Methods

We used data from the Infant Feeding Practices Study II (IFPS II)¹⁷ that was collected by the Centers for Disease Control and Prevention and the US Food and Drug Administration between 2005 and 2007. In the IFPS II, a longitudinal survey was distributed to a national consumer opinion panel of 500,000 households. The survey comprised 1 prenatal and 1 neonatal questionnaire and a series of 9 postnatal questionnaires; all questionnaires were sent via regular mail. A brief birth screening telephone interview was administered to achieve the sample goals of healthy-term or near-term singleton newborns. Of the 4900 respondents in the IFPS II study, approximately 3000 answered the neonatal questionnaire. In total, 2560 responded to the question pertaining to the type of feeding at hospital or birth center discharge (herein “hospital discharge”). Questionnaires that were either missing data about their MOD or other covariates adjusted in the multivariable model were excluded.

The main outcome of interest was exclusive breastfeeding at hospital discharge. The questionnaire asked, “When you left the hospital or birth center, how were you feeding your baby?” Options included “breastfeeding only,” “formula feeding only,” or “both breast and formula feeding.” A dichotomous exclusive breastfeeding variable was created by coding breastfeeding only responses as “yes” for exclusive breastfeeding and formula feeding only or both breast and formula feeding responses as “no” for exclusive breastfeeding.

The primary independent variable was MOD, which was coded as a dichotomous variable. The questionnaire asked “How was your baby delivered?” Responses indicating vaginally and not induced or vaginally and induced were categorized as “delivering vaginally,” and responses indicating a planned cesarean delivery or an unplanned or emergency cesarean delivery were categorized as “cesarean delivery.”

Covariates included mother’s age, marital status, race or ethnicity, education, prenatal smoking, gestational age, neonatal intensive care unit (NICU) stay, previous cesarean delivery, and prenatal intention to exclusively breastfeed. The percentage of federal poverty level in IFPS II was calculated using the household size and income. We calculated body mass index (BMI) using the formula BMI=(weight, lb/[height, in])²×703. To determine the gestational weight gain during pregnancy, we used the Institute of Medicine’s revised cutoffs for gestational weight gain: underweight (BMI<18.5; recommended weight gain: 12.5-18 kg), normal weight (BMI=18.5-24.9; recommended weight gain, 11.5-16 kg), overweight (BMI=25.0-29.9; recommended weight gain, 7-11.5 kg), and obese (BMI≥30; recommended weight gain, 5-9 kg).¹⁸

Results

Overall, 7 respondents were excluded because of missing data about their MOD, and 1059 respondents were excluded because of other covariates adjusted in the multivariable model, resulting in a final sample of 1494 respondents. The overall characteristics of the study sample are displayed in Table 1. Of 1494 respondents, 1117 (74.8%) gave birth vaginally and 377 (25.2%) gave birth by cesarean delivery. Among the overall respondents, 1159 (77.6%) exclusively breastfed at hospital discharge. A total of 666 respondents (44.6%) had a normal BMI, and 698 (46.7%) had an above-average gestational weight gain according to the Institute of Medicine’s revised guidelines for weight gain during pregnancy.

Table 2 includes the bivariate association between the characteristics of the study sample by the newborn feeding status. The 266 respondents who delivered by cesarean birth (23.0%) were less likely to exclusively breastfeed at hospital discharge compared with the 893 respondents who delivered vaginally (77.1%) (P<.01). Additionally, the 962 respondents (83.0%) who stated that they had prenatal intentions to exclusively breastfeed in the first few weeks were significantly more likely to exclusively breastfeed at hospital discharge compared with the 197 respondents (17.0%) who had no intention (P<.01). Exclusively breastfeeding at hospital discharge differed by status of previous cesarean delivery. The 258 respondents who had a previous cesarean delivery (22.3%) were less likely to exclusively breastfeed than the 901 respondents who never had a previous cesarean delivery (77.7%) (P=.037). With the exception of income, gestational age, and NICU stay, all other covariates were statistically significantly associated with exclusive breastfeeding at hospital discharge.

The unadjusted and multivariable adjusted odds of exclusively breastfeeding are shown in Table 3. After adjusting for potential confounders, respondents who delivered by cesarean birth were 59% less likely to exclusively breastfeed at hospital discharge compared with respondents who delivered vaginally (OR, 0.41; 95% CI, 0.24-0.71). The subgroup analysis shows that among respondents who delivered by cesarean birth, no statistically significant differences in the odds of exclusive breastfeeding exist between those who had a planned cesarean delivery and those who had an unplanned cesarean delivery (unadjusted OR, 1.03; 95% CI, 0.61-1.74) (multivariate adjusted OR, 0.99; CI, 0.46-2.10).

Discussion

Among the respondents who participated in IFPS II, cesarean delivery was negatively associated with exclusive breastfeeding at hospital discharge. This association was independent of age, marital status, education, income, BMI, gestational weight gain, prenatal smoking, gestational age, NICU stay, previous cesarean delivery, and prenatal intent to exclusively breastfeed. However, among those who delivered by cesarean birth, the odds of exclusive breastfeeding at hospital discharge did not differ by the type of cesarean delivery, indicating that any cesarean delivery (planned or unplanned) was negatively associated with exclusive breastfeeding at discharge.

