The rate of caesarean section (c-section) in Canada has increased from 17.6 % in 1995 to 27.1 % in 2012 [1]. Although worldwide consensus on the optimal rate of c-section is variable [2], the World Health Organization (WHO) [3] suggests 5–15 % of all live births, far below current c-section rates in Canada. C-sections are associated with poor maternal and infant health outcomes [4–6], such as higher incidences of maternal infection and uterine haemorrhage, and infant respiratory distress and hypoglycaemia [6, 7]; all of which may adversely impact breastfeeding success.

Researchers have shown that women who deliver by c-section delivery are less likely to breastfeed, or delay breastfeeding initiation [8–14]. Breastfeeding within the first hour post-delivery has been cited as an important predictor of continued breastfeeding [15–17]. An early study by Rowe-Murray and Fisher (2002), found that babies born via c-section were less likely to be have skin-to-skin contact immediately after birth and were more likely not to have attempted breastfeeding within the first 24 h post delivery [8]. Skin-to-skin contact has been suggested to improve breastfeeding initiation, maintenance and duration [18]. Delays in breastfeeding initiation accompanying c-section delivery are associated with maternal/infant separation, reduced suckling ability, decreased infant receptivity, and insufficient milk supply, which are predictive of shortened breastfeeding duration [8, 12–14, 19–23].

Breastfeeding is associated with infant health benefits, such as fewer childhood illnesses, lower blood pressure and cholesterol levels, lower prevalence of obesity, and improved intelligence as adults [24, 25]. Maternal benefits of breastfeeding include faster involution of the uterus and lower risk of haemorrhage after birth, in addition to lower a lower lifetime incidence of type II diabetes, and breast and ovarian cancer [26]. These studies have suggested that the benefits of breastfeeding act in a dose response relationship, whereby increasing breastfeeding duration results in more infant health benefits [24–26]. Since it has been established that breastfeeding confers several long-term maternal and infant benefits, it is important to understand the impact of c-section delivery on breastfeeding initiation, and duration.

Health Canada [27] and the WHO [16] recommend that all infants should be breastfed exclusively for the first six months of life, and continue breastfeeding up to and beyond two years of age. Perinatal Health Indicators for Canada (2013) state that breastfeeding initiation rates in Canada remain at approximately 87 % (2005–2010) [28]. However, the 2008 Perinatal Health Report found that only 53.9 % of babies were still breastfeeding at 6 months of age [29]. The rate of exclusive breastfeeding at 6 months has slowly increased from 20.3 % in 2005 to 25.9 % in 2010 [28],but the rate is still low. Only three of the 13 provinces and territories in Canada have designated facilities that meet the WHO Baby Friendly Hospital Initiative (BFI) guidelines [30]; thus, improvements in this area could help to meet the national and international breastfeeding recommendations [16, 27].

Research comparing the effect of c-section delivery to vaginal delivery on breastfeeding duration document conflicting results and often lack adequate sample sizes to support their findings [20, 31–33]. Pérez-Escamilla et al. (1996) indicated that c-section delivery did not affect breastfeeding duration if women initiated breastfeeding from birth and maintained breastfeeding for at least four weeks postpartum, [32]. However, Brown & Jordan (2013) contradict these results reporting that mothers who gave birth via c-section were more likely to discontinue breastfeeding compared to those who gave birth vaginally [33]. The Ontario Mother and Infant Study, by Watt et al (2012), found no differences between c-section and vaginal deliveries on breastfeeding initiation or duration to 6 weeks postpartum [14]. However, this study found that emergency c-sections and assisted vaginal deliveries were associated with an increased likelihood of initiation and duration of breastfeeding, when compared to planned c-section and unassisted vaginal birth [14]. Watt et al. (2012) acknowledged that these results were unexpected and suggested further research to explore all modes of delivery and impact on breastfeeding [14].

