The AAP (1) strongly recommends that human milk is the preferred nutrition not only for healthy term infants but for all infants, including premature and sick newborns, with rare exceptions (1). For an excellent in-depth examination of this topic, see Chapter 20. The reader is also referred to a recent review article on the use of human milk for premature infants (36).

There are, however, several points that are worth reiterating here. Preterm milk differs significantly from term milk (Table 21.3). Notably, preterm milk has higher concentrations of protein, fatty acids, sodium, and chloride (37,38), which, interestingly, are all components required in higher amounts by babies born early. This phenomenon was initially attributed to lower milk volumes produced by mothers of preterm babies, thus causing a concentrating effect on these nutrients. However, contrary to this theory, other components of preterm milk are present in the same concentrations as in term milk. It has subsequently been shown that preterm milk has similar volumes to term milk. Some initially speculated that this is a maternal adaptation to the delivery of her premature infant, although these differences are likely the end result of the interruption in maturation of the mammary gland during pregnancy. In a study looking at total nitrogen, fat, lactose, and carbohydrate concentrations in human milk, GA at birth was inversely related to carbohydrate concentration; postmenstrual age (PMA; an indicator of autonomous developmental processes not affected by the moment of birth) was not related to milk composition; and postnatal age (PNA) was related to a decrease in total nitrogen and an increase in lactose concentration. These data were interpreted to indicate that PNA strongly influences the development of the composition of very preterm human milk, GA affects carbohydrate content with a negligible effect on the nutritional value of the milk, and PMA has no effect (39) consistent with changes in maturation of the mammary gland.

The higher concentration of nutrients in preterm human milk decrease to approximately term milk levels over the course of the first postnatal month, regardless of the GA of the baby at delivery, although the premature baby’s increased needs continue until approximately term corrected GA. With the advantage of the earlier higher concentrations, particularly of protein and electrolytes, that premature milk affords initially no longer present, we are often required to fortify human milk for the smallest babies after reaching adequate volumes. Babies less than 1,500 g have been shown to require fortification with additional calories, protein, calcium, phosphorus, sodium chloride, and some vitamins to preclude poor growth rates, hyponatremia, hypochloremia, and osteopenia (36). Larger, more mature babies thrive on mother’s milk alone. Because human milk content differs not only over the course of lactation but also during the course of a feed or milk expression session, at different times of the day, and for those babies requiring feeding by alternative methods, by the method used, it is critical to monitor these VLBW babies for growth rates, serum sodium levels, and bone mineralization status (see also Chapters 20 and 33). Recent
discussions have suggested that due to variations throughout the day in some human milk components, it might be prudent to have mothers collect their milk in 24-hour aliquots, which, when gently mixed, would theoretically yield a less variable feed. This has yet to be studied.

Human milk protein is 80% whey, as opposed to bovine milk protein, which is 80% casein. The whey in human milk, α-lactalbumin, is much more easily digested than is bovine whey, β-lactalbumin, an important factor to consider for premature babies with immature gut function. Additionally, human milk protein also includes nucleotides, secretory immunoglobin A (sIgA) and other immunoglobulins, and an enzyme, lysozyme, all of which are thought to aid in host defense; growth factors that stimulate gut growth and maturation; a variety of hormones; and enzymes (e.g., mammary amylase, lipases) that enhance the ability of the immature intestinal tract to digest nutrients. The amino acid taurine, which serves many functions in the newborn including bile acid conjugation, osmoregulation, and neurotransmission, and acts as an antioxidant and a growth factor, is present in high concentrations in human milk and is almost absent in bovine milk. Hence, it is now added to artificial feedings. Unlike bovine milk, human milk is also low in phenylalanine and tyrosine, which the premature and newborn infant are poorly equipped to metabolize. For these reasons among others, human milk protein composition is well adapted to the needs of premature babies (see Chapters 20 and 22) (40).

Human milk lipids have sparked great interest in milk components, with the increasing body of literature on the positive neurodevelopmental effects of the long-chain polyunsaturated fatty acids (LC-PUFAs), in particular docosahexaenoic acid (DHA) and arachidonic acid (AA). They are found in phospholipids in the brain, retina, and red blood cell membranes. These LC-PUFAs are not readily synthesized by preterm babies and are normally delivered via the placenta throughout the third trimester. They occur naturally in human milk but are not found in bovine milk. For premature babies, LC-PUFAs must be delivered via an external source, in this case easily by human milk. Because of studies that show improved neurodevelopmental outcome and visual function in breastfed preterm babies and in formula-fed preterm babies supplemented with exogenous sources of LC-PUFAs, DHA and AA have now been added commercially to most term and preterm infant formulas. Of concern for ultimate function is that these additives are of plant origin and are structurally different than human LC-PUFAs (see “Neurodevelopmental Advantages”).

