An ever-growing body of evidence indicates that exclusive breastfeeding for the first 6 months must be considered a public health priority, since it promotes maternal-infant well-being and reduces healthcare costs. The World Health Organization (WHO) and the United Nations Children's Fund (UNICEF) strongly recommend exclusive breastfeeding for 6 months after delivery, followed by continued breastfeeding during weaning. However, breastfeeding rates remain low, settling in Europe between 13 and 39% (1).

Breastfeeding is a complex process, whose achievement implies the successful interaction between two individuals. It requires that both the mother and the newborn have an intact neurological status, a good state of health, and the ability to interact with each other. None of the methods currently used to assess the well-being of the newborn (e.g., Apgar score or neurobehavioral tests) is predictive of the success of this interaction, and many maternal variables, including age, parity, mode of delivery, infusion of intravenous fluids during delivery, attachment style (2–4), may interfere with it.

The currently available literature suggests that the second and third days following birth (48–72 h post-partum) appear to be the days of maximum weight loss (5). Postnatal weight loss up to 7% at nadir can be considered physiological (6, 7). It must be taken into consideration that the extent of postnatal weight loss is affected by many variables. On the basis of the available evidence, clinical guidelines caution that a postnatal weight loss greater than 7% indicates the need for careful assessment of the neonate and could indicate ineffective breastfeeding (6). However, to the best of our knowledge, to date no data are available on the predictive value of the magnitude of postnatal weight loss on exclusive breastfeeding continuation up to 6 months. The aim of this study was to investigate differences between full-term newborn infants showing a Weight Loss (WL) <7 or WL ≥7% at nadir both in terms of persistence of exclusive breastfeeding at 6 months and occurrence of events leading to its interruption (any integration, complementary feeding, shift to exclusive formula feeding, weaning).

This manuscript adheres to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines. This study project was reviewed and approved by the Institutional Ethics Committee (register 166798) of the Parini Regional Hospital, which is the only facility in Aosta Valley and, since 2010, has been repeatedly designated by UNICEF as a Baby-Friendly Hospital.

At our hospital, every mother receives standardized breastfeeding support, regardless of type of delivery (vaginal delivery or cesarean delivery) or labor analgesia. The decision to administer formula milk or glucose to newborns follows extremely strict criteria, in line with the recommendations of the WHO/UNICEF Ten Steps to Successful Breastfeeding. Infants are weighed between 2 and 6 h from birth and between 48 and 72 h from birth (at discharge) and WL is expressed as a percentage of birthweight, calculated on the lowest weight reached.

We conducted an observational cohort study, from January 1st, 2014 to September 22nd, 2015.

For the purposes of the present study, we considered eligible mothers of healthy newborns (to avoid bias related to underlying diseases), born at term (i.e., from 37 to 41 + 6 weeks gestational age) and with a birthweight between 2,500 and 4,000 g. We decided to exclude preterm and small for gestational age newborns, since they have been reported to be at increased risk for greater postnatal weight loss than full-term and appropriate for gestational age ones (8–10). Other infants' health conditions arisen after birth did not exclude newborns from eligibility but were considered as confounders. Newborns from single mothers were excluded because fathers' support can influence mothers' breastfeeding decisions and behavior (11, 12). Furthermore, several confounders considered in the analyses were relevant to both parents as a couple.

Lastly, we excluded women who asked for epidural analgesia for pain relief during labor and did not receive it due to technical problems. Indeed, it is not clear whether labor analgesia's repercussions on breastfeeding are related to the technique itself or to the conditions that lead to require it: the choice itself to request labor analgesia could be linked to difficulties in breastfeeding (13–16).

Maternal health conditions were not included in the exclusion or inclusion criteria but were considered as confounders.

Among those considered eligible at birth (n = 1834), at the time of discharge we enrolled 1,260 consecutive dyads: we recruited only those who were exclusively breastfeeding at 72 h from birth (in the case of twins, we considered the first born), and, according to the WHO definition of exclusive breastfeeding (17), did not receive formula milk or any other fluid during hospital stay, except for medicines, vitamins or minerals. Birth weight was collected between 2 and 6 h from birth, and at the nadir of weight loss, that is, between 48 and 72 h from birth (at discharge). On the basis of the calculated percentage of weight loss (WL) at nadir, newborns were subsequently divided into two groups: WL <7 and WL ≥7%.

The confounders that, according to the current literature, could be associated with early interruption of exclusive breastfeeding (18, 19) were entered into the analysis. Altogether, we identified 40 confounders, grouped as 23 antelabor and 17 peripartum confounders (Table 1), where peripartum confounders included intrapartum and postpartum ones. To be considered antelabor, a confounder had to: be present before labor initiation, have an alleged association with the endpoint (persistence of breastfeeding up to 6 months), and belong to a temporal sequence impossible to reverse.

