 6-8 Abstract 1 Objectives . Obese women have lower breastfeeding initiation and maintenance rates than 2 healthy weight women. Research generally focuses on biomedical explanations for this. 3 Psychosocial factors, including body image and well-being after childbirth are less well 4 understood as predictors of breastfeeding. In obese and healthy weight women, we 5 investigated changes in body image between 72 hours post-delivery and 6-8 weeks post- 6 partum, studying how women’s body image related to breastfeeding initiation and 7 maintenance. We also investigated how psychological distress was related to body image. 8 Design. Longitudinal semi-structured questionnaire survey. 9 Methods. Body image and psychological distress were assessed within 72 hours of birth and 10 by postal questionnaire at 6-8 weeks, for 70 obese and 70 healthy weight women initiating 11 exclusive (breastmilk only) breastfeeding or mixed feeding (with formula milk) in hospital. 12 Breastfeeding was re-assessed at 6-8 weeks. 13 were less to exclusively breastfeed in breastfeeding to Better image was to maintaining to psychological but most significant predictor maintenance multivariate including body and Body but not moderate the relationship weight and breastfeeding maintenance. Body image was lower overall in obese women, all women had low body image satisfaction around reducing further at 6-8 weeks. Conclusions . Health professionals should consider women’s body image when discussing breastfeeding. A focus on breast function over form may support breastfeeding for all women.


25
Breastfeeding promotes health 26 27 There is clear evidence that breastfeeding is beneficial for the health of mother and 28 infant (Victora et al., 2016). UK breastfeeding rates compare poorly with other western have also been related to postnatal depression (Watkinson, Murray & Simpson, 2016

Participants and Methods 135
This was a longitudinal questionnaire-based study. We compared healthy weight 136 (defined as body mass index (BMI) 18.5<25kg/m 2 ) and obese women (defined as BMI 137 >30kg/m 2 ). Recruitment was by a research midwife from January 2011 to March 2013. 138 Women were eligible for inclusion if they had breastfed at first feed, given birth to a single 139 baby at >37 weeks gestation and had a BMI at any stage of pregnancy of either 140 18.5<25kg/m 2 or >30kg/m 2 (from maternal records). Women were ineligible if they were 141 <18 years old, were not being discharged home with their baby, or were unable to give 142 informed consent. Women whose baby is admitted to a neonatal unit are likely to face 143 additional challenges in relation to postnatal recovery and in initiating breastfeeding, so were not included in this study. We specifically recruited equal numbers of 'healthy weight' 145 and 'obese weight' women to explore predictors of breast-feeding maintenance in these 146 groups. Although it is of interest to study underweight women, we wanted to focus on how 147 body image affected breastfeeding for obese and 'healthy' weight women in this research. 148 Eligible women were approached on the postnatal ward within 72 hours of giving birth. 149 Following informed consent, women completed a questionnaire prior to hospital discharge 150 (Time 1) and a second postal questionnaire at 6-8 weeks postpartum (Time 2). This time 151 point is selected in many studies as a period when women have established some routine in 152 their home/family context, and some stability in their infant feeding behaviour.

Body Image 194
The concept of 'body image' captures a range of attitudes towards the physical self. 195 The experience of childbirth and becoming a mother presents a significant challenge to 196 women's physical self-concept (Lupton, 2012). We were interested in women's appraisals of 197 their body (evaluations, affective reactions) and overall importance of (investment in) their 198 appearance, comparing women's attitudes in the immediate postnatal period with 6-8 199 weeks later, and in relation to breastfeeding maintenance. We used a short version of a

Analysis 232
We examined the distribution of study variables to check for non-normal distributions, 233 outliers and missing data. We investigated breastfeeding differences between obese vs. 234 healthy weight women using Chi square tests. Effect sizes were assessed using partial eta 235 squared. Although body image is generally stable over time, childbirth and initiating 236 breastfeeding are uniquely challenging embodied events, so it was important to consider 237 whether body image scores at Time 1 or Time 2 were likely to be best predictors of 238 breastfeeding maintenance at 6-8 weeks. We reasoned that Time 1 in hospital was unlikely 239 to provide a realistic 'baseline' since women's perception of their bodies at this time may be 240

