1. Introduction
Breastfeeding is a key strategy to improve maternal and child health. The benefits for infants include the reduction of risk of infectious disorders, obesity, allergic disorders, and developmental delays [
1,
2]. For mothers, breastfeeding could protect against postpartum weight retention, cardiovascular diseases, and malignancy, including breast, ovarian and endometrial cancers [
3].
Although the World Health Organization (WHO) and the American Academy of Pediatrics (AAP) have recommended exclusive breastfeeding for the first six months of life [
4,
5], recent data suggest that less than 37% of women could fulfill this goal worldwide [
2]. Factors associated with the low breastfeeding rate are multifaceted. Potential risk factors for breastfeeding cessation are maternal smoking behavior, delivery mode, parity, dyad separation, maternal educational status [
6], paternal support [
7], maternal perception of lacking sufficient milk supply, mastitis, infants’ failure to thrive [
8], predelivery breastfeeding education [
9], and maternal obesity [
10].
The epidemic of maternal obesity is a major public health concern today [
11]. More than 30% of American women of reproductive age are obese [
12]. Overweight and obese women may not only have a greater risk of adverse perinatal outcomes [
13], but may also have more difficulty in continuing to breastfeed after delivery, due to a decreased prolactin secretion response to suckling [
14] and delayed lactogenesis II [
15]. However, observational studies have been conducted, to determine the possible associations between pre-pregnancy body mass index (pBMI) and breastfeeding practices, and the evidence remains inconclusive [
16,
17,
18,
19,
20,
21]. Less is known about the effects of maternal pre-pregnancy underweight status on breastfeeding behavior. In one previous study, mothers with pre-pregnancy underweight status had a higher risk of any breastfeeding cessation within two months postpartum [
21]. However, Giovannini et al. reported no difference between underweight and normal-weight mothers for the duration of breastfeeding [
22]. The evidence of this association in Asian populations is scant, and there is not enough data to show the relationship between pBMI and breastfeeding practices. Only a few studies conducted in Japan and China have reported that pre-pregnancy obesity [
17,
23] and underweight [
21] may have a negative effect on breastfeeding duration.
The inconsistency in previous literature might result from differences in study designs, study populations, sample sizes and various breastfeeding outcome definitions. In our previous study, the prevalence of underweight was greater in Taiwan than in the United States and European countries [
11,
24]. It is unclear whether pre-pregnancy underweight status may have a certain effect on breastfeeding behaviors from a population with a higher proportion of women who are underweight. Therefore, to address these knowledge gaps and study limitations in previous literature, we conducted a secondary analysis study by using data from a prospective birth cohort in Taiwan, to explore the association between pBMI and breastfeeding practices.
4. Discussion
In the present study, our findings indicated that maternal obesity was associated with a shorter duration of breastfeeding compared with normal weight, and obese mothers also had a greater risk of breastfeeding cessation at 2 months postpartum. In the ordinal logistic regression model, maternal underweight and obesity were associated with a shorter duration of breastfeeding. Moreover, underweight mothers were less likely to sustain breastfeeding, and the risks of early breastfeeding cessation postpartum at different time points were consistent in the adjusted logistic regression model and subgroup analysis stratified by parity. Our results using data from a nationally representative cohort in Taiwan indicated that maternal underweight and obesity may be important risk factors for breastfeeding practices.
Our study findings are consistent with those of previous studies, demonstrating that maternal obesity was inversely associated with breastfeeding duration [
16,
19,
20,
23,
30]. Obese women were at greater risk of an inability to sustain any breastfeeding at 2 months postpartum. In our cohort, the obese rate is lower in primiparous mothers. However, the adjusted odds ratios for obese women to stop breastfeeding were not the same, and the aOR was greater in primiparous women. A potential explanation could be contributed to more breastfeeding experiences in multiparous mothers. In contrast to the effect of obesity, underweight women had an extensively greater risk of failure to continue breastfeeding at 2, 4, and 6 months postpartum.
Breastfeeding behavior is complex. It is well established that maternal obesity may contribute to many adverse perinatal outcomes, including preterm birth [
31] and cesarean delivery [
32], which are common risk factors for unsuccessful breastfeeding. Obese women may have other negative characteristics, such as lower socioeconomic status or educational level, that make it challenging for them to continue breastfeeding [
33].
Less is known about the association between underweight and breastfeeding. In contrast to most of the previous studies on obesity and breastfeeding in Western countries [
16,
19,
20,
30], our findings indicated that maternal pre-pregnancy underweight also has a negative effect on the duration of breastfeeding. In a systematic review and meta-analysis study, Huang et al. [
10] reported that maternal pre-pregnancy underweight status was associated with the initiation of breastfeeding only in sensitivity analysis involving studies with sample sizes of less than 500. In addition, only one Chinese study was enrolled in the meta-analysis study for investigation [
17]. Since the prevalence of underweight is low in most Western countries [
11], the effect of underweight on breastfeeding duration may need more studies to clarify the possible influences of pBMI from different countries and populations.