Human milk lipids are an easily digested source of energy, in part as a result of their composition and in part as a result of their packaging with bile salt-stimulated lipases present in the milk, providing approximately 50% of the total milk calories. Additionally, they provide cholesterol, which is an essential component of membranes. Human milk-fed babies show significantly higher plasma cholesterol levels than do formula-fed or mixed-fed babies (41). Although one might then expect human milk-fed babies to have higher cholesterol levels than do formula-fed babies as young adults, the opposite has been found to be true, in addition to lower C-reactive protein, a measure of the inflammatory process associated with atherosclerosis (42). Additionally, one study has shown men who were breastfed as infants to have better endothelial function in young adulthood than did men who were formula fed by showing that breastfeeding was inversely associated with atherosclerosis, measured by the thickness of the intima-media, arterial distensibility, or the prevalence of carotid plaques (43). It has been postulated that this early exogenous exposure to cholesterol, a necessary nutrient, keeps endogenous cholesterol production downregulated, thus resulting in lower cholesterol levels in adult life. The mechanisms remain to be elucidated.

The disaccharide lactose is the predominant carbohydrate in human milk. It is a ready source of energy and is broken down by the enzyme lactase, located in the brush border of the intestinal mucosa, to galactose and glucose, necessary for energy supply to the rapidly growing brain. Lactase activity is low in premature infants but is readily inducible by exposure to lactose, enabling them to absorb more than 90% from human milk. The remaining unabsorbed lactose contributes to softer stool consistency and colonization of the gut by nonpathogenic fecal flora. Lactose also enhances calcium absorption, critical to preventing nutritional rickets in prematures. Oligosaccharides, the third most abundant component of human milk, act in host defense by preventing bacterial attachment to intestinal mucosa, thus serving a protective role for the relatively immunocompromised preterm infant. More than a hundred different oligosaccharides have been identified, and their composition varies between mothers as well as over the course of lactation (44).

Several older studies have suggested that human milk-fed term (45,46) and preterm (47,48) infants have lower sleeping energy expenditure compared with formula-fed infants. A randomized, cross-over study of gavage-fed preterm babies showed significantly lower energy expenditure in the human milk-fed babies at prefeeding, during feeding, and postfeeding measurements (49). With the vast difference in composition of the nutrients and other factors between human and artificial milk, it is not possible to say what causes this difference, but it is an intriguing question that remains to be investigated.

In addition to the species specificity and superior digestibility of the nutrients in human milk for premature babies, human milk also favorably affects the function and maturation of the GI tract. Human milk has been shown in vivo to decrease intestinal permeability in preterm infants when compared to preterm formula (50). Several studies have found that human milk promotes more rapid
gastric emptying than does artificial formula (51,52), with one study revealing that on average, human milk emptied twice as fast as formula (52). This has implications for clinical practice. Delayed gastric emptying, which generally presents clinically as measurable gastric residuals or vomiting, prevents advancement of enteral nutrition. Babies who cannot reach full enteral feeds require longer periods of parenteral nutrition, with the concomitant risks inherent in both prolonged intravenous catheter usage (e.g., infection, thrombosis, chemical infiltrates) and prolonged intravenous nutrition (e.g., mineral or electrolyte imbalance, hepatic damage). Any of these can impact length of stay (LOS), which in turn has economic and social/family implications, all of which are important to consider in the therapeutic plan in our NICUs today.

Another related finding is the induction of lactase activity by feedings. Lactase, the enzyme responsible for the digestion of lactose, is present in the fetal intestine early in gestation, showing the greatest increase during the third trimester. Premature babies are relatively lactase deficient at birth. In one study, lactase activity was induced in preterm babies (26 to 30 weeks of gestation) by the initiation of enteral feeds. Of greatest significance, the highest levels of enzyme activity were seen with the introduction of “early” feeds (4 days old) as opposed to “standard” feeds (15 days old) and in human milk-fed versus formula-fed babies (53). There was also an inverse correlation between lactase activity at 28 days and the time to achieve full enteral feeds. It appears that the level of lactase activity may be a marker of intestinal maturity, with a direct relationship between human milk use and the progression of gut maturity.

