Disclosure: RD is senior clinical researcher for FWO Vlaanderen (2010-2015).
Anxiety during early pregnancy predicts postpartum weight retention in obese mothers
Version of Record online: 25 MAY 2013
Copyright © 2013 The Obesity Society
Volume 21, Issue 9, pages 1942–1949, September 2013
How to Cite
Bogaerts, A. F. L., Van den Bergh, B. R. H., Witters, I. and Devlieger, R. (2013), Anxiety during early pregnancy predicts postpartum weight retention in obese mothers. Obesity, 21: 1942–1949. doi: 10.1002/oby.20352
Funding agencies: AB was supported by a PWO project from Flanders. BVDB is funded by the Netherlands Organisation for Scientific Research (NWO; Brain and Cognition Program; 2008-2012), and by EU 7th Framework Program (FP7-Health-2011; BrainAGE; 2012-2017).
This article has been presented orally at the 1st Nordic Congress on Obesity in Gynaecology and Obstetrics (NOCOGO), 22-24 October 2012, Billund, Denmark and was awarded with price for “best oral poster presentation” 24 October 2012, Billund Denmark.
- Issue online: 23 SEP 2013
- Version of Record online: 25 MAY 2013
- Accepted manuscript online: 14 FEB 2013 07:19AM EST
- Manuscript Accepted: 5 DEC 2012
- Manuscript Revised: 11 NOV 2012
- Manuscript Received: 5 OCT 2012
- AB was supported by a PWO project from Flanders
- BVDB is funded by the Netherlands Organisation for Scientific Research (NWO; Brain and Cognition Program; 2008-2012)
- EU 7th Framework Program (FP7-Health-2011; BrainAGE; 2012-2017)
We aimed to describe the weight status of obese mothers 6 months after delivery and examine its relationship to important sociodemographical, behavioral, and psychological variables.
Design and Methods
Postpartum data from an interventional trial in obese pregnant women (n = 197), conducted in three regional hospitals, between March 2008 and June 2012, were available from 150 mothers. Obesity was defined as body mass index (BMI) ≥ 29 kg/m2. Predictors were examined from the pregnancy and postpartum period. Descriptive statistics were performed and linear regression models constructed.
Postpartum weight retention (PPWR) 6 months after delivery ranges from −17 to + 19 kg with a mean of −1.28 kg (SD 6.05). Thirty-nine percent showed PPWR (>0 kg) and 13% of obese mothers reported a high PPWR (≥5 kg). Gestational weight gain (GWG) and psychological discomfort were significantly higher in obese mothers with PPWR compared to those with no or low PPWR. Mean duration of breastfeeding in this cohort of obese mothers was 9.5 weeks (SD 8.7), with 17.3% breastfeeding for at least 6 months. At 6 months after delivery, prepregnancy BMI (β = −0.283; P = 0.001), GWG (β = 0.337; P = 0.001), and maternal trait anxiety in the first trimester of pregnancy (β = 0.255; P = 0.001) were significantly associated with PPWR in obese mothers.
PPWR in obese mothers is associated with psychological discomfort during early pregnancy. Besides the importance of adequate prenatal weight management, focused psychological support should be an important cue to action in obese women, to prevent maternal obesity on the long run.
The prevalence of obesity [body mass index (BMI) 30 kg/m2 or more] among women of reproductive age is increasing in most developed countries to worrying percentages. In the region of Flanders (Belgium), 31.7% pregnant women were overweight (21.6%) or obese (10.1%) at the start of pregnancy in 2009, compared to 33% (22.3% overweight and 10.7% obese) in 2011 (unpublished data from the Flemish Study centre for Perinatal Epidemiology (SPE)). Obese pregnant women show more perinatal complications compared to normal weight women [1, 2]. They are also at increased risk of excessive gestational weight gain (GWG). More than 50% of obese pregnant women report an excessive GWG according to the Institute of Medicine (IOM) guidelines [3-5]. Moreover, excessive GWG is significantly associated with a range of unfavorable pregnancy, labor, and delivery outcomes and longer-term weight problems in women and their infants [4, 6].
