Introduction

Caesarean sections are a major obstetrical surgery that can save lives when medically indicated. The World Health Organization (WHO) recommends a Caesarean rate of 10–15% (WHO, 2015), yet rates of Caesarean section vary widely around the globe and have gradually risen over the past several decades (Betrán et al., 2016). In Canada, Caesarean section increased from 18.7% of births in 1997 to 29.9% of births in 2019 (Canadian Institute for Health Information, 2020). This procedure is not without short- and long-term risks for maternal and newborn health. Caesarean delivery is associated with higher risk of maternal mortality and morbidity (e.g. hemorrhage, uterine rupture, urinary infection; Korb et al., 2019) and neonatal respiratory distress (Li et al., 2019). Longer-term risks include greater likelihood of complications in subsequent pregnancies and health conditions in children, including asthma and obesity (Keag et al., 2018).

Research has consistently shown that Caesarean delivery negatively impacts breastfeeding; women delivering by Caesarean have lower odds of initiating breastfeeding and generally discontinue breastfeeding earlier than women who give birth through vaginal delivery (Prior et al., 2012). Some evidence suggests this disparity is specific to planned Caesarean, while others suggest both planned and unplanned Caesareans can impede breastfeeding efforts (Cohen et al., 2018; Hobbs et al., 2016; Prior et al., 2012). Caesarean surgery can place high stress on both the mother and infant, and post-operative recovery is often characterized by maternal pain, limited mobility, and separation from the infant to encourage mothers to rest and heal (Consales et al., 2020; Tully & Ball, 2014). Mothers who undergo planned or unplanned Caesarean delivery are less likely to experience skin-to-skin contact immediately after delivery, to breastfeed their infants within an hour after giving birth, and to share a room with their infant after delivery compared to mothers who give birth vaginally (Prado et al., 2018).

Breastfeeding and lactation difficulties are the most frequent reasons women cite for stopping breastfeeding earlier than desired, of which the most common in the early postpartum period are difficulties with infant latch, milk supply, suboptimal infant weight gain, and underlying medical conditions (Li et al., 2008; Odom et al., 2013). In Canada, the most common reasons for early breastfeeding discontinuation are insufficient milk production and difficulty with techniques in breastfeeding, such as infant positioning and latch at the breast (Public Health Agency of Canada, 2018). Yet few studies have examined how the presence of breastfeeding difficulties differs in the context of a vaginal versus Caesarean birth. For example, Hobbs et al. (2016) reported descriptive data on the prevalence of breastfeeding difficulties varying by mode of delivery. Compared to mothers delivering vaginally or by planned Caesarean, a higher proportion of mothers who delivered by unplanned Caesarean reported infant difficulties such as trouble latching and other difficulties such as not producing enough milk. However, additional research that accounts for the severity of difficulties and potential for confounding bias is needed. Moreover, the timing of difficulties is important to study, as challenges that present shortly after birth may undermine early maternal confidence to breastfeed and pose the greatest detriment to breastfeeding establishment if not promptly addressed with hospital and public health support. Evidence on differences in early breastfeeding difficulties based on mode of delivery may help to identify interventions that can reduce feeding disparities in mothers delivering by Caesarean. Therefore, our objective was to explore the association between Caesarean delivery and types of breastfeeding difficulties experienced by mothers of full-term infants during the delivery hospitalization.

Methods

Study design

We conducted a secondary analysis of data from the Happy Baby Study, which was a prospective community-based cohort study in Calgary, Alberta, Canada (Chaput et al., 2016). The objective of this parent study was to assess the association between breastfeeding difficulties and postpartum depression risk and the moderating role of breastfeeding support. Data on demographic characteristics, peripartum events, breastfeeding difficulties and outcomes, mental health, and health service use were collected via self-report questionnaires at three time points postpartum: the delivery hospitalization, 6 weeks postpartum, and 6 months postpartum. This analysis only used cross-sectional data from the delivery hospitalization to focus on breastfeeding difficulties that emerge immediately after birth when in-hospital and early community-based intervention is most readily available through Canada’s universal healthcare and public health system.

