Influence of Psychosocial Factors on Postpartum Weight Retention

Introduction

Postpartum weight retention (PPWR) is common, and the mean weight at 1 year postpartum has previously been found to be 0.5–3 kg higher than the prepregnancy body weight, but with large variation (1,2,3,4,5,6). Several studies have shown that up to one-fifth of mothers retained 5 kg or more 1 year postpartum (1,5,6). This may be a risk factor for overweight 15 years later (3).

The transition to motherhood is a period of social, psychological, and behavioral changes in women's lives. It is expected to be a joyful experience, but is also known to cause a certain amount of distress. In pregnancy and early motherhood, it is common that a woman's mood varies from happiness to sadness and anxiety (7,8). Little is known, however, about whether a woman's mental state during pregnancy may be related to her weight renormalization after birth.

Although depression and anxiety during pregnancy have traditionally been associated with loss of appetite and subsequent weight loss (9), recent studies suggest that depressive symptoms in pregnant women may result in either increased or decreased caloric intake (9,10). However, most studies (9,11,12), but not all (13), have failed to find an association between psychological distress in pregnancy and PPWR. Social support, which can help women to cope with pregnancy and motherhood, may also play a role. The level of perceived social support was found to be inversely related to distress, depression, and anxiety (8,9,12,14), and lack of social support has been associated with weight retention after birth (9,12).

The current evidence on the association between psychosocial factors during pregnancy and PPWR has several limitations. Most previous studies were small, and data on the exposure and outcome were collected at the same time and for the same time period which makes the direction of the association unclear (9,11,12,15). Finally, a woman's lifestyle and social status may be related to her mental well-being during pregnancy and also her risk of weight retention, which further complicate the interpretation of these associations.

In this study, we used data from a large prospective cohort study of pregnant women to examine the association between psychosocial factors in pregnancy and weight retention at 6 and 18 months postpartum, while taking into account important lifestyle factors.

Methods and Procedures

Outcome measures

Table 1 gives an overview of the study variables and the interviews they derived from. A woman's weight retention was measured in whole kilograms and calculated as the difference between her self-reported prepregnancy weight, as reported in the first pregnancy interview and her weight 6 and 18 months postpartum, as reported in the interviews at these time points. To describe a large weight retention, we defined the outcome measure “PPWR ≥5 kg.” This cut point has been frequently used in similar studies (12,18,19).

Table 1. Study variables from the four telephone interviews in the Danish National Birth Cohort

Exposure variables

The main exposures were psychosocial factors based on self-reported information about psychological well-being, psychosocial burdens, social support, social status, and family income.

In the second pregnancy interview, nine questions about psychological well-being covered different aspects of distress, anxiety, and depression with three questions each. The six questions on anxiety and depression came from Symptoms Checklist-92 (20,21), while the three questions about distress came from the General Health Questionnaire 60 (22) (Table 2). For each question, the possible answers were classified into three categories (“no” = 1, “a little” = 2, and “a lot” = 3), which for some were different from the classification in the original instruments. For each woman, a Likert score was created by the sum of scores for each of the three questions for distress, anxiety, and depression, respectively. We found that the score for depression and the score for anxiety had an internal consistency corresponding to a Cronbach's α of 0.7. The six questions on anxiety and depression were therefore combined into a single scale. For distress, we combined the score from the three questions into a separate single scale. For both distress and anxiety/depression, we divided the women into two groups by using a cutoff value closest to the 80th percentile. The same questions about depression/anxiety and distress were included in the interview taken 6 months postpartum. Questions on depression/anxiety were categorized in the same way as described above.

Table 2. Questions about psychological well-being included in the second pregnancy interview: “Have you.…”

Questions on psychosocial burdens were included in the second pregnancy interview where the women were asked if they felt burdened by their “economy,” “housing situation,” or “job” while pregnant. Women were categorized as either being burdened or not burdened. We generated the same three burden variables based on information from the interview 6 months postpartum. The same questionnaire was used, only now in relation to early motherhood.

Information on social support came from the first pregnancy interview where the women were asked how often they were in contact with a family member, either by telephone or in person. The answers were categorized as “several times a week” or “once a week or less.”