Postpartum weight retention (PPWR) plays an important role in the development of obesity among women of childbearing age. If women are already obese at the start of pregnancy, high PPWR maintains or even exacerbates maternal obesity , contributing to an increase of transgenerational obesity and its associated health problems. Interpregnancy weight gain—often as a result of PPWR—results in higher odds for adverse pregnancy outcomes (i.e., pre-eclampsia, caesarean delivery, stillbirth, large-for-gestational-age babies) for obese women compared to women whose BMI remained rather stable between two consecutive pregnancies [8, 9]; with linear increases in risk following interpregnancy weight gain.
One Canadian , two American [7, 11] ,and two European studies [12, 13] report excessive GWG to be a significant predictor of PPWR, with measured ranges of PPWR between 1.5 and 18 months. One American study reporting about PPWR over a long time (5 years)  and one from the United Kingdom  indicate socioeconomically disadvantaged young women as well as ethnic minority groups within a Dutch cohort  and American cohort  to be at greater risk for PPWR. Behavioral factors such as breastfeeding, reported in a Danish cohort , and smoking, reported in an American cohort , can have an impact on PPWR, although study results are inconclusive, ranging from negative to weak positive or no significant association at all [13, 18]. An American study from Gunderson et al. indicated that women sleeping five or fewer hours per day at 6 months after delivery showed a threefold increased risk of high PPWR 1 year after delivery . A large Danish prospective cohort study reported psychosocial factors (i.e., feeling distressed/anxious/depressed) during pregnancy to be related with higher PPWR at 6 and 18 months after delivery . Most of these studies were performed in a nonobese population. In this study, we focused on obese mothers, as they are more vulnerable to psychosocial and weight-related problems than normal weight women . We recently showed that attention to the psychosocial context of obese women was effective in limiting GWG and could be of importance in PPWR .
Given the scarse literature on the topic, we aimed to study predictors of PPWR in obese mothers, at 6 months after delivery, in a longitudinal study to provide clues for the design of interventions aimed at preventing interpregnancy weight gain in this high-risk obstetric population. Our hypothesis is that psychosocial (i.e., levels of anxiety and depression) and behavioral (smoking, sleeping, and breastfeeding behavior) variables during the perinatal period influence 6-month PPWR in obese mothers.
We performed a prospective longitudinal cohort study to identify and understand factors that could be used to predict PPWR in obese mothers. The Central Medical Ethics Committee for Human Experimentation of the Faculty of Medicine, Catholic University of Leuven, and the Medical Ethics Committees of the participating hospitals approved the study design. All participants provided written informed consent at study entry before 15 weeks of pregnancy, knowing that they would be followed up until 6 months after delivery.
An interventional trial aimed at reducing GWG and psychological discomfort in obese pregnant women had recruited 197 women. The postpartum longitudinal data were available for 150 of these women (76%). These were included in this study. No difference in mean and categories of PPWR and in levels of psychological variables at baseline (i.e., early in pregnancy) and 6 months after delivery were shown between the three randomized pregnancy groups. We therefore decided that the impact of our pregnancy intervention was extinguished 6 months after delivery, so we could handle this cohort as a rather naturalistic prospective cohort. The detailed inclusion procedure and characteristics of included obese pregnant women were described elsewhere . Briefly, prepregnancy obesity was defined as BMI ≥ 29 kg/m2, according to the IOM criteria of 1990. This cutoff point was used because inclusion started in 2008, before the release of the adapted cutoffs for maternal obesity from 29 to 30 kg/m2. Exclusion criteria were: gestational age > 15 weeks, pre-existing type 1 diabetes, multiple pregnancy, primary need for nutritional advice, and insufficient knowledge of the Dutch language. No specific postnatal care programs for obese women were organized during the time of the study.
PPWR was calculated as maternal weight 6 months after delivery minus prepregnancy weight. Prepregnancy weight was self-reported at the time of study inclusion as descriped before  and maternal weight after delivery was self-reported by questionnaires obtained 6 months after delivery. For descriptive analysis, PPWR was summarized by two categorical variables, defined as weight retention (>0 kg) versus weight loss (including “zero” kg), and high (≥5 kg) versus low (<5 kg) PPWR. This cutoff of 5 kg was based on earlier publications [12, 21, 25]. To perform multivariate analysis, PPWR was used as a continuous variable. We chose to explore PPWR at 6 months after delivery because failure to lose pregnancy-related weight in an appreciable time of 6 months after delivery is an important indicator of obesity in midlife .