The potential explanations for the inconsistent results in previous literature may include differences in study designs, definitions of breastfeeding, definitions of pre-pregnancy BMI, adjusted confounding factors, race/ethnicity and sample sizes. Zhu et al. [
21] reported that underweight before pregnancy increased the risk of termination of any breastfeeding within 2 months postpartum, while another study by Giovannini et al. [
22] showed no association between underweight and breastfeeding duration. It is important to note that these two studies defined underweight mothers according to different criteria (Zhu: <18.44 kg/m
2; Giovannini: <19.8 kg/m
2). In another study, Guelinckx et al. [
20] found that underweight in pre-pregnancy was associated with low intentions and low initiation of breastfeeding, and they defined underweight status according to the WHO classification (<18.5 kg/m
2), which is the same definition we used. However, they only enrolled 50 women in each category, and could not assess the risk by a comprehensive adjusted model, due to the small sample size. An insufficient number of participants may limit the ability of the evidence to unveil the effect of underweight on breastfeeding practices, and to fully investigate this issue at the population level. Currently, the evidence of the association between maternal pre-pregnancy underweight and breastfeeding duration is still inconclusive. Researchers in the abovementioned studies all applied the same measurement as self-reported body weight and measured body height for pBMI calculation [
20,
21,
22]. Moreover, women tend to underestimate their body weight before pregnancy [
34], which may bias the effect of pre-pregnancy underweight on the duration of breastfeeding. In the future, consistent and reliable definitions of pBMI and criteria for BMI statuses should be emphasized, to compare the differences between each study accurately.
The potential mechanisms for lower breastfeeding rate in obese women might be due to decreased prolactin secretion response to suckling [
14] and delayed lactogenesis II [
15]. However, the underlying mechanism behind underweight and breastfeeding behavior remains unknown. The factors contributing to breastfeeding behaviors are multifaceted, and possible reasons could be considered for the low breastfeeding rate among underweight women. Underweight women may have more unfavorable background characteristics to continue breastfeeding [
9]. In our study, 62.4% of underweight women were primiparous; multiparous women may have more experience with breastfeeding, and parity may be another important contributor to breastfeeding. However, in our subgroup analysis by parity, the effect of underweight persisted, which means that there might be more unmeasured risk factors for evaluation. Notably, regarding the possibility of a higher prevalence of low-birth-weight (LBW) infants among underweight women in the TBCS cohort [
24], these vulnerable babies are also at an increased risk of dyad separation after birth and early breastfeeding cessation, and these interactions among SGA, LBW and underweight mothers may contribute to early breastfeeding termination [
9,
35].
Strengths and Limitations
To our knowledge, the present study is the first to comprehensively investigate the association between pBMI and breastfeeding duration in a large-scale population cohort in an Asian population. Furthermore, in contrast to previous literature, the main strength of our study is its focus on differences in the effects of underweight on breastfeeding behaviors. The richness of information also allows us to analyze the contribution of pBMI in the multivariable adjusted model, stratified by primiparous and multiparous mothers.
Our study has several limitations. First, information on pBMI and breastfeeding duration was obtained from self-reported data. Although previous studies have indicated that maternal recall for pre-pregnancy weight, height and GWG were reliable [
36,
37] and that maternal recall for breastfeeding duration was also reliable after 3 years postpartum [
38], recall bias may still exist, and the misclassification of outcomes should be considered. Given that bias of outcome misclassification is usually toward the null, our findings may underestimate the impact of pBMI on breastfeeding behaviors [
39]. Second, we could only provide information on any breastfeeding duration, and not on exclusive breastfeeding, which is recommended by the WHO [
4] and the AAP [
5]. Given the restrictive definition of exclusive breastfeeding [
27], we could not determine from the information in the questionnaires whether infants consumed any liquids or food in addition to breastmilk or formula. We can only define the outcome as any breastfeeding duration in our analysis. Third, prenatal data for maternal breastfeeding plans, detailed dietary behaviors, employment status before 6 months postpartum, treatment and diagnosis for dyad separation [
6], paternal support [
7], maternal perception of lacking sufficient milk supply, mastitis, infants’ failure to thrive [
8], predelivery breastfeeding education [
9] and psychological factors may influence the initiation and duration of breastfeeding. These factors might contribute to the causality between pBMI and breastfeeding duration. Given that our study design was an observational study, causality for the association could not be discussed in detail. Additional studies to explore the causal mechanisms between them are warranted. Fourth, our data were collected from 2006 to 2007; therefore, our results might not be fully generalized to current clinical practices.
Considering that improvements in breastfeeding rates may benefit both mothers and children, and that current breastfeeding rates are still not sufficient worldwide, our research has important public health implications to encourage healthy weight before conception, especially for underweight women of reproductive age, which may help them maintain breastfeeding for a longer period of time postpartum. Healthcare providers should raise concerns about the threat of low breastfeeding rates, regarding the long-term impact on maternal and child health. Maternal undernutrition and obesity remain significant global health issues in women of reproductive age. Nutritional status before conception has major implications for the long-term health outcomes for mothers and children [
40]. Evidence-based maternity care practices and policies could help to improve breastfeeding [
41]; however, underweight and obese mothers might need additional and special assistance to increase their willingness to initiate breastfeeding and their success of continuing breastfeeding for longer periods. Optimizing maternal BMI during the pre-conception period is essential, and nutrition-specific interventions should be considered, to improve maternal and child health outcomes in many areas.