There are a large number of “bioactive components” in human milk that are not present in formulas. They variously provide anti-inflammatory effects, provide protection from infectious agents, are hormones and growth factors that influence development, or are immune function modulators (54,55,56). This is an active area of research. At least a few of these bioactive factors have activity suggesting that they may be involved in GI maturation, growth, and motility (54). Epidermal growth factor (EGF) is a major growth-promoting cytokine that stimulates proliferation of intestinal mucosa and epithelium and strengthens the mucosal barrier to antigens (54). In an animal model in 1993, EGF isolated from human milk was shown to facilitate gut healing after induced injury (57). Since then, EGF has been shown to be critical to the maturation and healing of the intestinal mucosa. Resistant to low pH and digestive enzymes, it stimulates intestinal enterocytes to increase deoxyribonucleic acid (DNA) synthesis, cell division, absorbance of water and glucose, and protein synthesis. EGF inhibits programmed cell death and corrects alterations in intestinal and liver tight junction proteins induced by proinflammatory tumor necrosis factor (TNF)-α. Heparin-binding EGF is the primary growth factor responsible for damage resolution following hypoxia, ischemia-reperfusion injury, hemorrhagic shock/resuscitation injury, and necrotizing enterocolitis (NEC). The average EGF level in colostrum is 2,000-fold higher, and in mature milk is 100-fold higher than in maternal serum. Preterm milk contains higher levels of EGF than does term milk (54). Other factors identified in human milk, known as human growth factors I, II, and III and insulin-like growth factor, have been shown to have growth-promoting functions, including stimulation of DNA synthesis and cellular proliferation. In vivo studies in animal species have shown remarkable increases in the mass of intestinal mucosa after feedings with colostrum, which contains these factors, but not after feedings with artificial milk (58,59).

Another of the extraordinary advantages of the use of human milk over formula in the preterm and NICU patient is the effect on host defense and infections. This advantage alone is enough to make the use of human milk the standard in this population. The myriad of hormones, factors, cytokines, proteins, enzymes, nucleotides, antioxidants, immunoglobulins, and live, functioning cell types such as lymphocytes, macrophages, neutrophils, and natural killer cells, as well as probiotic bacteria contained within human milk (54,55,56), their interactions, and the milieu in which they effect action, is not and cannot ever be duplicated in an artificial milk. The more we discover, the more we find we have yet to uncover.

One can think of human milk not only as the perfect form of nutrition but also as “baby’s first immunization.” One type of immunization is the passive transfer of antibody. Examples include the use of intravenous immune globulin, tetanus immune globulin, or rabies immune globulin. The immunoglobulins represented in a mother’s milk are a history of many of the infectious agents she has been exposed to, and developed antibody to, during her lifetime. By transferring these antibodies to her baby through her milk, she is in effect “immunizing” her baby against those organisms. This process is a continuum from the one begun in utero by the passage of maternal antibody across the placenta. Taking this one step further is the concept of the enteromammary immune system, first proposed in 1979 by Kleinman and Walker (60) (Fig. 21.2). In this schema, antigen presented to the maternal mouth and gut is brought into the proximity of lymphoid follicles in the maternal GI tract known as gut-associated lymphoid tissue (GALT). The antigen’s presence commits maternal lymphoblasts to specific IgA production against that antigen. These lymphoblasts migrate via the mesenteric nodes and the thoracic duct into the systemic circulation, in which they find their way into the active mammary tissue. There, these cells manufacture sIgA, which is secreted into the milk. When the infant ingests the milk, the immunoglobulin present functions in the infant’s gut as protection against a specific pathogen. Most sIgA is not absorbed in the infant’s intestine but rather plays an active role in mucosal defense. Although intact immunoglobulin has been found in infant urine, signifying some systemic absorption, most survives intact in the GI tract and is excreted into the feces. The environment of the NICU is full of potentially pathogenic organisms. What makes this concept of the enteromammary immune system even more appealing is that during skin-to-skin (kangaroo) care, a neonate is held by the mother at her chest and is touched, kissed, and caressed. The mother exposes herself to any potential pathogens to which her baby has also been in contact. Through the enteromammary immune system, she makes antibody to those particular organisms and then, in subsequent feedings, she passes that antibody to her baby. Imagine—individually prepared immunizations to help protect each infant in the NICU!