Sample

Convenience sampling was used to recruit women from postpartum units at all tertiary care hospitals providing delivery care in the Calgary Health Region between June and October 2010. All of the hospitals had integrated some or all of the steps from the WHO Baby-Friendly Hospital Initiative; however, none were accredited at the time of the study. Research assistants approached women and invited them to participate in the study within 72 h of giving birth. Women were eligible if they were aged 18 or older; delivered a full-term, singleton infant; and planned to breastfeed. Exclusion criteria were prenatal plans to formula-feed, inability to communicate in English, and a neonatal intensive care unit stay > 24 h. Overall, 446 women enrolled in the study and completed the first questionnaire, and 360 women completed all 3 questionnaires. Our sample for this analysis excluded participants with incomplete data on mode of delivery (n = 23) or incomplete data on breastfeeding difficulties (n = 5), resulting in an analytic sample size of 418.

Mode of delivery

The exposure was mode of delivery, measured categorically as vaginal, planned Caesarean, or unplanned Caesarean (with or without spontaneous labour). Planned Caesarean was defined as Caesarean section based on the participant’s request or advice of the health team due to medical indications prior to the onset of labour, which occurred on a scheduled date. Unplanned Caesarean (also termed emergency) was defined as Caesarean section that was not scheduled prior to the delivery hospitalization and was due to peripartum complications.

Breastfeeding difficulties

The outcome was presence of breastfeeding difficulties measured using the Breastfeeding Experiences Scale (BES; Chaput et al., 2016; Wambach, 1997). The BES is a validated, 18-item measure for breastfeeding difficulties with a 5-point rating scale for each difficulty’s experienced severity ranging from “not at all” (0) to “unbearable” (4). For this analysis, individual difficulties were grouped together into types based on clinical similarities in the nature of the difficulty; the grouping of difficulties into six types is outlined in Table 1. A composite variable was developed for each difficulty type based on the presence of one or more pertinent difficulties rated as moderate, severe, or unbearable.

Table 1 Grouping of breastfeeding difficulties from the Breastfeeding Experiences Scale (BES) into types
Full size table

Covariates

We developed a directed acyclic graph (DAG) using literature searching and subject matter expertise to identify potential confounders for this analysis (Bandoli et al., 2016; Supplemental Figure 1). Continuous covariates were body mass index (kg/m2) and maternal age at time of delivery (years). Dichotomous covariates were psychosocial vulnerability, socioeconomic status, and parity. Psychosocial vulnerability was defined as the presence of depression (a score of ≥ 10 on the Edinburgh Postnatal Depression Scale) and/or low social support (dissatisfaction with support received from partner, family, or friends). Socioeconomic status was measured using a composite variable developed previously to capture confounding related to breastfeeding in the Happy Baby Study, defined as experiencing ≥ 2 of the following: not owning a home, being single, annual household income < $60,000 (the threshold for subsidized housing at time of recruitment), and/or having a low education (high school or less; Chaput et al., 2016; Braveman et al., 2005). Parity was defined as the number of live births and categorized as primiparous (no previous births) or multiparous (one or more previous births). Gestational age was controlled for by restricting the sample to full-term birth (37 weeks gestation or greater) at the time of data collection. Data on pregnancy complications were not collected.

Statistical analysis

We used descriptive statistics including proportions, means, and standard deviations to summarize participant characteristics by mode of delivery. We then used multivariable logistic regression modelling to quantify the association between mode of delivery and types of breastfeeding difficulties shortly after birth, using vaginal delivery as the referent group. Composite variables for each of the six difficulty type groupings were used to ensure adequate statistical power, due to low cell counts for some individual difficulties. Models were adjusted for confounders identified using the DAG. Missing covariate data were handled using multiple imputation. Statistical significance was defined as results with a 95% CI that did not include a null value of one (corresponding to a p value < 0.05). All data cleaning and analyses were completed using Stata/IC 13.1.

Ethics approval

The Happy Baby Study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (REB-ID 21754).

Results

Sample characteristics

Overall, 37.1% of women in the Happy Baby Study had a Caesarean delivery, of which most were unplanned. Table 2 displays sample characteristics by mode of delivery. The majority of participants had postsecondary education, had a household income ≥ $60,000, were white, and were married or co-habiting. The mean age was 31.4 years. Approximately half of the participants were multiparous; a greater proportion of women with an unplanned Caesarean delivery were primiparous (63.6%), whereas a greater proportion of women with a planned Caesarean delivery were multiparous (81.2%), compared to the vaginal delivery group. Approximately 42.9% of participants received lactation consultant support, which was more common among the Caesarean delivery groups. The prevalence of breastfeeding difficulties is displayed in Table 1. Overall, 80.9% of women reported experiencing at least one moderate to unbearable breastfeeding difficulty shortly after birth. The most common difficulty type was maternal physical difficulties (58.9%), followed by infant behaviour/health difficulties (54.9%) and infant latch difficulties (39.8%).

