Our main findings are that our infants were equally likely to be exclusively breastfed by day 100 whether or not they reached extreme weight loss, but we show that newborns whose birthweight loss at discharge from the maternity ward was below the median were more likely to discontinue exclusive breastfeeding by days 15, 30 or 100 of life, but not by day 7 of life. To the best of our knowledge, this study is the first to examine the link between in-hospital birthweight loss and breastfeeding duration beyond month three of life.
From the physiological standpoint it is unknown whether extreme weight loss results solely from voiding and insensible water loss during those days when milk oral intake is still low or it is due mainly to inadequate intake [
21]. However, it is worth remembering that the average amount of breastmilk ingested during the first day of life by full term neonates is 15 mL [
22]. In addition, very recent research [
4] shows that once newborns started gaining weight, similar patterns of weight gain emerged between the group above and the group below the threshold for extreme weight loss. Moreover, not only delayed onset of lactogenesis, but also intrapartum fluid net balance and infants’ rooting movements on day two of life may be involved in food intake regulation and predict birthweight loss [
15,
16,
23]. Taken together, current data do not support hospital routines that are associated with both spurious increase in birthweight loss and poor breastfeeding outcomes of extreme weight loss infants. Depending on what kind of feeding directions are provided to mothers when newborns lose > 7% of their birthweight, some babies will continue breastfeeding and others will be finally switched to formula.
On the other hand, our findings expand on the association between milder degrees of birthweight loss and breastfeeding continuation. We did not find a difference of exclusive breastfeeding discontinuation at the more extreme end of birthweight loss, but we show that newborns whose birthweight loss at discharge was below the median were more likely to discontinue exclusive breastfeeding in the medium term. Our results are in line with Flaherman et al. findings [
24], they have reported that duration of exclusive breastfeeding through 1 month was less likely among newborns whose discharge birthweight loss was >50th percentile, and we find the same by days 15, 30 and 100. It was expected that birthweight loss would be associated with decreased rate of exclusive breastfeeding in the very short term. In contrast, this was not true by day 7 but it was true thereafter. The underlying assumption is that birthweight loss trajectory is a marker of suboptimal breastfeeding efficiency in the medium term but not a trigger for imminent formula supplementation.
Finally, we found in our study that known predictors of breastfeeding cessation such as sex of child, parity, impaired glucose tolerance, and mode of delivery, were associated with breastfeeding cessation. Accordingly, the literature reports that sex of child impacts breastfeeding duration in India, Australia, Scandinavia, Latin America, and the US. It varies from country to country. Within the US, Hispanic women have lower odds of breastfeeding duration if they have sons compared to Hispanic women who have daughters [
25]. The same applies to Brazil [
26]. Also, Kenyan boys are more likely to be introduced to complementary feeding early compared with girls [
27]. Conversely, Indian boys are breastfed more than girls, yet having few or no older brothers results in earlier weaning of daughters [
28]. Our study shows that parity < 1 is an independent predictor of breastfeeding cessation from day 7 to day 100 of life. Previous studies show that breastfeeding duration increases with increasing parity [
29,
30], which might be related to previous breastfeeding experiences. Cesarean section is a recognized risk factor for breastfeeding initiation failure. However, there is not absolute agreement on this point, the UK 2005 Infant Feeding Survey found that mothers were equally likely to be breastfeeding at 1 week regardless of delivery methods [
31]. We report that cesarean section is significantly associated with breastfeeding cessation by weeks 1 and 2 of life, but our previous research found that mode of delivery was not associated with breastfeeding cessation [
20], we do not have an explanation for this. Our findings add to the evidence that a woman’s impaired glucose intolerance adversely affect her lactation outcomes, this important association may account for some of the association with weight loss and exclusive breastfeeding outcomes. It is well established that overt gestational diabetes is associated with significantly increased risks of adverse breastfeeding outcomes [
32]. In addition, emerging data suggest that lesser degrees of hyperglycemia also increase offspring risks, including early breastfeeding cessation [
18,
33]. It has been shown that insulin-sensitive gene expression tends to be upregulated during the lactation cycle, that is why insulin plays a direct role in lactation [
34].
Limitations
First, data for patients’ characteristics were collected retrospectively in addition to routinely; therefore, lack of accuracy in this area is an expected shortcoming of the survey. Conversely, lack of accuracy of breastfeeding duration is not expected because these data were collected prospectively. Second, we have no detailed data on the many possible reasons for cessation of exclusive breastfeeding. We were not able to adjust for exclusive breastfeeding intention. Third, another limitation is that data are not available at the time of maximum weight loss, but reliability of our results is corroborated by the fact that the reported median weight loss of 6% corresponded to the 5.95% mean weight loss reported in a large birth cohort of term, healthy, exclusively breastfed neonates from a breastfeeding friendly birth center [
9]. Finally, the studied population came from general care pediatric clinics in a middle-class neighborhood in Spain, which may not be representative of the general population. The data should be verified in a larger study encompassing more centers. We hope this study will help debunk some of the myths surrounding the rush for supplementation in cases of extreme weight loss.