Similarly to our previous discussion of concentrations of specific nutrients, many of these immune modulators are in higher concentration in preterm milk than term milk, helping to compensate for the premature baby’s immature immune function (Table 21.4). When major factors were quantified and compared between colostrum of mothers delivering both prematurely (28 to 36 weeks) and at term (38 to 40 weeks), mean concentrations of IgA, lysozyme, lactoferrin, and absolute counts of total cells, macrophages, lymphocytes, and neutrophils were found to be significantly higher in the preterm colostrum (61). The degree of prematurity did not influence the anti-infective levels in the colostrum. However, the total cells and macrophages were significantly higher in the colostrum of mothers delivering at 28 to 32 weeks of gestation compared to 33 to 36 weeks of gestation (p < 0.05) (61). These differences make the use of early colostrum and human milk critical in the care of premature and sick babies, both as prevention and possibly in treatment of infectious disease.

Another factor in improved immune defense with the use of human milk is the affect on fecal flora. The normal flora of the intestines of a breastfed infant is predominantly the grampositive bacterium Lactobacillus bifidus. Non-human milk-fed babies are colonized with many more types of bacteria, most of which are potentially pathogenic gram-negative bacteria. With establishment of Lactobacillus as the predominant flora inhabiting
the premature infant’s GI tract, the likelihood of a serious or life-threatening gram-negative infection would be expected to decrease.

It has long been known that breastfed babies are at decreased risk of a number of infectious diseases—respiratory infections, otitis media, gastroenteritis, and diarrhea—and are at decreased risk of mortality. These protections accrue in a dose-dependent fashion —that is, the more human milk a baby receives, statistically the better protected they are. Although much of the earlier work was done in developing countries, there are now also many studies that show significant impact in populations in the developed world (32,33). As discussed previously, one study estimates 911 excess deaths per year in the United States due to lack of breastfeeding, nearly all of whom are infants (34). It is probably fair to assume that babies admitted to NICUs, once attaining term corrected GA and discharged to home, will accrue similar advantages from breastfeeding/human milk as will their healthier term counterparts in these studies. But even more importantly for our population, an increasing body of research has accumulated examining the effects of a diet of human milk compared to preterm formula in premature and low-birth-weight babies, with respect to clinical infections. The data are clear—premature infants fed human milk are at significantly less risk of serious diseases, including NEC, urinary tract infections, sepsis, and meningitis.

As early as 1971, a report from Sweden showed a protective effect of breastfeeding against sepsis in the newborn (62). Then, in 1980, Narayanan and her group in India reported that even partial use of human milk (63) and subsequently exclusive use (64) could significantly decrease the incidence of infection in a premature and low-birth-weight population. Infections noted were sepsis, diarrhea, pneumonia, meningitis, conjunctivitis, pyoderma, thrush, and upper respiratory infections. The strongest effect was noted in those babies exclusively fed human milk, followed by those fed partial feedings of human milk, thus indicating a dose-response effect. Hylander et al. examined the incidence of infection among VLBW infants, in relationship to their type of feeding, while controlling for confounding factors. They found that the incidence of infection (human milk 29.3% vs. formula 47.2%) and sepsis/meningitis (human milk 19.5% vs. formula 32.6%) differed significantly by type of feeding (Fig. 21.3) (65).

In 1999, Schanler et al. (66) showed that in a group of premature infants between 26 to 30 weeks of gestation, those fed predominantly fortified human milk were discharged earlier (73 ± 19 vs. 88 ± 47 days) and had lower incidence of NEC and late-onset sepsis than did infants fed preterm formula. These data on NEC
confirm a previous study by Lucas and Cole (67), who showed that in a cohort of 926 infants with birth weights below 1,850 g, formula-fed babies were 6 to 10 times more likely to develop NEC than were those fed human milk exclusively and three times more likely than were babies who received a combination of human milk and formula supplements, again demonstrating the dose-response effect of human milk. Recent studies show that using an exclusively human milk diet (human milk, either mother’s own or donor with human milk-derived fortifier if required) can reduce the incidence of NEC to virtually zero (68,69). Further study of the use of donor milk compared to mother’s own milk and exploration of issues pertaining to the high cost of the currently single manufacturer human milk-derived fortifier (which precludes its use by many NICUs) is required. While reviewing the data on NEC, it is also important to point out that, in addition to these significant clinical decreases in incidence, there are also concomitant significant economic savings. This is an enormous issue to practitioners, to the health care system as a whole, and certainly to the individual families we care for. A reduction in the cases of NEC would engender significant savings in medical charges and LOS. Bisquera et al.(70) showed that infants with surgical NEC exceeded LOS by 60 days over matched controls and those with medical NEC by 22 days over controls. Based on LOS, the estimated total hospital charges per baby for surgical NEC averaged $186,200 more and for medical NEC $73,700 more than controls. This translated into yearly additional hospital charges at their institution for NEC of $6.5 million or $216,666 per survivor. A model using data from this study showed that, given these costs of NEC and the cost of providing human milk-derived fortifier, significant cost savings could be appreciated by treating all infants less than 1,250 g birth weight with exclusively human milk products (71).