Table 2 Characteristics of participants (N = 418)
Full size table

Caesarean delivery and breastfeeding difficulties

Results for logistic regression modelling of Caesarean delivery (either planned or unplanned) and types of breastfeeding difficulties are shown in Table 3. There was a significant association between Caesarean delivery and low milk supply difficulties (AOR = 1.62, 95% CI = 1.16–2.28) and infant behaviour/health difficulties (AOR = 1.33, 95% CI = 1.01–1.75) in the adjusted models. No association was observed for maternal physical difficulties, high milk supply, social difficulties, or infant latch difficulties.

Table 3 Association between Caesarean delivery versus vaginal delivery and breastfeeding difficulties shortly after birth
Full size table

Results for the logistic regression modelling of planned and unplanned Caesarean delivery and types of breastfeeding difficulties are shown in Tables 4 and 5, respectively. Both planned Caesarean (AOR = 2.42, 95% CI = 1.19–4.92) and unplanned Caesarean (AOR = 2.21, 95% CI = 1.16–4.22) were significantly associated with low milk supply difficulties in adjusted models. No associations were observed for the remaining maternal or infant difficulties. Of note, the association between planned Caesarean and infant behaviour/health difficulties (AOR = 1.66, 95% CI = 0.93–2.98) was near the threshold of significance.

Table 4 Association between planned Caesarean delivery versus vaginal delivery on types of breastfeeding difficulties at delivery
Full size table
Table 5 Association between unplanned Caesarean delivery versus vaginal delivery and types of breastfeeding difficulties at delivery
Full size table

Discussion

Women of full-term infants who delivered by Caesarean were roughly twice as likely to report low milk supply difficulties shortly after birth than women who delivered vaginally. This is likely capturing a later onset of lactogenesis II (i.e. copious milk production) given the timing of data collection within 72 h after birth. A delay in lactogenesis II that persists beyond 72 h has shown to be associated with premature cessation of breastfeeding (Brownell et al., 2012). Many circumstances surrounding a Caesarean delivery may contribute to later onset of lactogenesis II, including maternal physiologic stress from undergoing surgery (Tully & Ball, 2014; Uvnäs Moberg et al., 2020), delayed initiation of mother-infant skin-to-skin contact and delayed breastfeeding initiation (Dewey et al., 2003; Prado et al., 2018), and possible lower sucking capacity in infants born by Caesarean (Zhang et al., 2016). Low breastfeeding self-efficacy may also be a factor, given that low self-efficacy has been linked to both Caesarean delivery and perceived low milk supply, separately (Huang et al., 2021). The association with low milk supply was slightly larger in magnitude for planned Caesarean (AOR = 2.42) compared to unplanned Caesarean (AOR = 2.21), though confidence intervals were wide and overlapping. Planned Caesarean can occur due to medical concerns necessitating a scheduled delivery to reduce the risks of continued pregnancy on maternal or fetal health (Hanley et al., 2010). Pregnancy complications have been shown to adversely impact breastfeeding outcomes (Scime et al., 2021). It is possible that these complications might act in combination with circumstances surrounding Caesarean surgery to delay lactogenesis, explaining the slightly larger association we observed with planned Caesarean and low milk supply. Alternatively, this finding may be explained by the common obstetrical practice of performing repeat Caesarean (Gu et al., 2020), given that a larger proportion of women with planned Caesarean delivery were multiparous compared to women with unplanned Caesarean delivery. Additional research that further categorizes planned Caesarean based on maternal request, medical indication, or repeat Caesarean would be helpful to explore these interpretations further.