Data were obtained from the database filled by the hospital's midwives assisting childbirth.

In most cases, previously published and validated cut-offs were used. If the percentile cut-off was used, the 10th and 90th percentiles were calculated for the entire population of women who delivered at our facility.

Our main endpoint was the duration of exclusive breastfeeding defined according to the WHO definition. The secondary endpoint was the occurrence of each event that led to the interruption of exclusive breastfeeding, according to the WHO definition: sporadic integration with formula milk (day on which an integration, even single, with formula milk occurs for the first time), complementary feeding (day on which breast milk begins to be supplemented daily with formula milk), shift to exclusive formula feeding (day on which exclusive formula feeding is started), and weaning (day on which any food other than breast milk or formula milk is introduced, including fruit).

After discharge, all mothers were reached on the phone by an anesthesiologist skilled in obstetric anesthesia and an International Board Certified Lactation Consultant (IBCLC) himself, at 10 days, 42 days and 6 months (≥183 days) from birth. Mothers were asked about what the infant's mode of feeding during the previous week had been, and time of introduction of liquids other than breast milk, or weaning. Mothers were asked the exact date on which the events under examination occurred, and these dates were reported in the database as follows: time of interruption of exclusive breastfeeding (identified as the time of occurrence of sporadic integration), start of complementary feeding, start of exclusive formula feeding, weaning.

Newborns were considered lost to follow-up if mothers could not be reached on the phone at any time point.

Figure 1 shows dyads' flow through the study. No differences in basal characteristics were found between mothers who were excluded from the study and those who were included (data not shown).

In our study population 68.7% of newborns lost <7% of birth weight at 72 h and mean WL at 72 h was 5.8% (+/−2.29). The WL <7% group lost on average 4.6% (+/−1.62) and the WL ≥7% group lost on average 8.4% (+/−1.07) of birth weight. As for mode of delivery, newborns born through cesarean delivery lost significantly more weight than newborns born through vaginal delivery (6.18 vs. 5.39%, respectively, p <0.001).

Table 1 lists antelabor and peripartum confounders. Missing data are declared variable by variable.

As demonstrated by Pearson chi-square test (Table 2), infants with WL ≥7% were exclusively breastfed at 6 months (day 183) from birth in a significantly lower percentage of cases than infants with WL <7% (p ≤ 0.001) whereas they received either sporadic integration with formula milk, or complementary feeding, or exclusive formula feeding, or were weaned through the first 6 months of life in a significantly higher percentage of cases.

Both the interruption of exclusive breastfeeding and each of the four events that determined it, occurred at a significantly earlier age in the WL ≥7% group than in the WL <7% group (Table 3).

The logistic regression confirmed the association between WL ≥7% and the interruption of exclusive breastfeeding before 6 months, reaching statistical significance on all considered outcomes [exclusive breastfeeding at day 183 from birth (95% CI 0.563 to 0.734, p < 0.001), sporadic integration with formula milk (95% CI 0.589 to 0.836, p < 0.001), start of complementary feeding (95% CI 0.460 to 0.713, p < 0.001), start of exclusive formula feeding (95% CI 0.587 to 0.967, p = 0.02), start of weaning (95% CI 0.692 to 0.912, p = 0.001)]. Among the confounders, no one was associated at the same time with all four outcomes (Table 4; relative risk, confidence interval and statistical significance are shown in Tables 5, 6).

By means of a Cox regression, a significantly different temporal course was observed between the two groups, for both time of interruption of exclusive breastfeeding and occurrence of sporadic integration, start of complementary feeding and start of exclusive formula feeding, but not for weaning (Figure 2).

We observed that WL ≥7% is significantly associated not only with a global reduction in the exclusive breastfeeding rate at 6 months, but also with all the four considered outcomes (i.e., increased frequency of sporadic integration, start of complementary feeding, shift to exclusive formula feeding, and weaning) up to 6 months. No one of the known described factors influencing exclusive breastfeeding at discharge has the same strong association at 6 months, significantly influencing only some of the four outcomes.

At our facility, certified as Baby Friendly Hospital (BFH) by UNICEF for 10 years in a row, we strictly follow the Ten Steps to Successful Breastfeeding, which are considered the optimal standard of in-hospital breastfeeding management. We observed that the average weight loss among healthy term exclusively breastfed newborns was 5.58%, inferior to that reported in other studies (20, 21), reflecting the importance of a breastfeeding supporting environment.