Demographic (maternal age, education and parity), socioeconomic (ethnicity, household income), behavioral (smoking behavior, duration of maternal sleep, and breastfeeding), and psychological data (levels of anxiety and depressed mood) were obtained through interviews and self-administered questionnaires at entry into the study, before 15 weeks of gestation, and from self-administered questionnaires, 6 months after delivery. Smoking behavior (yes/no), monthly household income (low, <1,500€; medium, 1,500-3,000€; high, >3,000€), duration of maternal sleep (mean hours/night, < 5 h; 5-7 h, and ≥7 h), and duration of breastfeeding (in weeks) were assessed at 6 months after delivery. Levels of anxiety and depressed mood were assessed at entry into the study and 6 months after delivery by self-administered questionnaires. Medical data were obstained from hospital records. Women self-reported prepregnancy weight and height before 15 weeks of gestation to calculate the prepregnancy BMI. GWG was defined by weight at delivery minus prepregnancy weight.
Anxiety symptoms were measured with the State-Trait Anxiety Inventory (STAI; Spielberger). STAI is designed to study general anxiety and is composed of two self-report scales for measuring two distinct anxiety concepts, state anxiety and trait anxiety. High scores mean more state or trait anxiety. Although a cutoff point for high anxiety has not been properly defined, most studies consider a score above 40 (range 20-80) as being highly anxious, as also used in this study. Feelings of depression were assessed by using the ten-item Edinburgh Postnatal Depression Scale (EPDS). Items are rated on a four-point Likert scale (0-3) and address the intensity of depressive symptoms in the previous 7 days. A cutoff of 13 normally discriminates between minor and major depression. More information regarding these validated instruments was previously published . The Cronbach's alpha analyses for this cohort for state and trait anxiety were high; at least 0.94 for both scales. Cronbach's alpha for EPDS was 0.90, reflecting a high internal consistency. To calculate Cronbach's α, one measurement for each score was taken at random for all women to ascertain independent measurements.
All analyses were performed using the SPSS for Windows statistical software package release 20.0 (SPSS, Inc., Chicago, IL). A two-sided level of significance of 0.05 was defined. The standard alpha value of 0.05 was used, which results in a 5% chance of type 1 error. Based on our hypothesis of relevant covariates, only a priori determined tests were performed. The continuous variables were analyzed using an unpaired t-test. If equality of variance was not assumed, the corrected unpaired t-test was used. Mann-Whitney U test was used if the continuous variables were not normally distributed. Categorical variables were analyzed using the χ2 test, and if cells were too small, the Fisher exact test was calculated.
As the distribution of PPWR was normal (Figure 1) and to use all the available information from PPWR, we decided to perform a multiple linear regression analysis, with the continuous variable “PPWR” as a dependent variable. When checking PPWR for normality, the normal Q-Q plot was a straight diagonal line, and the Kolmogorov-Smirnov test confirmed normality (P = 0.22). Independent variables, based on earlier research and entered in the model to control for confounding and to assess their individual association with PPWR were: GWG, prepregnancy BMI, maternal age and education, parity, smoking behavior, duration of maternal sleep at 6 months after delivery, duration of breastfeeding, feelings of anxiety and depressed mood, ethnicity, and household income. Maternal age, prepregnancy BMI and GWG, levels of anxiety and depressed mood, and duration of breastfeeding were used as continuous variables in the multivariate analysis. All the other covariates were entered creating dummy variables to distinguish different categories.
First, univariate linear regression models were performed to search for a significant influence of all the above-mentioned possible predictors for PPWR individually. The significance level to withhold predicting variables in the model was <0.20. The multivariate model was built with the forward method and the significant predicting variables from the univariate analysis. During model building, we controlled for multicollinearity between predicting variables by scanning a correlation matrix of all the predictor variables and by analyzing the variance inflation factor and tolerance statistic. The final multivariate model was built with the remaining significant predicting variables. To obtain a reliable regression model with sufficient power (0.8), depending on numbers of predictors and a medium expected effect size (r = 0.3), we had to include approximately 150 cases .