In addition, studies have found that the type of c-section, either emergency or planned, influences breastfeeding outcomes [3, 14, 22, 34]. A meta-analysis performed by Prior et al. (2012) pooling 48 studies (n = 553,306) found that women who delivered by planned c-section were less likely to initiate any breastfeeding; whereas emergency c-section delivery had no significant impact on initiation. The authors found no association between any type of c-section delivery and exclusive breastfeeding up to six months [31]. However, the authors of this meta-analysis acknowledge that the clarity of the association between mode of birth and several breastfeeding outcomes is limited given that it was not the primary outcome of most of the studies included and that there was evidence of heterogeneity between studies, often resulting from small sample sizes and unadjusted point estimates from individual studies.

Guidelines by the WHO BFI [16] state that keeping mother and baby together for at least the first hour post birth leads to improved initiation and duration of breastfeeding [34, 35]. Although findings support that both emergency and planned c-section affect breastfeeding initiation and duration, the direction of the association among studies is inconclusive and warrants further exploration of breastfeeding, stratified by type of c-section. Our large population-based study examined the effect of planned or emergency c-section and vaginal delivery of singletons, comparing breastfeeding initiation, duration, and difficulties in the early postpartum period, while taking into account for key indicators of breastfeeding success.

Significant associations were found across mode of birth compared to parity (p<0.001) and gestational age (p<0.001). Of the emergency c-section births, approximately 70 % (n = 304) were nulliparous compared to 49 % n = 1094) of vaginal births and 20 % (n = 60) of planned c-sections. Conversely, the majority of women who delivered by planned c-section (80 %, n = 235) were multiparous. Although 93 % (n = 2784) of the sample delivered at full term (≥ 37 weeks gestation), preterm deliveries were found in 13 % (n = 55) of women who delivered by emergency c-section, versus a smaller proportion of vaginal births (6.5 %, n = 146) and planned c-section (3.4 %, n = 10).

Our study found associations between mode of birth and a successful first attempt at breastfeeding (p<0.001), breastfeeding within the first 24 h (p<0.001), and success with breastfeeding when leaving the hospital (p<0.001). Almost 40 % (n = 167) of women who had an emergency c-section were unable to successfully breastfeed their baby on the first attempt compared to approximately 25 % of women who either delivered vaginally (n = 542) or by planned c-section (n = n = 74). A higher proportion of women who delivered by emergency c-section did not successfully breastfeed within the first 24 h of birth (7.3 %, n = 31), when compared to vaginal births (3.1 %, n = 69) and planned c-sections (2.1 %, n = 6). Of the women who had an emergency c-section, 7.3 % (n = 31) reported not being able to successfully breastfeed within the first 24 h post delivery, and 17 % (n = 72) of the women were less likely to breastfeed prior to being discharged from the hospital.

Of the women who initiated breastfeeding (n = 2954), 62 % (n = 1832) reported having more than one difficulty with breastfeeding. For the women in our study, there were significant differences across mode of birth and breastfeeding difficulties with baby (e.g., latching or a sleepy baby) (p = 0.002), personal discomfort with breastfeeding (e.g., sore nipples, swollen breasts) (p<0.001), and other difficulties (e.g., perceived low milk supply or having flat or inverted nipples) (p<0.001). More than half of the women (54 %, n = 236) who delivered by emergency c-section, reported having had breastfeeding difficulties with baby and 41 % (n = 180) had described other difficulties. Conversely, approximately 45 % of women who delivered either vaginally (n=1016) or by planned c-section (n = 135) reported difficulties with baby and only around 30 % (vaginally n = 658 or planned c-section n = 98) reported other difficulties with breastfeeding.

The women in our sample were asked about breastfeeding supports accessed in and out of the hospital. Mode of birth was associated with seeing a lactation consultant in hospital (p<0.001) and seeking additional support from a breastfeeding clinic or visiting a health care provider after discharge (p<0.001). Approximately 67 % (n = 284) of women who delivered by emergency c-section saw a lactation consultant and 58 % (n = 245) found additional supportive care after discharge. Conversely, about 41 % (n = 117) of women who had a planned c-section sought additional breastfeeding support upon discharge from the hospital.