Our findings are consistent with and add to existing literature that has reported a negative association between cesarean delivery and breastfeeding.^11,13 The mechanisms for having lower odds of exclusive breastfeeding at hospital discharge could be multifactorial, including hormonal, social, or procedural causes.^19-22 For example, several studies have shown that certain types of anesthesia can affect lactation.^23-25 Other possible causes include delayed breastfeeding initiation after cesarean delivery, maternal-newborn separation, and possible delayed lactogenesis.^26-28

The 2 widely known hormones that play a major role in lactation are the pituitary hormones prolactin and oxytocin.²⁹ Milk production is controlled by prolactin levels, and milk ejection occurs in response to a surge in oxytocin.²⁹ Prolactin periodically surges to stimulate many processes during pregnancy, including lactation and breast tissue development.³⁰ Several studies have explored the specific roles and mechanism of action of the hormone prolactin during pregnancy and postpartum period.^31-33 After delivery of the placenta, the inhibitory factors of prolactin decrease, including progesterone and other placental hormones.³⁴ Further investigation of the biochemical mechanism of action associated with MOD on hormonal changes that affect lactation is important.

The current study found other factors that showed statistically significant associations with increased odds of exclusive breastfeeding at hospital discharge. These factors included race or ethnicity, education, gestational age, and prenatal intention to exclusively breastfeed, all of which have been previously noted.^26,35-37 A strong positive association was found between maternal prenatal intention to breastfeed and breastfeeding at hospital discharge (OR, 11.03; 95% CI, 8.19-14.84), supporting the need for education about breastfeeding. Several studies evaluating the effectiveness of prenatal breastfeeding education have found that prenatal breastfeeding education has a potential to improve breastfeeding rates.^38-41 Providing extra prenatal education for mothers who are planning to have a cesarean delivery could improve exclusive breastfeeding by increasing their intentions to breastfeed. Mothers need to be aware of the possible effect cesarean deliveries have on breastfeeding before making a decision. Early postnatal breastfeeding education and assistance for women who have either a planned or emergency cesarean delivery can also support improved breastfeeding outcomes.

A limitation of the current study is that the IFPS II data were collected by women volunteering to complete a mailed survey, which makes it susceptible to volunteer bias, recall bias, and selection bias. This problem is further exacerbated by a loss of 41% of the eligible study sample due to incomplete information. The final analytical sample statistically significantly differed from those with incomplete data by key sociodemographic characteristics, indicating that the results may not be generalizable to the IFPS II study sample (Table 4). However, IFPS II is not a nationally representative dataset and may not produce results that can be generalized to all newborns, pregnant women, and new mothers in the United States. In addition, IFPS II data did not make any distinction between “exclusive breastfeeding at the time of maternal discharge” and “exclusive breastfeeding at the time of infant discharge.” The lack of this distinction could lead to a potentially misleading assumption that newborns who were discharged after their mothers would have the same access to breastfeeding as those who were discharged with their mothers. Likewise, the dichotomous variable of NICU stay (≤3 days) assumes that newborns who spent 3 days or fewer in the NICU could have similar access to breastfeeding as newborns who did not spend any time in the NICU.

The dataset also lacks important factors like timing of breastfeeding initiation. Furthermore, the cross-sectional design does not allow inference regarding causal relationships. In addition, given that this research is based on a secondary analysis of nearly decade-old data, the results should be interpreted with caution because guidelines for breastfeeding mothers, including those who have had a cesarean delivery, have improved after 2005 at most labor and delivery centers as a result of consumer demand, The Joint Commission, the Surgeon General’s Call to Action to Support Breastfeeding, the Centers for Disease Control and Prevention’s Maternity Practices in Infant Nutrition and Care, the World Health Organization, and the United Nations International Children’s Emergency Fund’s Baby-Friendly Hospital Initiative. A large national sample size, the diversity of the women who responded to the survey, and our ability to adjust for as many confounding variables as possible constitute some of the strengths of the current study.

Breastfeeding provides many nutritional, immunologic, and health benefits for newborns and infants.⁹ Regardless of the type of delivery, mothers and newborns should be able to take advantage of this natural form of nutrition. Future research should focus on identifying specific causes of poor breastfeeding rates in women who undergo cesarean delivery. Further research is needed to understand the mechanism of lactation among women who have undergone a cesarean delivery as well as the role of conditions surrounding unplanned cesarean delivery. Understanding this mechanism is important because it can lead to the development of techniques to improve exclusive breastfeeding among these mothers.

Conclusion

The current study demonstrated lower rates of exclusive breastfeeding at hospital discharge among women who underwent cesarean delivery compared with those who underwent vaginal delivery. More effort is needed to improve early initiation of breastfeeding in women who have cesarean deliveries, including developing and implementing hospital policies that support breastfeeding, training health care professionals to support breastfeeding after cesarean delivery, and providing prenatal and postnatal education and counseling on the importance of breastfeeding.