Feeding during the first 3 days after birth other than breast milk is associated with early cessation of exclusive breastfeeding

Abstract Exclusive breastfeeding (EBF) has significant effect on morbidity and mortality. EBF is established when breastmilk alone is exclusively fed from birth until 6 months of age. However, feeding during the first 3 days after birth is often ignored for various reasons. We aimed to assess the role of feeding during the first 3 days in respect of early cessation of EBF. Data of 1,040 children aged under 6 months was derived from the baseline survey of Suchana, a large‐scale nutrition program, conducted in Sylhet, Bangladesh, and subsequently analysed. Guidelines established by World Health Organization were used to define EBF and feeding during the first 3 days. The strength of the association between feeding during the first 3 days and early cessation of EBF was established using multiple logistic regression after adjusting for other covariates. Among all children, around 62% and 13% were exclusively breastfed and were fed something other than breastmilk within the first 3 days of birth, respectively. Feeding during the first 3 days was independently and significantly associated with early cessation of breastfeeding (adjusted odds ratio: 1.94, 95% confidence interval [1.31, 2.88], p = .001). Less than four antenatal care (ANC) visits, increased child's age and increased household size were also independently associated with early cessation of EBF. Feeding during the first 3 days of birth is a significant predictor of early cessation of EBF. Simple counselling activities to discourage feeding anything within the first few days of birth may increase the prevalence of EBF in rural Bangladesh without investing additional resources.

introduction of complementary feeding, the two critical components of optimal infant feeding practices are promoted globally to avert child mortality and morbidity (Nguyen, Withers, Hajeebhoy, & Frongillo, 2016;Pak-Gorstein, Haq, & Graham, 2009). The Global Nutrition Targets to be achieved by 2025, is increasing the rate of EBF to, at least, 50% (World Health Organization & UNICEF, 2014) from the current global statistic of around 40% (UNICEF, 2015), whereas according to Demographic and Health Surveillance data, the current prevalence of EBF in Bangladesh is 55.3%, a decrease of 8.2 percentage points from 2011 (NIPORT, Mitra and Associates, & ICF International, 2016) despite significant stakeholders' effort to promote EBF (Ministry of Health and Family Welfare, 2016), raises concern. Reasons behind termination of breastfeeding early are multifaceted and include sociodemographic, biophysical, and psychosocial reasons. Existing evidence suggests that maternal age, father's occupation, and mode of child delivery are significantly associated with discontinuation of breastfeeding (de Oliveira & Camelo, 2017;Sun, Chen, Yin, Wu, & Gao, 2017;Venancio & Monteiro, 2006). Other aspects such as mothers' selfefficiency, post-natal depression, anxiety, intention to breastfeed, comfort in breastfeeding, desires or attitude of mothers regarding breastfeeding, mother-infant bonding, and family support is important in sustaining breastfeeding (Sun et al., 2017).
However, it is to be noted that EBF for children under 6 months of age as per the Demographic and Health surveys or as per World Health Organization's recommendations is established if the child did not receive anything other than breastmilk or necessary medicines in the 24 hrs preceding the survey (Greiner, 2014;NIPORT et al., 2016;World Health Organization, 2008). However, using the 24-hr recall criteria often misclassifies children who have deviated from the criteria of EBF at some point in their life especially during the first 3 days after birth. Children delivered through caesarean section or if the child is born with low birth weight and needs special medical attention may often cause delayed initiation of breastfeeding due to delayed first mother-child contact (Dewey, Nommsen-Rivers, Heinig, & Cohen, 2003;Rowe-Murray & Fisher, 2002). Delayed onset of lactation may also cause children to be fed other foods until lactation normalizes, and it may take over 72 hrs or 3 days (Dewey et al., 2003). Thus, during first few days after birth, it has often been observed that many children are being shifted from EBF to predominant breastfeeding and then back to EBF (Greiner, 2014;Pak-Gorstein et al., 2009). Also, a common phenomenon in many culture including that of Bangladesh is to provide prelacteal feed to children during the first 2 to 3 days after birth even though they are continued to be exclusively breastfed from thereafter (Ahmed, Rahman, & Alam, 1996;Greiner, 2014;Pak-Gorstein et al., 2009;Taranum & Hyder, 1995). Nonetheless, it is also possible that easiness in feeding children other foods during early days and if it does not have any adverse effect on the child's health may discourage mothers to continue EBF. Although there is literature on a diverse range of variables explaining early cessation of EBF, there is lack of evidence clarifying the effect of feeding during the first 3 days of childbirth on early EBF cessation.
This phenomenon provides the scope for this paper, which is to assess the relationship between early cessation of EBF and introduction of any feeding within the first 3 days after birth.