It appears clear that an exclusively human milk diet is protective in fragile preterm babies. In the previously discussed Hylander et al. study (65), the human milk-fed group also received supplements of formula and still showed significant benefit. Lucas and Cole showed that looking at NEC, partial human milk feeding was protective but less so than exclusive human milk feedings (67). Furman et al. (72), looking at 119 VLBW babies, identified a daily threshold amount of at least 50 mL/kg/d of maternal milk through week 4 of life as needed to decrease rates of sepsis in this group. It has also been shown that upper respiratory symptoms are reduced for those low-birth-weight babies who continue to receive human milk after discharge from the NICU through 7 months corrected age and also decreased occurrence of rehospitalization in the first year of life (73). Although not significant, a trend also appears to exist for otitis media, bronchiolitis, and gastroenteritis. More studies with larger cohorts are needed to confirm these data.

There is no longer any doubt that by providing human milk, we do make a difference in the incidence and impact of serious potential morbidities in even the tiniest of our patients in a dose-response manner. That being the case, with all the potential morbidities that our patients face, and the costs both economically and emotionally for these families and society, it is imperative upon us to make human milk the standard for nutrition in neonatal intensive care.

Improved cognitive development has been demonstrated in premature babies who receive human milk. In 1988, Morley and associates showed an 8-point cognitive advantage using the Bayley Scales of Infant Development for 771 infants with birth weights less than 1,850 g (74). After controlling for demographic and perinatal factors, a 4.3-point advantage remained. When this cohort was followed up at 7.5 to 8 years of age, infants who had received human milk by tube (rather than breastfeeding) continued to show an 8.3-point IQ advantage (over half a standard deviation) even after adjustments were made for difference in mother’s education and social class (75). They also showed a dose-response relationship between the proportion of human milk in the diet and subsequent IQ. There has been one meta-analysis of controlled studies looking at the question of human milk and cognitive development, which demonstrated a 3.16-point higher score for cognitive development in human milk-fed babies compared with formula-fed babies after adjustment for significant covariates (76). This difference was observed as early as 6 months and was sustained through 15 years of age, the last time of reliable measurement. Longer duration of breastfeeding was accompanied by greater differences in cognitive development (dose-related response). Whereas normal weight infants showed a 2.66-point difference in IQ scores between breastfed and formula-fed groups, a 5.18-point difference was demonstrated in low-birth-weight infants. The results of this meta-analysis suggest that not only does human milk contribute to optimal neurodevelopment but the effect is even more striking in the higher-risk low-birth-weight infants.

Vohr et al. (77,78) followed a cohort of 1,035 babies who were less than 800 g at birth, enrolled in the National Institute of Child Health and Human Development Neonatal Research Network Glutamine Trial at 18 and then 30 months corrected age. Multivariate analyses, adjusting for confounders, confirmed a significant independent association of human milk on all four primary outcomes: the mean Bayley Mental Development Index, Psychomotor Development Index, Behavior Rating Scale, and incidence of rehospitalization. At 18 months, for every 10-mL/kg/d increase in human milk ingested, the Mental Development Index increased by 0.53 points, the Psychomotor Development Index increased by 0.63 points, the Behavior Rating Scale percentile score increased by 0.82 points, and the likelihood of rehospitalization decreased by 6%. In an effort to identify a threshold effect of human milk, the mean volume of human milk per kilogram per day during the hospitalization was calculated, and infants in the human milk group were divided into quintiles of human milk ingested adjusted for confounders. As every 10 mL/kg/d human milk contributed 0.53 points to the Bayley Mental Development Index, the authors suggest that the impact of human milk ingestion
during the hospitalization for infants in the highest quintile (110 mL/kg/d) on the Bayley Mental Development Index would be 10 × 0.53 or 5.3 points. They postulate that an increase of 5 IQ points in this population (one-third of a standard deviation) could be significant enough to impact need for early intervention and special education services (77). At 30 months corrected age, the benefits remained. For every 10 mL/kg/d increase in human milk, the Mental Developmental Index increased by 0.59 points, the Psychomotor Developmental Index by 0.56 points, the total behavior percentile score by 0.99 points, and the risk of rehospitalization between discharge and 30 months decreased by 5% (78).