We looked at standardized residuals and showed 4% of cases with an absolute value greater than 2 (1.96). When looking at residual statistics, Cook's distance, a measure of the overall influence of the effect of a single outlier on the model as a whole, was <1, indicating that there is no real need to delete any case, since no large effect on the regression analysis is expected. Scatter plots assessed the linearity of the significant predictors for PPWR and showed no nonlinear relations for levels of trait anxiety, prepregnancy BMI, and GWG.
Of the 197 available obese women, 47 did not provide information on maternal weight at 6 months after delivery or after a telephone reminder, so they were excluded. The excluded women (n = 47) differed from those with completed PPW (n = 150) in percentage of preterm birth (12.8% versus 4%, P = 0.04) and duration of breastfeeding (1.9 weeks versus 9.5 weeks, P = 0.001). They also reported significantly more higher levels of anxiety (42.6% versus 23.5%; P = 0.01 for state anxiety; 46.8% versus 23.8%; P = 0.005 for trait anxiety) and depressed mood (27.5% versus 10.9%; P = 0.02) at the time of study entry. Those who were excluded due to absence of reporting their PPW had a higher mean prepregnancy BMI (35.5 versus 34.4 kg/m2) and lower level of education (29.8% versus 46% for bachelor/master degree), but these differences were not statistically significant (Table 1).
|PPW available||PPW not available (excluded)||P|
|n = 150||n = 47||value|
|Maternal age, y (SD)||28.9 (4.1)||29.5 (5.7)||0.72|
|BMI, kg/m2 (SD)||34.4 (4.5)||35.5 (4.8)||0.14|
|Class of BMI||0.30|
|Class I (until BMI 34.99kg/m2)||59.3%||46.8%|
|Class II (BMI 35–39.9 kg/m2)||27.3%||34%|
|Class III (BMI ≥40 kg/m2)||13.3%||19.1%|
|Until 18 year||54%||70.2%|
|Master or higher||6%||2.1%|
|Gestational weight gain, kg (SD)||10.9 (6.8)||12.3 (8.2)||0.23|
|Gestational weight gain (IOM categories)||0.48|
|Less than recommended||19.3%||14.9%|
|More than recommended||56%||66%|
|Preterm birth (<37 weeks) (n) %||(6)4%||(6)12.8%||0.04|
|Macrosomia (≥4,000 g)||14.7%||12.8%||1.00|
|Birth weight, g (SD)||3477 (591)||3348 (561)||0.14|
|Infant weight at 6 months, g (SD)||7851 (985)||7552 (1083)||0.24|
|Induction of labor||20.7%||21.7%||0.84|
|Admission to neonatal unit||11.3%||20%||0.14|
|Breastfeeding longer than 8 weeks||46.7%||5.3%||0.000|
|Breastfeeding at least 6 months||17.3%||0%||0.05|
|Breastfeeding duration, weeks (SD)||9.5 (8.7)||1.9 (4.7)||0.000|
|First trimester of pregnancy:|
|High (≥40) state anxiety||23.5%||42.6%||0.01|
|High (≥40) trait anxiety||23.8%||46.8%||0.005|
|High (≥13) depressed mood||10.9%||27.5%||0.02|
|State anxiety, mean (SD)||34.9 (8)||38.4 (12.7)||0.12|
|Trait anxiety, mean (SD)||35.2 (8.1)||39.6 (12.1)||0.04|
|Depressed mood, mean (SD)||6.8 (4.3)||8.3 (5.8)||0.24|
|6 months postpartum|
|High (≥40) state anxiety, 6 months postpartum, n (%)||22%||33.3%||0.53|
|High (≥40) trait anxiety, 6 months postpartum||27.6%||33.3%||1.00|
|High (≥13) depressed mood, 6 months postpartum||14%||33.3%||0.38|
|State anxiety, mean (SD)||33.7 (11.4)||39 (29.4)||0.77|
|Trait anxiety, mean (SD)||34.3 (12.4)||37 (21.6)||0.96|