| METHODOLOGY
The analyses are based on the baseline survey data of Suchana, a new program designed to improve the nutritional status of children and targeted the poorest segment of the population in Sylhet, Bangladesh.
Cross-sectional data of 1,040 beneficiary households including data of the youngest child aged below 6 months using a structured ques- villages from the unions using participatory rural appraisal strategy. At the final stage of the selection process, systematic sampling using a calculated "interval" was used to select the households for the survey.
Data were recorded digitally using personal digital assistants and maternal height, and weight was recorded using locally made wooden scales used previously in national surveys and SECA 874 model weight scale with 1-gm resolution. Child's length was measured using SECA 416 infantometer with a precision of 1 mm.

| Variables of interest
EBF: Infants 0-5 months of age, who received only breastmilk, and due to medical condition(s) required medicine(s) including Oral Rehydration Solution (ORS) during the previous day (World Health Organization, 2010). The variable was dichotomized into responses "yes" and "no." Feeding during first 3 days after birth: Whether the infant was fed anything other than the breastmilk and required medicine(s) including ORS within 3 days of birth. The variable was dichotomized into responses "yes" and "no."

Key messages
• Around 60% children are exclusively breastfed in rural Bangladesh.
• Feeding children anything except breastmilk during the first 3 days of birth is significantly and independently associated with the child being not exclusively breastfed later.
• Increased child's age less than the optimal number of antenatal care visits and household size are significantly associated with early cessation of breastfeeding.
Other variables used for analyses are all pertinent to the cessation of EBF and were identified after consulting similar literature. The previous nationally representative survey questionnaires and also demographic survey questionnaire were consulted for selection of the variables relevant for the analyses (James P Grant School of Public  et al., 2016). The selected variables are initiation of breastfeeding within first hour of birth, whether the child received colostrum, child's sex, child's acute morbidity status (Whether the child was sick in past 15 days before the survey), place of child delivery, maternal antenatal care (ANC) visit status, mother receiving any support on at least three household chores (cooking, gathering water/firewood for the house, cleaning, child care, agricultural activities, selling produce or going to market, homestead gardening, and homestead poultry rearing), whether mother has ever been threatened by the husband or other family members (threatened for divorce, taking another wife, and verbal or physical abuse by husband or any other family member), maternal education status, paternal education status, whether the mother received any message on breastfeeding or complementary feeding practices, maternal knowledge on EBF, maternal body mass index (BMI), household food insecurity status measured by household food insecurity access scale (Coates, Swindale, & Bilinsky, 2007), father's education status, household size, number of sleeping rooms in the household, total yearly income from all sources, maternal age, maternal age during the marriage, maternal age during the first pregnancy, maternal dietary diversity 1 (Kennedy, Ballard, & Dop, 2011), the child's age, the child's birth order, and asset score 2 (NIPORT et al., 2016).
Additionally, we have examined bivariate relationship between caesarean section delivery and delayed initiation of breastfeeding and also the proportion of children who were born in institutions and were fed other food or liquid during the first 3 days after birth and the proportion of children with delayed initiation of breastfeeding (>1 hr after birth) and were born in institutions.

| Analysis procedure
All analyses were done using STATA, version 15. As the data were collected from multiple clusters, STATA command svyset was used for adjusting the clusters. The primary analysis involved producing descriptive statistics mentioning mean and proportion with 95% confidence interval (CI) using appropriate cut-off values. Simple logistic regression was used to establish crude bivariate relationship between EBF and other variables. Multivariate regression was used to determine the strength of association between feeding during the first 3 days and early cessation of EBF after adjusting for necessary covariates. Variables were considered significant only if the p value was less than .05. Variables with a p value less than .25 (Bursac, Gauss, Williams, & Hosmer, 2008) were used in the final multiple logistic regression model.

| RESULTS
Of the 1,040 children assessed, 62.4% of children were exclusively breastfed and 13.0% of children had consumed anything other than breastmilk within the first 3 days of birth. All children had breastfeeding initiated within 24 hrs after birth. On other criticalindependent variables assessed, 12.8% of children did not have breastfeeding initiated within the first hour of birth, and 47.9% had acute morbidity during the last 15 days preceding the survey, whereas 83.7% of mothers had less than four ANC visits during last pregnancy and 74.0% of children were delivered at home. In total, 12% of children were delivered through caesarean section, and among all children born in institutions, 20.59% (n = 34) children had fed during the first 3 days after birth. However, for children whose breastfeeding was initiated at least 1 hr after birth, 49.62% (n = 66) were born in institutions. Additionally, 16.3% of mothers received no support from other household members for at least three household chores, and 33.9% mothers received threat from their husband. In total, 32.0% of mothers had a BMI of less than 18.5, and 20.8% mothers and 40.6% fathers had no formal education. The average age of the children was 3.77 months, whereas the average household size was 6.48 and average household yearly income from all sources was Tk. 108,809. 3 Details of characteristics are presented in Table 1