256
Demographics 257 140 participants were recruited (n=70 healthy weight and n=70 obese). The overall response 258 rate at follow-up was 84%, (n= 117) with no difference in response rates between healthy 259 weight and obese women (86% vs 81%, respectively, p=0.65). 260 Healthy weight women were more likely than obese women to be exclusively breastfeeding 261 at recruitment at Time 1 (p=0.018), and to have maintained exclusive or 'mixed' 262 breastfeeding at Time 2, (p=0.052) as shown in Table 1. 263 Correlations between breastfeeding maintenance at 6-8 weeks, weight status, body image 279 and psychological distress at Time 2, are shown in Table 3. Education level and delivery 280 method (vaginal vs. caesarean) were included as potential covariates. Body image variables 281 were positively or negatively inter-correlated as expected. Appearance evaluation and body satisfaction were highly inter-correlated (r=.83). After exploratory analyses we excluded 283 'appearance evaluation' from further regression analysis. Psychological distress was related 284 to less breastfeeding maintenance, lower appearance evaluation and body satisfaction, and 285 higher weight self-classification. 286 Wald =7.98, p=.005). To assess moderation effects we included interactions (product 293 terms) of weight status (healthy weight/obese) and body satisfaction, appearance 294 orientation, overweight preoccupation and self-classified weight in a final block. This model 295 was non-significant, (Nagelkerke R 2 =.21, n.s) suggesting no moderation effect.  Table 4. 300 Demerath, 2012) we tested Time 2 body image variables as potential mediators of the 310 relationship between weight status and breastfeeding maintenance at 6-8 weeks (Baron & 311 Kenny 1986). Breastfeeding outcome (any breastfeeding vs no breastfeeding) was the 312 dependent variable in separate logistic regression analyses with weight status entered first, 313 followed by the potential mediator, using bootstrapping (BCa95%CI) to evaluate effects 314 (Table 5). 315 Weight status was a significant predictor in model 1, but became non-significant with 318 appearance evaluation, body satisfaction, overweight preoccupation and weight self-319 classification, suggesting partial mediation effects of these variables. 320

Changes in Body Image and Psychological Distress after Childbirth 322
We considered how body image changed from Time 1 to Time 2 for obese and healthy 323 weight women. Table 6 shows mean body image scores at both time points for MBSRQ 324 components comparing healthy weight and obese women. 325 Obese women had poorer body image at both time points, in relation to appearance 328 evaluation and body satisfaction. Obese women recorded more overweight preoccupation and higher weight self-classification with medium to large effects (partial eta 2 ). There was 330 no difference, or change over time in appearance orientation. 331 Appearance evaluation and body satisfaction were significantly lower, irrespective of 332 weight status at Time 2. Interactions (between weight status and time) were not significant. 333 Overweight preoccupation increased significantly over time for both healthy weight and 334 obese women, with no significant interaction effect. one-sample t-tests, as shown in Table 6. 341 Women's appearance evaluation, appearance orientation and overweight preoccupation, Breastfeeding is an important health behaviour, which is rewarding for most women 357 but can be challenging to initiate and maintain, contributing to the challenges of early 358 motherhood. Health psychology has much to add to our understanding of this behaviour. 359 The experience of childbirth has a huge influence on how women perceive their body, and  The primary research question focused on predicting breastfeeding maintenance 373 (exclusive or mixed) at 6-8 weeks postnatal. We found all body image components, except 374 appearance orientation were correlated with breastfeeding maintenance and weight status. 375 Higher satisfaction and appearance evaluation were positively related to breastfeeding and 376 negatively related to weight status. Greater preoccupation with overweight and weight 377 self-classification were negatively related to breastfeeding and positively related to weight 378 status. In regression analysis including weight status, body image components as a block 379 added significantly to the variance predicted. Individually, body satisfaction was a significant 380 predictor, and appearance evaluation, body satisfaction, overweight preoccupation and 381

study. 446
We deliberately compared discrete groups of healthy weight and obese women but 447 recognise underweight and overweight women may also have different experiences of 448 breastfeeding. We categorised women's weight according to case notes at booking in early 449 pregnancy, acknowledging that this may have changed at Time 2. We followed up the 450 sample at 6-8 weeks, a common time-point chosen in research, however longer follow-up of 451 breastfeeding maintenance would have been useful in relation to assessing body image. 452 Our main outcome measure constructed to represent 'breastfeeding' versus 'no 453 breastfeeding' at Time 2, also did not reflect how long women had breastfed on discharge 454 from hospital, and may have lacked sensitivity, as many women switch to formula in the 455 first 2 weeks post-discharge (McAndrew et al., 2012). Additionally, the study was slightly 456 over-powered at Time 2, with potential for Type 1 error given a relatively large number of 457 variables and multiple testing.