|Depressed mood, mean (SD)||6.1 (5.4)||7.7 (10)||0.96|
|Outdoor job time (6 months):||0.59|
|No outdoor job||29.3%||12.5%|
|Part-time outdoor job||31.3%||37.5%|
|Full-time outdoor job||39.3%||50%|
|Duration of sleep (hours)/night:||0.36|
|Less than 5 h||5.3%||0%|
|Between 5 and 7 h||44.7%||25%|
|Equal or more than 7 h||50%||75%|
In the study group of remaining 150 obese mothers, the mean prepregnancy BMI was 34.4 kg/m2 (SD 4.5), mean maternal age was 28.9 year (SD 4.1). The mean GWG was 10.9 kg (SD 6.8, range −9, +30 kg), maternal height 1.66 m (SD 0.06), and maternal weight at 6 months after delivery 93.6 kg (SD 13.6, range 69, 135 kg). Most obese mothers (59.3%) in this cohort were of class I (BMI 29-34.99 kg/m2), 27.3% of class II (BMI 35-39.99 kg/m2), and 13.3% of class III (BMI ≥ 40 kg/m2) before pregnancy (Table 1).
PPWR at 6 months after delivery ranges between −17 and +19 kg with a mean of −1.28 kg (SD 6.05; median −1 kg). Overall, 39% showed PPWR (>0 kg) and 13% of obese mothers reported a high PPWR (≥5 kg). Mean GWG and psychological discomfort (i.e., levels of anxiety and depressed mood) were significantly higher in obese mothers with high and any PPWR compared to those with low and no PPWR. Mean duration of breastfeeding in this cohort of obese mothers was 9.5 weeks (SD 8.7); no significant difference was demonstrated for any duration of breastfeeding between those with PPWR compared to those with low or no PPWR. Those with high PPWR were more likely to sleep less (15.8% versus 3.8%; P = 0.09) and smoked less (5.3% versus 17.7%; P = 0.31) compared to those with low PPWR, although these differences were not statistically significant (Table 2).
|PPWR < 5 kg n = 131 (87%)||PPWR ≥ 5 kg n = 19 (13%)||P value|
|Maternal age, y (SD)||28.9 (4.1)||28.6 (3.9)||0.71|
|Maternal length, m (SD)||1.65 (0.06)||1.66 (0.05)|
|BMI, kg/m2 (SD)||34.7 (4.7)||32.6 (2.8)||0.12|
|Class of BMI, n (%)||0.16|
|Class I (until BMI 34.99 kg/m2)||75 (57.3%)||14 (73.7%)|
|Class II (BMI 35–39.9 kg/m2)||36 (27.5%)||5 (26.3%)|
|Class III (BMI ≥40 kg/m2)||20 (15.3%)||0 (0%)|
|Gestational weight gain, kg (SD)||10.05 (6.6)||16.6 (6.1)||0.000|
|Gestational weight gain (IOM categories), n (%)|
|Less than recommended||28 (21.4%)||1 (5.3%)||0.03|
|As recommended||35 (26.7%)||2 (10.5%)|
|More than recommended||68 (51.9%)||16 (84.2%)|
|Smokers in postpartum, n (%)||23 (17.7%)||1 (5.3%)||0.31|
|Maternal education, n (%)||0.16|
|Until 18 year||73 (55.7%)||8 (42.1%)|
|Bachelor||49 (37.4%)||11 (57.9%)|
|Master or higher||9 (6.9%)||0 (0%)|
|Belgian/Dutch||106 (80.9%)||16 (84.2%)|
|Other||25 (19.1%)||3 (15.8%)|
|Primiparous women||60 (45.8%)||12 (63.2%)||0.22|
|GDM||13 (9.9%)||1 (5.3%)||1.00|
|PIH||14 (10.7%)||3 (15.8%)||0.45|
|PET||8 (6.1%)||1 (5.3%)||1.00|
|Preterm birth (<37 weeks)||6 (4.6%)||0 (0%)||1.00|
|Gestational age (weeks)||39.3 (1.8)||39.6 (0.9)||0.56|
|Macrosomia (≥4,000 g)||19 (14.5%)||3 (15.8%)||1.00|
|Birth weight, g (SD)||3462 (606)||3583 (470)||0.41|
|Infant weight at 6 months, g (SD)||7816 (992)||8099 (923)||0.25|
|Induction of labor, n (%)||28 (21.4%)||3 (15.8%)||0.76|