Additionally, only 31.3% of newborns in our study showed a WL ≥7%, similar to what has been reported by Bertini (22), whose hospital complied with the WHO guidance on infant feeding. Conversely, other researchers reported higher percentages, going as far as suggesting to raise the cut-off for physiological weight loss (21).

Breast problems only occurred in 15 of the 1,260 mothers enrolled, significantly contrasting with the incidence of mastitis (ranging from 2.5 to 20%) (23) and cracked nipples (24) published in previous studies. We can explain this interesting observation by specifying that in Aosta Valley women are followed extensively on the territory for a year after childbirth, by midwives trained on breastfeeding. Again, support seems to be crucial, even in preventing breast problems.

Interestingly, a WL ≥7% significantly impacts weaning when considering the absolute number of infants exclusively breastfed at 183 days from birth, yet survival curves describing the trend over time of the event “weaning” do not significantly differ. This discrepancy could be explained taking into account the incorrect but unfortunately widespread practice to start weaning in any case as early as 5 months. On the other hand, it seems to suggest an influence of weight loss on this same cultural factor: both the mother and health professionals may be less in a hurry to wean a baby with lower initial weight loss, thus reaching the goal of 183 days of exclusive breastfeeding.

Mode of delivery (cesarean delivery vs. vaginal delivery) does not influence persistence of exclusive breastfeeding at 6 months, as well as the other considered outcomes. We may infer that a strong breastfeeding support, offered by a multidisciplinary team (obstetricians, gynecologists, pediatricians, anesthetists, nurses) trained according to the UNICEF standards of the BFH project, can overcome the difficulties determined by the circumstances of birth (25).

In this regard, we observed that, consistently with current literature, in our cohort newborns from cesarean delivery lost significantly more weight than vaginally delivered newborns (6.18 vs. 5.39%), but average WL measured in our cohort was lower than that reported in the literature (26, 27).

It is interesting to note that authors describing major WL after cesarean delivery do not describe intraoperative fluid balance (26–28).

The observation that mode of delivery does not impact the persistence of exclusive breastfeeding at 6 months further reassures us on the validity of our intraoperative protocols, which consist in the administration of crystalloids (<100 ml per h of mother's preoperative fast before the umbilical cord clamping) and recommending mothers to drink clear fluids up to 4 h before surgery, thus avoiding excessive fluid administration during delivery. Limiting the maternal volume expansion may also protect from delayed lactogenesis II, which has been described after cesarean delivery (29).

We must also specify that at our hospital we weigh newborns for the first time from 2 to 6 hours after birth, that is after the initial period of diuresis, obtaining what Flaherman defines “the baseline dry weight” (30) and limiting bias related to intravenous fluid administration to the mother (31).

Among maternal illnesses, only severe preeclampsia seems to be related to the introduction of formula milk and start of exclusive formula feeding, perhaps because mothers in those conditions are often admitted to the intensive care unit and/or experience a delayed and less effective lactogenesis.

Being the only facility in the region is both a strength and a limit of our hospital: if on the one hand it allows us to study the entire population born in the region and ensure that every dyad of our cohort received the same standardized care, on the other hand it determines a population selection that could compromise the reproducibility of our results in other settings.

Additionally, the BFH designation of our hospital could have introduced a further selection, inducing foreign women to choose our facility to deliver. To control this selection bias, we included residence as a confounder in the multivariable analyses.

Another limit of our study is that the choice of considering, in the case of twins, only the first born, could introduce a bias, since a mother who has to care for two newborns cannot be as dedicated as one who only has one.

Finally, since the dates of occurrence of the events analyzed were provided by mothers in retrospect, they could be inaccurate, since a recall bias cannot be excluded.

However, the great sample size, the large number of confounders considered, and the long follow-up period may be considered the main strengths of our study.

In conclusion, a WL ≥7% in the first 72 h after birth is a simple and reproducible predictive factor for the early interruption of exclusive breastfeeding in healthy term exclusively breastfed newborns.

This parameter could aid the timely identification of dyads at risk of early interruption of exclusive breastfeeding, thus allowing health professionals to offer targeted care.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Ethics Committee from Azienda usl Valle d'Aosta. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

ED and RW contributed to conception and design of the study. RW and FV organized the database. LP performed the statistical analysis. ED, RN, and AC wrote the first draft of the manuscript. MG, MB, DM, FM, and LM wrote sections of the manuscript. All authors contributed to manuscript revision, read, and approved the submitted version.

This research was received funding from University of Milan, Dipartimento Scienze Cliniche e di Comunità, for open access publication fees.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.