Postulated to be responsible for at least a portion of these neurodevelopmental advantages of human milk is the presence of LC-PUFAs, which until recently, were not present in formulas (see also Chapter 20 for a more in-depth review). DHA normally accounts for greater than one-third of the total fatty acids of the gray matter of the brain and the retina of the eye (79). Most of the prenatal accumulation of DHA in these tissues occurs in the third trimester —thus by definition, premature infants are deficient compared to their term counterparts. Animal studies have shown that deficiency of DHA in neural tissues during development leads to behavioral and retinal changes.

Other examples of the effects of human milk on neurologic maturation have also been observed. Premature infants receiving human milk have been shown to have faster brainstem maturation than do those receiving formula (80). Visual acuity and the development of retinopathy of prematurity (ROP) have also been studied. A number of studies were performed before the routine addition of the LC-PUFAs to preterm formulas. In one, improved retinal function was present in LC-PUFA sufficient VLBW neonates fed human milk or supplemented formula, as compared to a group receiving unsupplemented formula (81). Better visual evoked potentials and acuity in both preterm and term infants at 57 weeks postconception have been seen in those fed human milk as opposed to those fed formula (82). Additionally, another interesting study in which healthy term infants who breastfed to 4 or 6 months, and then were weaned, were randomly assigned to commercial formulas with or without DHA and arachidonic acid (AA) supplements. At 1 year of age, the level of DHA as measured in the red blood cells was reduced by 50% from weaning level in the unsupplemented group, although there was an increase of 24% in the supplemented group (83). The conclusions drawn from this study were that the critical period during which a dietary supply of DHA and AA can contribute to optimizing visual development in term infants extends through the first year of life. This supports the AAP recommendation that breastfeeding continue through the first year of life (1), and begs consideration then of the length of time breastfeeding should be encouraged and supported for the baby born prematurely. The incidence of ROP in relation to human milk feedings has been investigated. A secondary analysis of data collected during two multicenter randomized controlled trials (RCTs) in Italy revealed that ROP incidence (at any stage) was significantly lower in infants fed maternal milk as compared to formula-fed neonates; the same difference in incidence held for threshold ROP (84). Another recent study looking at short-term outcomes in preterm infants detected lower rates of ROP in a subgroup of breastfed infants born at 24 to 28 weeks’ GA; the difference did not reach statistical significance using univariate analysis (p = 0.06). However, using multivariate analysis, the incidence of ROP stage III among this subgroup was significantly lower (p = 0.022) (85). The etiology of ROP is clearly complicated and is not based simply upon the nature of the enteral feedings provided. The components in human milk that have the protective effects, be they the LC-PUFAs or bioactive factors, require further study.

Breastfeeding is widely assumed to be more stressful than is bottle-feeding for premature infants; an assumption that continues to lead, in many intensive care units, to the introduction of bottle-feedings as the first oral feeds, and postponing attempts at breastfeeding until babies “can prove themselves with the bottle.” There are also often concerns for the small premature baby’s ability to maintain temperature while breastfeeding, leading to “rules” restricting initiating breastfeeding until a certain weight is achieved. Additionally, there is the widely held belief that the “suck-swallow-breathe” mechanism is not mature until approximately 34 weeks of gestation. The fear that premature babies will choke, desaturate, and aspirate leads to more “rules” dictating that oral feeds, including breastfeeding, not be initiated until at least 34 weeks of corrected GA. It is important to point out that there is no scientific evidence to back any of these statements. On the contrary, the opposite is the case as we will discuss. It is also important to understand that these policies, in addition to being unsupported scientifically, are also harmful in that they (a) preclude premature babies and their mothers from early breastfeeding experiences; (b) allow babies whose mothers want to breastfeed to learn to suck from a bottle, which then in many cases makes it difficult for them to transfer to the breast; and (c) introduce breastfeeding so late in the hospital stay that, in addition to struggling to overcome what they have learned with an artificial nipple, the mother and baby are often discharged before they have had time to learn together and develop the confidence and skills necessary to allow successful breastfeeding, especially with the help they can receive from the NICU lactation staff. It is important to reemphasize our medical dictum in this circumstance: Primum non nocere.