|Caesarean delivery||37 (28.2%)||4 (21.1%)||0.60|
|Admission to neonatal unit||14 (10.7%)||3 (15.8%)||0.45|
|Breastfeeding initiation||105 (80.8%)||14 (73.7%)||0.54|
|Breastfeeding longer than 8 weeks||59 (45%)||11 (57.9%)||0.33|
|Breastfeeding at least 6 months||22 (16.8%)||4 (21.1%)||0.74|
|Breastfeeding duration, weeks (SD)||9.3 (8.7)||10.9 (9.1)||0.60|
|First trimester of pregnancy:|
|High (≥40) state anxiety, trimester 1, n (%)||30 (23.1%)||5 (26.3%)||0.77|
|High (≥40) trait anxiety, trimester 1||27 (20.8%)||8 (47.1%)||0.03|
|High (≥13) depressed mood, trimester 1||11 (9.9%)||3 (17.6%)||0.40|
|State anxiety, mean (SD)||34.8 (8.3)||35.6 (6.3)||0.44|
|Trait anxiety, mean (SD)||34.6 (7.7)||39.7 (10)||0.04|
|Depressed mood, mean (SD)||6.6 (4.2)||8.2 (4.8)||0.18|
|6 months postpartum|
|High (≥40) state anxiety, 6 months postpartum, n (%)||15 (17.6%)||7 (46.7%)||0.02|
|High (≥40) trait anxiety, 6 months postpartum||17 (20.5%)||10 (66.7%)||0.001|
|High (≥13) depressed mood, 6 months postpartum||9 (10.6%)||5 (33.3%)||0.03|
|State anxiety, mean (SD)||32.4 (11.3)||41.1 (12.5)||0.008|
|Trait anxiety, mean (SD)||32.5 (11.3)||44.4 (14.2)||0.003|
|Depressed mood, mean (SD)||5.5 (5.2)||9.7 (5.7)||0.005|
|Outdoor job time (6 months):||0.69|
|No outdoor job||37 (28.2%)||7 (36.8%)|
|Part-time outdoor job||53 (40.5%)||6 (31.6%)|
|Full-time outdoor job||41 (31.3%)||6 (31.6%)|
|Duration of sleep (hours)/night:||0.09|
|Less than 5 h||5 (3.8%)||3 (15.8%)|
|Between 5 and 7 h||59 (45%)||8 (42.1%)|
|Equal or more than 7 h||67 (51.1%)||8 (42.1%)|
|Low (<1,500€)||9 (7.3%)||1 (6.2%)|
|Medium (1,500–3,000€)||75 (61%)||12 (75%)|
|High (>3,000€)||39 (31.7%)||3 (18.8%)|
Univariate associations with PPWR were significant (P < 0.20) for prepregnancy BMI, GWG, trait anxiety during the first trimester of pregnancy and at 6 months after delivery, depressed mood at 6 months after delivery, parity, household income, and maternal education. Collinearity statistics indicated problems between household income and maternal education (r = 0.39; P = 0.0001 between maternal education, bachelor degree, and highest level of household income). Also a high correlation coefficient between trait anxiety at study entry and trait anxiety 6 months after delivery (r = 0.68; P = 0.0001) and between trait anxiety at study entry and levels of depressed mood (EPDS) 6 months after delivery (r = 0.64; P = 0.0001) was shown. The final multivariate analysis showed that GWG (β = 0.337; P = 0.001) and maternal trait anxiety in the first trimester of pregnancy (β = 0.255; P = 0.001) were significant positive predictors for PPWR at 6 months after delivery. Prepregnancy BMI (β = −0.283; P = 0.001) was a significant negative predictor for PPWR (Table 3). For each kilogram gained during pregnancy, there was a risk of approximately 0.300 kg increase of PPWR at 6 months after delivery. If trait anxiety in obese pregnant women increased by one unit at the start of pregnancy, there was a 0.190 kg increase of PPWR 6 months after delivery. In the multivariate model, all P values were smaller than 0.005 (Table 3), indicating a chance of type 1 error of less than 0.5%.