We also found that infant behaviour/health difficulties were 30% more common in women who delivered by Caesarean. Our definition of infant behaviour/health difficulties was heterogeneous, and future research with sufficient sample sizes to examine each individual difficulty would help clarify the specific infant behaviours driving this association. Low milk supply difficulties following Caesarean likely contribute to infants showing heightened cues of hunger and distress. Perceived low milk supply may also influence maternal perceptions of infant behaviour, including a greater likelihood of reporting a lack of infant satisfaction, a higher need for formula, slow infant weight gain, and lack of breast fullness (Kent et al., 2021). Mothers who deliver by Caesarean report significantly higher pain compared to vaginal delivery, require more analgesic or anti-inflammatory drugs during their hospital stay (Zanardo et al., 2018), and have reported experiencing distress and shock following their operation, particularly when it was unplanned (Tully & Ball, 2013). It is therefore possible that maternal pain or stress following Caesarean surgery also contributes to women’s heightened sensitivity to infant fussiness or irritability in response to breastfeeding.

We found no significant associations between Caesarean delivery and maternal physical difficulties (e.g. sore or cracked nipples, maternal fatigue), high milk supply difficulties (e.g. swollen breasts, blocked milk ducts), social difficulties (e.g. inconvenience or embarrassment when breastfeeding), or infant latch difficulties. Given our study’s focus on experiences shortly after birth and prior to hospital discharge, null findings are expected as many of these difficulties (e.g. high milk production, breast pain, and blocked milk ducts) often present later in the postpartum period (Berens, 2015; Odom et al., 2013). It is reassuring that latch difficulties did not differ by mode of delivery, despite a high prevalence (approximately 40%) in this sample.

Our findings indicate that prompt hospital or public health support targeted towards promoting milk supply shortly after birth might help to improve women’s breastfeeding experiences following Caesarean delivery. Hospital policies that promote rooming in, skin-to-skin contact, and pre-discharge lactation assessment, as well as community-based educational and counselling initiatives, have been shown to be effective at improving self-efficacy, actual and perceived low milk supply, and breastfeeding outcomes (Galipeau et al., 2018; Sinha et al., 2015). Pharmacological intervention through use of galactogogues, such as domperidone, may also be considered in certain circumstances where low milk supply has been objectively confirmed (Brodribb, 2018). Support with interpreting and responding to infant behaviours and cues may also ameliorate breastfeeding challenges shortly after birth (Committee on Obstetric Practice Breastfeeding Expert Work Group, 2021).

Our study has several limitations to consider. The Happy Baby Study was comprised of adult mothers of full-term infants, and thus, our results are not generalizable to adolescent mothers or the preterm population. Recruitment dates back to 2010, and while the Happy Baby Study’s sample was representative of the general maternal population in Alberta at that time (Chaput et al., 2016), findings should be considered in light of changing Caesarean and lactation practices that may have occurred over the last decade, including rising rates of Caesarean deliveries and increased implementation of breastfeeding-friendly practices (Alberta Health Services, 2014; Canadian Institute for Health Information, 2020). Data on pregnancy complications were not collected and may be a source of unmeasured confounding. Despite the modest sample size, cell sizes were insufficient to model each of the 18 items in the BES separately. Grouping difficulties into types introduced heterogeneity into outcome measurement, which may have affected or obscured the strength and significance of associations across difficulties. We also acknowledge limited statistical power for some of the models stratified by Caesarean subtype, as evidenced by low precision (wide CIs) and associations at the threshold of significance (for example, infant behaviour/health).

Conclusion

Women who deliver by planned or unplanned Caesarean section are approximately twice as likely to report low milk supply shortly after birth than women who deliver vaginally. Infant behaviour/health difficulties were also more commonly reported following Caesarean, though to a lesser degree. Supportive hospital and public health interventions that can address low milk supply, whether perceived or actual, are recommended to ameliorate the breastfeeding experiences of mothers following Caesarean.

Contributions to knowledge

What does this study add to existing knowledge?

• Caesarean delivery is associated with reduced breastfeeding initiation and earlier discontinuation compared to vaginal delivery, but differences in breastfeeding difficulties by mode of delivery have seldom been explored.

• We found that mothers who experience Caesarean delivery have significantly higher odds of reporting low milk supply and perceived infant behavioural/health difficulties shortly after birth than women who deliver vaginally.

What are the key implications for public health interventions, practice, or policy?

• Interventions that are effective at improving actual or perceived low milk supply in the hospital and community may help to alleviate breastfeeding difficulties for mothers following a Caesarean delivery.