In an unadjusted analysis, women who delivered by emergency c-section were more likely to discontinue breastfeeding ≤12 weeks postpartum (OR=1.35; 95 % CI: 1.03, 1.76; p = 0.029), when compared to those who delivered vaginally; there was no significant difference between planned c-sections and vaginal birth. However, in the adjusted multivariable logistic regression model (Table 3), women who had a planned c-section were more likely to have early cessation of breastfeeding (≤12 weeks) (OR=1.61; 95 % CI: 1.14, 2.26; p = 0.006) when compared to those who delivered vaginally. There was no significant difference in breastfeeding cessation between women who had an emergency c-section and women who delivered vaginally in the adjusted analysis. Covariate patterns and resulting confounding structure in the multivariable model explains the change in magnitude of the effect estimates and the reversal in the association between the levels of exposure in mode of birth in the unadjusted versus adjusted analysis. An important confounder in the adjusted model was parity as it was strongly associated with both early breastfeeding cessation and mode of birth (Table 1 and Table 3); a significantly higher proportion of multiparous women were found among those who delivered by planned c-section (80 %, n = 235), compared to vaginal (51 %, n = 1155)) or emergency deliveries (30 %, n=131) (p<0.0001).

Our findings demonstrate that planned c-sections are associated with reduced breastfeeding success in the first 4 postpartum months, when compared to vaginal births. In particular, we have successfully added to the growing evidence that planned c-sections negatively affect breastfeeding initiation and duration.

All of the Calgary area hospitals support breastfeeding but may not have a specific breastfeeding protocol instituted, nor are they BFI accredited. Each hospital in Calgary has a lactation consultant and post discharge breastfeeding resources with physician and nursing support available. In our study, women who delivered via c-section were more likely to report breastfeeding difficulties when compared to women having had vaginal deliveries. Brown and Jordan (2013) found similar findings; women in their study who delivered via c-section also had more problems with latching, positioning and more pain when compared to those women with vaginal birth [33].

In our study, women who had an emergency c-section were more likely to have had an unsuccessful first breastfeeding attempt, were unable to breastfed their baby within the first 24 h, and were unable to breastfeed upon leaving the hospital. Zanardo et al. (2010) found similar results, reporting that women who delivered via emergency c-section were more likely not to have been able to breastfeed their baby at delivery or at discharge [39]. It has been hypothesized that the maternal and fetal stress response associated with complications during delivery, particularly related to c-sections, may be related to increased difficulty and early cessation in breastfeeding during the early postpartum period [14, 33]. Lactogenesis may be equally affected by the insult of abdominal surgery in both planned and emergency c-sections, although the notion of an “emergency” may invoke a greater or prolonged maternal stress response.

Our data showed that women with planned c-sections were more likely to indicate they did not intend to breastfeed and therefore did not initiate it compared to those with other delivery modes. This concurs with a large meta-analysis that found an association between a lower likelihood of initiation of breastfeeding and planned c-sections [31]. Although our study did not analyze maternal or medical reasons for planned c-section, reasons for planned c-section coupled with no intention to breastfeed should be explored given that the rate of planned c-sections has been rising since the year 2000 [1, 28]. The Canadian Maternity Experiences Survey in 2008 reported that of all 6,421 women who completed the survey, 26 % delivered by c-section and 13.5 % of all c-sections were planned [29]. Of those planned c-sections, the majority (86.9 %) were planned for medically indicated reasons [29]. The Canadian Perinatal Surveillance System acknowledged that c-section rates are increasing and recognize that c-sections on maternal request could be part of this growing burden [21]. However, a study by Hutton and Kornelsen (2012) found that in British Columbia, one of the provinces with the highest rates of c-section delivery, very few women requested c-sections with no medical indication and concluded that the increased rate of c-section delivery in Canada may not be related to maternal choice [40]. Similarly, a review of 17 articles on the topic of maternal decision making found that only a small portion of women requested c-sections for non-medically indicated reasons in each study [41]. However, the studies included in this review were susceptible to a high risk of bias due to small sample sizes, warranting further studies specifically investigating maternal decision making surrounding requested c-section deliveries for non-medical indication.