|PPWR||Unstandardized β||Standardized β||Standard error||95% CI for β|
|Trait anxiety, T1||0.19||0.25||0.05||0.08||0.30|
This article demonstrates an association between psychological discomfort (i.e., levels of anxiety and depressed mood) and PPWR in obese mothers. Considering the association between failure to lose weight 6 months after delivery and excessive weight gain 15 years later , exploring clinically important predictors of PPWR, especially in obese mothers, can give opportunities to develop and examine effects of targeted interventions aimed at reducing childbirth-related weight in this high-risk group. Maternal psychological comfort should be an important issue in counseling obese mothers. We only found one prospective cohort study of pregnant women examining the association between psychosocial factors in pregnancy and weight retention at 6 months after delivery, showing similar results as ours . While women in that study answered questions about anxiety and depression during a telephone interview at approximately 30 weeks of gestation and based on recall of psychological wellbeing for the entire pregnancy, we used validated anxiety and depression scales that were self-administered by the pregnant woman herself.
Higher perinatal risks in obese pregnant women are well described in the literature [1, 2], and knowledge of these risks may induce anxiety in obese women . In an earlier study , we found that obese pregnant women showed significant higher levels of anxiety and depressive symptomatology throughout pregnancy compared to normal-weight pregnany women, and gestational diabetes mellitus (GDM), which is more often prevalent in obese than in normal weight women , is an important contributor to levels of anxiety in this high-risk group . Maternal characteristics like lower levels of education, psychological history and history of miscarriages, higher parity, and being of another ethnicity, which are mostly more prevalent in obese than in normal weight mothers are all predictive factors for higher levels of psychological discomfort during pregnancy . From other studies, we know that pregnant women with elevated levels of stress and anxiety consume more food, such as fats, oils, sweets, and snacks and also have a decreased intake of vitamins . Additionally, the quality of diet in general  or more specific intake of fiber and calcium, decreases with an increasing BMI in pregnant women . This increased psychosocial vulnerability in obese women and the weight-related problems  can be an explanation for the association of high trait anxiety and increased PPWR. Concurrently, altered lifestyle characteristics (including eating behavior) during the postpartum period, influenced by many physical, psychosocial, and emotional changes, may be associated with the mother's mental wellbeing and risk of PPWR.
After adjustment, we showed GWG to be a significant predictor for PPWR. This was in accordance to previous reports [4, 10, 12, 22, 33]; and in cohorts with obese mothers only , even related to BMI after 15-year follow-up . Prepregnancy BMI, on the contrary, was negatively correlated to PPWR, indicating that obese women of class I are more prone to PPWR compared to the extreme and morbid obese mothers (mean PPWR 0.06, −2.5, −4.8 kg respectively; P = 0.001). GWG and psychological discomfort during pregnancy were each independently associated with PPWR in obese mothers. Earlier analysis in our population revealed that 54% obese women had GWG in excess of the IOM recommendations . Furthermore, in this cohort, one in four (24%) and nearly one in three (28%) obese women reported high (≥40) levels of trait anxiety at the onset of pregnancy and 6 months after delivery, respectively. In view of the relatively high prevalence of these significant determinants for PPWR and the well-known associated perinatal risks of a stepped weight gain after delivery [1, 2, 9], intensive and effective psychoeducation for obese pregnant women seems justifiable .
Although hypothesized, we could not find an association between breastfeeding behavior and PPWR in this cohort, which was contrary to the study of Ostbye et al.  and Baker et al. , who reported a positive effect of breastfeeding on postpartum weight reduction in overweight and obese mothers. If obese women exclusively breastfeed for at least 6 months as recommended by the WHO, breastfeeding could potentially eliminate PPWR by 6 months postpartum in women with reasonable GWG (<11 kg) [13, 20]. In our cohort, the mean duration of lactation was 9.5 weeks and only 17.3% (12.6% exclusive) breastfed for 6 months as prescribed by the WHO. This is much less than the 37% (26.5% exclusive) found in a general term cohort of infants from the Flanders Growth Survey . It is known that obese mothers breastfeed for a shorter period than normal weight mothers [36, 37] and have more lactation difficulties .