Data do exist that enable us to develop breastfeeding policies that are physiologic, safe, and supportive of breastfeeding. As early as the 1980s, Meier was publishing data concerning physiologic stability at breastfeeding compared to bottle-feeding. She was able to show that in babies less than 1,500 g at the time of first feeding, different sucking mechanisms were employed for breast and bottle, with better coordination of suck-swallow-breathe during breastfeeding, particularly in the smaller, less mature babies. Concomitantly, there are markedly different patterns of transcutaneous oxygen pressure (tcPO₂) recorded for the two methods of feeding. TcPO₂ patterns suggest less ventilatory interruption during breastfeeding than during bottle-feeding, with greater declines in tcPO₂ recorded during bottle-feeding than during breastfeeding over the course of a feeding session. Additionally, infants stay significantly warmer during breastfeeding than during bottle-feeding (86,87,88). Similar studies done by Bier and colleagues in first VLBW (89) and then extremely low-birth-weight (ELBW) (90) infants showed that they could tolerate beginning both breast- and bottle-feedings at the same PNA, that they were less likely to have oxygen desaturations to less than 90% during breastfeeding, and that they had lower intakes during breastfeeding. The lower intakes have been observed in a number of studies. It is postulated that this is due to several factors: better paced control of suck-swallow-breathe during breastfeeding than during bottle-feeding where the artificial nipple releases fluid regardless of whether the infant is ready to accept a bolus or not, possible lower milk volumes in some of the breastfeeding mothers, and weaker sucks in some of the infants resulting in less milk released with breastfeeding. These may all be factors in the relatively safer physiologic state when breastfeeding than bottle-feeding. Given that there are no data supporting the safety of initiating feeding with artificial nipples, and that there are data to show that during breastfeeding premature babies are more physiologically stable, it is therefore reasonable, safe, and scientifically sound for mothers to place their babies nuzzling at breast and beginning early steps toward breastfeeding, once they are physiologically ready, and there is no basis for requiring bottle-feeding as first feeds, if mothers intend to breastfeed.

Expressing human milk and subsequently breastfeeding a premature infant has been described as a vital contribution to the care of her infant that only a mother can make. This is a tangible opportunity for these mothers, in a situation in which they otherwise feel powerless. Additionally, this feeding choice has been considered the one aspect of natural caregiving that the mother does not have to forfeit when she delivers a sick or premature infant. In the past, the decision to provide expressed human milk for a premature or sick infant was regarded as a matter of parental choice, with professional responsibility limited to implementing the parent’s decision. With increasing and irrefutable evidence that human milk supports the optimal health, development, and nutrition of sick and premature infants, obstetric and NICU care providers must actively encourage mothers to initiate lactation and provide milk to their infants, even if they do not plan to ultimately feed their infant at the breast.

Parental decisions related to infant feeding must be based on informed choice. Emphasizing the importance of human milk to the health of the infant assists mothers and their support persons in making an informed decision about feeding method. Health care providers should share research-based evidence about the superiority of human milk and the critical value of this intervention for the sick or premature infant. Mothers should be assured that all efforts to support her expressing colostrum and developing a milk supply will be made (see section on “Initiating and Maintaining Lactation with a Breast Pump” below). Optimally, donor milk (see “The Use of Donor Human Milk” below) is available in the NICU to initiate feeds if mother’s milk is not yet available or to supplement mother’s milk later as volumes advance. It is also important to communicate the maternal benefits of lactation to these mothers, both in terms of her immediate ability to be actively involved in the care of her baby, no matter how sick he or she is, and in terms of the health benefits the mother will accrue (1,2,3,4,5). Optimally, this information should be discussed with the family as soon as the premature birth and/or the NICU admission of their baby becomes a possibility. Sometimes, time will allow for a prenatal consultation; sometimes, this discussion occurs in the labor and delivery unit, with the baby’s birth imminent. With the understanding that these options may not be feasible, the discussion should occur as soon after the infant’s birth as realistic, so that milk expression can be initiated and colostrum expressed when the hormones controlling lactation are optimal. Even mothers who have indicated prior to the birth of their child that they desire to feed formula are frequently very willing to initiate milk expression to obtain colostrum for their babies when they understand all the advantages it confers. Written information should emphasize that early human milk is considered “medicine” that prevents infection and helps mature the GI tract in this vulnerable population. Although staff may worry that this approach is coercive or may make mothers feel guilty about initial feeding plans, mothers who are educated to make an informed choice to provide human milk for their infant report that they are thankful for the information that assisted them to support the health of their infant. Mothers who are not given this information initially have later reported being angry that they were not given the knowledge to make a choice. It is important that families understand that a decision to initiate milk expression does not commit them to breastfeeding; they can decide to stop at any point along the continuum. This knowledge is very empowering to these women who, with the birth of a baby who requires intensive care, acutely feel the loss of normal maternal control. Research has demonstrated that structured education and support programs within NICUs are effective in increasing lactation initiation rates, often in mothers who had intended to formula feed (91). Consistent messaging cannot be overemphasized. One review looked at types of support required by mothers providing their milk for infants in the NICU. Results revealed that emotional and practical support for NICU mothers differs from that provided for other breastfeeding mothers. Mothers of infants in the NICU require significant ongoing emotional support from health care providers who also need to monitor their milk production and provide educated encouragement that anticipates breastfeeding challenges, especially when the mother is pumping for an extended period of time while her infant is hospitalized (92). According to parents’ perspectives, successful milk supply in the NICU depends on coherent and accurate knowledge about techniques and benefits, reinforcement of mothers’ motivation, and alignment between NICU routines and parental needs. Parents perceived their own current milk supply experience and parent-professional relationships as either helpful or not helpful, depending on the circumstances (93). The positive breastfeeding outcomes of NICU programs that concentrate on the consistent support of breastfeeding families clarify the critical role of health care providers in sharing the science of human milk with mothers so that they can make an informed choice related to their infant’s feeding.