| DISCUSSION
This cross-sectional study examined the relationship between early cessation of EBF and introduction of any feeding within the first 3 days after birth in rural Bangladesh. Our findings dictate that around 60% of children were exclusively breastfed and feeding children anything except breastmilk or required medicine during the first 3 days of birth is significantly and independently associated with the child being not exclusively breastfed later. The odds of being not exclusively breastfeed is around twice for children if they had consumption of anything except breastmilk and medicines during the first 3 days after birth. Our findings also showed that the increased child's age less than the optimal number of ANC visits and household size were also significantly associated with early cessation of breastfeeding.
As to explain the occurrence of EBF cessation within the first 3 days, it could be mentioned that traditional and cultural factors play a pivotal role in introducing other feedings within first few days after birth and that may influence early discontinuation of EBF (Ahmed et al., 1996;Pak-Gorstein et al., 2009). Anecdotal evidences of few traditional beliefs that may led to cessation of EBF are giving honey to the new born after birth, giving food so that child becomes normal and quiet, milk secretion does not happen after delivery so food is given, and mustard oil is given to clear the child's oral cavity (Taranum & Hyder, 1995). Infant feeding practices in communities and maternal decision to breastfed are strongly influenced by people's perception, knowledge, and beliefs about everyday life (Ahmed et al., 1996;Pak-Gorstein et al., 2009;Roy, Dasgupta, & Pal, 2009). Cultural aspects regarding feeding foods and liquids other than breastmilk should also be taken into consideration when studying early introduction of other foods (Ahmed et al., 1996;Lakati, Makokha, Binns, & Kombe, 2011;Pak-Gorstein et al., 2009;Taranum & Hyder, 1995). breastfeeding issues to build confidence can be incorporated in programs for promoting EBF discussion on this issue (Ku & Chow, 2010).
Discussions and other EBF promotional activities can be inclusive of friends, other senior members of the family, and in-laws to improve the uptake of the concept of EBF (Chandrashekhar et al., 2007;Khanal, Adhikari, Sauer, & Zhao, 2013).
On other covariates examined, increased age of the child was also found to be significantly associated with cessation of EBF in our study. Other studies also reported similar findings (Perera et al., 2012;Victor, Baines, Agho, & Dibley, 2013). As per a report published on an initiative to improve infant feeding practices in Bangladesh also showed the rate of EBF to decline sharply with increased child's age (Saha et al., 2011). A cohort study from neighbouring Sri Lanka indicated that maternal anxiety over inadequate breastmilk production and not achieving optimal weight gain for age to be the most important reasons for early EBF cessation (Perera et al., 2012). In our study, less than four ANC visits was found to be significantly associated with early cessation of EBF. It is possible that the mothers who attended the standard number of four ANCs have retained the counselling they received on EBF and acted accordingly (Nekatebeb, Guyon, Beyero, & Stoecker, 2010). Stated by a previously conducted study, ANC visit is an important access point that can improve maternal breastfeeding behaviour (Tariku et al., 2017) and antenatal feeding intention was T A B L E 2 Relationship between early EBF cessation and feeding during first 3 days after birth (n = 995) significantly associated with early breastfeeding cessation (Oakley, Henderson, Redshaw, & Quigley, 2014).
On the significance of increased household size being associated with early cessation of EBF, it could be said that the reasons are multifaceted; increased household size reflects narrow interval between births, and mothers may have to cease EBF as a preparation for pregnancy (Akter & Rahman, 2010a;Akter & Rahman, 2010b;Reddy & Abuka, 2016). Additionally, other factors pertaining to higher household size and parity such as previous adverse breastfeeding experience (de Oliveira & Camelo, 2017) and poor health of mother due to high parity (Giashuddin, Kabir, Rahman, & Hannan, 2003) may cause early cessation of EBF. Elderlies in the family especially grandmothers play a critical role in support and care-seeking of children (Saha et al., 2011). A study in Bangladesh has shown that in around 42% of EBF cessation cases, food was introduced by the grandmother, followed by the traditional birth attendant (24.5%) (Taranum & Hyder, 1995).
It is evident that breastfeeding is the ideal nourishment for a child up to 6 months of age. It remained crucial to achieving fourth millennium development goals and a priority in the second sustainable development goal (Kaur, 2017 (Giashuddin et al., 2003;Saha et al., 2011) and also by ensuring more focused monitoring of the pertaining activities by the stakeholders.

| Limitations and strengths
For the purpose of collecting data on infant and young child feeding, we had to depend solely on the respondents' response. There is a probability that some of the responses are prone to recall bias; however, the data collectors were specifically trained to collect those responses as accurately as possible. The cross-sectional nature of the study prevents it from drawing any causal inference; however, a large sample size was added to the strength of the study.

ACKNOWLEDGMENTS
This manuscript is based on data collected through the baseline survey of Suchana; the intervention itself is to be implemented by Save the Children, with support from DFID and EU. icddr,b acknowledges with gratitude the commitment of Save the Children, DFID, and EU to its research efforts. We acknowledge with gratitude the commitment of the Government of the People's Republic of Bangladesh to icddr,b's research strength. We also acknowledge the following donors for providing unrestricted support to icddr,b's effort and advancement to its and icddr,b. This project itself is a joint partnership between DFID and European Union (EU), Save the Children, and other consortium.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.