Although earlier studies found that women reporting shorter sleep duration (≤5 h/24 h) were two to three times more likely to retain at least 5 kg at 3 months , 1 year , and 3 years after delivery , we could not find any significant association in obese women. A limitation here was that the cross-sectional registration of sleep duration at 6 months after delivery did not take into account evolutions since the early postpartum period, quality of sleep, and daytime naps. Smoking behavior had no influence on PPWR in our cohort, as was also the case in a large Danish study .
Study strengths include the prospective longitudinal design in obese women, taking into account psychological variables measured during pregnancy and the postpartum period. The influence of psychological variables on PPWR in obese women is new and an interesting finding targeting effective lifestyle interventions. The limitations to address are the women's self-reported prepregnancy weight and weight 6 months after delivery, as they can be subject to bias. Additionally, we assumed that possible bias in the reporting of prepregnancy weight was similar to possible bias in the reporting of maternal weight 6 months after delivery, so imprecision would be in parallel. We lacked postpartum weight data from 47 (24%) obese mothers who can bias the results. However, analysis of maternal characteristics between those with complete and incomplete data regarding maternal weight at 6 months after delivery gave us an insight in these women's compliance and motivation to report health-related behavior (e.g., reporting their weight status). Those who are least motivated to register their weight after delivery are most vulnerable for higher levels of anxiety and depressed mood, and for quitting breastfeeding earlier, making them vulnerable for higher PPWR. Simultaneously, we need to take into account these differences in sample characteristics when generalizing our results. But as the 47 women with missing PPW had higher levels of anxiety compared to the 150 women with complete data, our results are likely an underestimation of the effects in the total group. The marital/partnered status was not incuded in the multivariate analyses, because only three women in our cohort were single. Perhaps changing psychological variables around the time of delivery and during the early postpartum period may also have predicted PPWR and should be taken into account as missing information when interpreting these results. And lastly, we did not control for interaction terms as there was no anticipated hypothesis about possible relevant interaction terms, and we only included obese mothers. Moreover, assessing all possible interaction terms would inflate type 1 error. Nevertheless, the complexity of associations between psychosocial factors during pregnancy and PPWR in obese mothers limits our understanding of causal mechanisms. We should take this into account when interpreting these results.
In conclusion, besides adequate prenatal weight management, targeted psychological support should be an important aspect of the care to be provided to obese women of reproductive age, to prevent maternal obesity in the long run, as this is an original and interesting finding that has received little attention in the literature so far. Prenatal psychoeducation seems justified to prevent weight gain in obese women ; effects on PPWR should be explored in further randomized controlled trials with obese mothers. Implementing multidisciplinary support and follow-up programs based on psychoeducation and focused on targeted care in obese pregnant women and mothers is an urgent challenge in this obstetrically high-risk group.
The authors thank all the pregnant women for their participation and the gynaecologists and midwives from St Janshospital Genk, Jessa Hospital Hasselt, and St Franciscus Hospital Heuden-Zolder, Belgium, for their cooperation in enrolment of women. The authors also thank M. Mead for the conscious reading of the manuscript and L. Ameye for looking at the statistical analysis.
- 10Gestational weight gain and early postpartum weight retention in a prospective cohort of alberta women. J Obstet Gynaecol Can 2012;34:637-647., , , .
- 24Effects of lifestyle intervention in obese pregnant women on gestational weight gain and mental health: a randomized controlled trial. Int J Obes (Lond) 2012, in press doi: 10.1002/ijo.2012.162., , , , , .
- 27Discovering Statistics Using SPSS, 3rd ed., London, SAGE Publications Ltd., 2009..
- 32Pregnancies complicated by obesity: clinical approach and nutritional management. Verh K Acad Geneeskd Belg 2010;72:253-276., , .