Mothers who provide their milk and eventually breastfeed their premature infants must develop their milk supply through manual or mechanical milk expression techniques. These techniques are suboptimal compared to actually breastfeeding in stimulating and maintaining a full milk supply. As a result, compromised milk production is a well-documented constraint to extended breastfeeding in the premature mother-infant dyad. An understanding of the physiology of lactogenesis and the effects of various pumps and pumping styles on milk production is critical to understanding this issue and supporting mothers to establish and maintain optimal milk volumes.

The transition from pregnancy to lactation is called lactogenesis. The first half of pregnancy is characterized by growth and proliferation of the mammary ductal system; during the second half of pregnancy, secretory activity increases and the alveoli become distended by accumulating colostrum (40). The capacity of the mammary gland to secrete milk after approximately 16 weeks of gestation is referred to as lactogenesis I. The onset of copious milk secretion 2 to 8 days after birth is called lactogenesis II. Triggered by a rapid drop in serum progesterone levels after the delivery of the placenta, lactogenesis II results in a rapid increase in milk volumes from 36 to 96 hours postpartum. Ongoing lactation is dependent on a delicate interplay of hormones and effective stimulation and emptying of the breast. Interference with these processes can delay and/or suppress milk production (40).

Prolactin and oxytocin are the predominant hormones of lactation (Fig. 21.4). Prolactin, the milk-making hormone, is essential for both initiating and maintaining milk supply. As progesterone and estrogen levels abruptly drop after birth, the anterior pituitary gland is no longer inhibited by these hormones and releases pulsatile surges of prolactin in response to suckling at the breast. Frequent breastfeeding in early lactation stimulates the development of prolactin receptors in the mammary glands and results in a faster increase in milk synthesis (40). The number of prolactin receptors increases in early breastfeeding and remains constant thereafter (94,95). Therefore, early breast stimulation during this critical period is a predictor of later milk production. Oxytocin is the hormone responsible for the removal of milk from the breast. Excreted by the posterior pituitary gland in response to suckling, oxytocin causes the myoepithelial cells surrounding the alveoli to contract and milk to be ejected into the ducts from which it is available to the newborn. The secretion of oxytocin is negatively affected by stress and pain. Therefore, the birth of a premature
infant (often by cesarean delivery) and the subsequent stressors associated with this event can interfere with the milk ejection reflex or “let down” of milk from the breast and ultimately interfere with milk production.

After the first few days postpartum, lactation shifts from endocrine control (hormone driven) to autocrine control (driven by milk removal). Galactopoiesis (the maintenance of milk production) is driven by the quality and quantity of milk removal. As long as milk is being removed from the breasts, the alveolar cells will continue to make milk. This supply-demand phenomenon regulates milk production to match intake by the infant (40).

The birth of a premature infant can negatively influence milk production. If a mother has delivered very prematurely, mammary development may be poor because the mother may not have received the full component of pregnancy-related hormones to prepare the breasts for lactation (96). Additionally, the close infant contact most frequently experienced by mothers following term delivery is usually limited or absent after premature delivery. As a result, the neurohormonal stimulus of the lactogenic hormones is impaired. An additional barrier to optimal milk expression is the fact that anxiety, fatigue, and emotional stress, all powerful inhibitors of lactation, are experienced almost universally in mothers of premature infants (97). As a result, for mothers of premature infants, concerns related to milk production and transfer of adequate milk to the infant are primary reasons for discontinuing expressing/breastfeeding or providing supplements. Early, frequent and optimal stimulation of maternal milk supply, with emptying of the breasts via a pump and/or manual expression, must replace and replicate as much as possible the natural breastfeeding process to ensure adequate milk production and duration of breastfeeding in this population.