Social status was based on information from the first pregnancy interview. Here, the woman gave information about her current or most recent job within the last 6 months. If the woman attended school, she was categorized according to the occupation matching this education. Thus, a nursing student was categorized as a nurse. We then defined two social groups: the highest group included women in management or in jobs requiring higher education, generally >4 years beyond high school. Office workers, service workers, skilled manual workers, unskilled workers, and unemployed women were classified in the “low” group.

In the interview at 18 months postpartum, the women provided information about family's annual gross income, which was categorized as greater or less than 500,000 Ddk (∼100,000 US dollars).

Results

The mean weight retention at 6 months postpartum was 1.4 kg (s.d. = 4.7 kg) with 7,708 women (20.8%) reporting a weight retention of ≥5 kg. At 18 months postpartum, the women weighed on average the same as their prepregnancy weight (0.00 kg; s.d. = 4.8kg), and only 4,492 women (12.1%) retained at least 5 kg. Obese women, however, weighed on average 4 kg less than their prepregnancy weight at 18 months postpartum.

Women who were likely to report symptoms of anxiety/depression or of distress during pregnancy had significantly higher mean weight retention both at 6 and 18 months postpartum compared to other women (Table 3). This was also found for women who felt burdened by psychosocial problems while pregnant. Being in contact with family members once a week or less during pregnancy was associated with higher weight retention, but only at 18 months postpartum. In women with less education as well as in women with a family gross income below 500,000 Ddk, the mean weight retention was higher at 6 months postpartum compared to other women, but no difference was observed at 18 months postpartum.

Table 3. Psychosocial factors measured in pregnancy and weight retention 6 and 18 months postpartum

The mean weight retention 6 months postpartum decreased with increasing age, but this pattern was not present 18 months after birth (Table 4). Primiparous and multiparous women had the same weight retention after 6 months, but at 18 months, multiparous women had the lowest weight retention. Increasing prepregnancy BMI was associated with lower weight retention both at 6 and 18 months postpartum, whereas GWG showed the opposite pattern with increasing retention with increasing gain. Women who quit smoking during pregnancy had a higher weight retention compared to smokers or nonsmokers, while women who exercised during pregnancy seemed to have a slightly lower weight retention. Breastfeeding was associated with lower weight retention at 6 months postpartum, but the association was attenuated at 18 months.

Table 4. Maternal characteristics and lifestyle factors and weight retention 6 and 18 months postpartum

Women with the highest score of symptoms of depression/anxiety or distress during pregnancy had an increased risk of PPWR (retaining ≥5 kg) both at 6 months and 18 months postpartum, even after adjustment for important confounders (Table 5). This was also observed for women who felt burdened by financial troubles or by their work situation. For women with a family annual gross income <500,000 Ddk, the increased risk of PPWR was almost the same after 6 and 18 months, while for women with less education, the risk after 18 months was somewhat higher than after 6 months. When all psychosocial factors were mutually adjusted, the odds radios were all somewhat attenuated except for the one for educational level. When adjustment for GWG and length of breastfeeding were carried out in an additional analysis, it only attenuated the observed observations marginally.

Table 5. Psychosocial factors in pregnancy and risk of PPWR ≥5 kg 6 and 18 months

Compared to women who had low scores for depression/anxiety or distress both during pregnancy and in early motherhood, the risk of PPWR was increased in women who had high scores in only one of these periods, and the highest excess risk was found in women who had high scores in both periods (1). These findings were present at both 6 and 18 months postpartum, however, with a tendency for stronger associations after 18 months. Also, women who felt depressed/anxious or distressed during pregnancy tended to have greater weight retention at 6 months than women who only reported these conditions in the postpartum period. This pattern was not present 18 months postpartum.

For women who felt burdened by psychosocial problems, a similar pattern was less pronounced, and only 18 months postpartum did women who felt burdened in both periods have the highest risk of PPWR (2). Although women with financial troubles in any period had significantly higher risk of PPWR than women without any troubles, this tendency was less pronounced in women with a troubled housing situation. Also, women with a troubled work situation during pregnancy had had higher PPWR than women who only reported work-related problems in early mother hood.