Lower breastfeeding initiation and increased difficulties with breastfeeding in women with c-section deliveries may be related to a physiologic influence on lactogenesis [42, 43]. A study by Evans et al. (2003) found that breast milk transfer among women with c-sections was significantly lower in the first 5 days postpartum, compared to women with vaginal births [42]. Similarly, Scott and Binns (2007) found that delayed onset of lactation was significantly higher in mothers that delivered via c-section compared to those that delivered vaginally [43]. It is postulated that the hormonal pathway that stimulates lactogenesis is disrupted by c-section delivery, either from maternal stress or decreased oxytocin secretion, and can hinder milk production [34, 39, 42, 44]. Late preterm infants, those born between 34 and 36 6/7 weeks gestation, often have a multitude of issues including disorganized sucking skills, lower birth weight and low level of alertness, all of which can adversely affect breastfeeding initiation [34, 45]. Since planned c-sections are performed prior to the onset of labour and before 40 weeks gestation, it is possible that physiological reasons are responsible for the decreased likelihood or troubled breastfeeding initiation. In response, the American College of Obstetricians and Gynecologists (2013) recommended that c-sections performed prior to onset of labour should not be conducted prior to 39 weeks unless medically indicated [46].

When controlling for other variables potentially associated with breastfeeding, we found that planned c- section predicts early breastfeeding cessation, defined as weaning at or prior to 12 weeks. Although Zanardo et al. (2013) found that c-section delivery was associated with lower rates of exclusive breastfeeding at 6 months, no difference between planned versus emergency c-sections was found [19]. Women who had planned c-sections in our study were less likely to intend to breastfeed or initiate breastfeeding, and less likely to seek lactation support. Although we controlled for parity in our multivariable model, the bivariate associations with mode of birth and breastfeeding intention and initiation may be related to the other factors such as variable infant feeding choices or return to work. Multiparous women may have had a previous challenging breastfeeding experience and chose not to breastfeed. Previous breastfeeding experiences in women are known to impact future decisions surrounding breastfeeding. Although this study did not look at prior experiences with breastfeeding, women who have had difficult breastfeeding experiences with prior births may require particular attention to promote future positive breastfeeding experiences. More targeted interventions towards women planning on c-section delivery from the perinatal period and into the postpartum time, may encourage breastfeeding initiation and continuation of breastfeeding.

As our study found that women who have emergency c-sections were more likely to have difficulties with breastfeeding and to seek additional lactation support, focused education surrounding breastfeeding techniques and encouragement to these mothers could be given immediately post birth in order to provide the appropriate anticipatory guidance to reduce difficulties. In Canada, there are few hospitals that have breastfeeding practices that adhere to these standards set by the BFI. We recommend that health care professionals should be aware that those women who have delivered via c-section may require more supportive care immediately post-delivery and in the early postpartum period. In particular, our study identified that those who delivered via emergency c-section may require access to lactation consultation and other supports for breastfeeding. These supports could be made available in the immediate post-partum period in order to encourage breastfeeding within 1 h post birth and ensure successful breastfeeding within the first 24 h, as recommended by the BFI and many research studies as being indicative of long term breastfeeding success [15–17]. As women who delivered via emergency c-section were more likely not to have successfully breastfed before leaving the hospital, the postpartum home visit presents an ideal opportunity to provide additional breastfeeding assessments and support to overcome any difficulties for these women.

Our study suggests that c-sections are associated with more breastfeeding difficulties, greater use of resources, and shorter breastfeeding duration compared to vaginal deliveries. Planned c-sections coupled with no intention of breastfeeding may be an emerging trend among women that requires further exploration. As planned c-section delivery may be associated with lower intention to and initiate breastfeeding as well as early cessation of breastfeeding, health care provider support and counselling in order to provide anticipatory guidance surrounding breastfeeding could be given prenatally to pregnant mothers and families considering a planned c-section. As well, additional supportive care made available to lactating women with emergency c-sections during the immediate to early postpartum period would be recommended in order to ensure early success with breastfeeding.