Discussion

In this study, we found that women who felt depressed/anxious or distressed during pregnancy had significantly higher PPWR than other women. Similar findings were found in women who felt burdened by their economic or working situation, who were in the lowest income group, or who had the least education. Notably, the highest risk of PPWR after 18 months was observed in women who were not feeling well psychosocially, neither during pregnancy nor in early motherhood.

Linné et al. found that weight retention at the end of the postpartum year predicted future overweight 15 years postpartum (3). Considering the obesity epidemic, it may be an important issue for health-care professionals to be aware that a woman is in higher risk of gaining excessive weight related to childbearing if she is burdened by psychological or social problems.

Our observations indicate that high levels of distress during pregnancy may occur more frequently among women in the lowest social groups, which is in accord with the findings of others (5,8,11,12,24,25). It may be that feeling burdened, distressed, or anxious during pregnancy elicits a stress response in the hypothalamic pituitary, adrenal, and sympathetic nervous system, that may eventually result in general and visceral obesity as suggested originally by Björntorp (26). Alternatively, the distress may simply be associated with overeating as a comfort, and hence lead to larger energy intake and higher weight gain (15).

By virtue of its large size, its prospective design and its plenitude of information, the DNBC provides a unique opportunity to study the association between psychosocial factors and PPWR. The participation rate of the women invited to the DNBC was about 60%. Nohr et al. (27) found that the women in the cohort were somewhat healthier than the source population, but they were generally representative for white women and able to provide a sufficient exposure contrast for internal comparisons. To avoid problems seen in previous studies with interpretations of causal directions, we chose to focus on psychosocial factors measured during pregnancy, when the woman did not know her weight retention.

Our study has, however, also some limitations. Measures of psychological well-being were taken from the Symptoms Checklist-92 and the General Health Questionnaire 60, but not all the original items could be included as the questionnaires used in the DNBC had to cover a wide range of exposures. Also, the women in the DNBC were asked to recall psychological well-being for the entire pregnancy at ∼30 weeks of gestation and for the postpartum period 6 months after birth. If the period of recall had been shorter, maybe more women would report having been exposed. On the other hand, the results show that our measures of psychosocial well-being were able to detect a difference in PPWR. Although all the main psychosocial variables came from the second pregnancy interview, information about social status, family contact, and income were collected at other points in time which should be considered when interpreting the results. Finally, the results relied on the woman's self-reported weight that is subject to recall error and perhaps wishful thinking. Women with large weight retention may have had a tendency to underestimate body weight, but we assume that any bias in reporting of prepregnancy weight were similar to biases in the reporting of GWG and weight postpartum. If this is true, the differences between prepregnancy weight and postpartum weight may be relatively unbiased (28).

The complexity of the relationship between psychosocial factors during pregnancy and PPWR limits our understanding of causal mechanisms and calls for caution when interpreting the results. GWG is a strong risk factor for PPWR, also in this study where it far overruled the effects of prepregnancy BMI. However, we judged GWG to be on the causal pathway between psychosocial well-being and PPWR. Thus, we did not adjust for it in the main analyses. Yet, the information about psychosocial factors was collected when the woman was about 30 weeks pregnant where a large GWG may have influenced her mental state. A similar situation applies to breastfeeding, which we considered to be an intermediate factor, yet it may also be a risk factor for psychosocial problems postpartum, e.g., if a woman is not able to breastfeed her child. In an additional analysis, we adjusted our findings for GWG and breastfeeding, although it may be considered overadjustment. We found that these factors explained a small part of the observed associations.

It is likely that the psychosocial factors that we measured were highly correlated. Thus, a low social status may explain why a woman feels burdened or psychologically distressed during pregnancy, and likewise distress, may explain if a woman is more easily burdened. We included all psychosocial factors in an additional analysis and found that while all associations were attenuated, the factors most closely related to social status, such as financial situation and education, remained the strongest risk factor for PPWR. This suggests that social inequity, to some extent, may explain the findings.

Feeling distressed, depressed, or anxious during pregnancy led to higher PPWR as did personal and economical burdens. Adverse psychosocial exposures were relatively common, while PPWR at 18 months were seen in 12%. Thus, even small relative risks associated with psychosocial adversity may have a significant public health impact if PPWR is associated with overweight in later life. Adverse psychosocial characteristics may be a common determinant of weight retention after childbirth.

DISCLOSURE

The authors declared no conflict of interest.