Cover Image
close this book Food and Nutrition Bulletin Volume 17, Number 4, 1996
close this folder Establishing lactation
View the document Breast development and control of milk synthesis
View the document Discussion of paper by Hartmann et al.
View the document Constituents of human milk
View the document Discussion of paper by Prentice
View the document Problems of establishing lactation
View the document Discussion of paper by Woolridge
View the document Social and biological determinants of lactation
View the document Discussion of paper by Villalpando et al.

Establishing lactation

Breast development and control of milk synthesis

Breast development and control of milk synthesis

Peter E. Hartmann, Robyn A. Owens, David B. Cox, and Jacqueline C. Kent


We have developed a computerized breast measurement system that can quantitate both long-term (lactation cycle) and short-term (between breastfeedings) changes in breast volume. The increase in breast volume during pregnancy was not related to milk production at one month of lactation, whereas milk production from one to six months of lactation remained constant and was not controlled directly by the suckling-evoked secretion of prolactin. From the measurement of circadian changes in breast volume, it was concluded that infants rarely emptied the breasts at a single breastfeeding and that short-term variation in the rate of synthesis during the day and between the left and right breasts was closely related to the degree of breast fullness. Furthermore, differences between women in the storage capacity of the breasts dictated their flexibility in frequency of breastfeeding. These observations are consistent with the autocrine (local) control of milk synthesis during established lactation in women.


Breastfeeding and breastmilk must be considered in the context of maternal physiology and infant development rather than just the narrow role of optimizing infant nutrition (table 1). The lactating breast has a high metabolic activity. Indeed, the energy output in milk represents approximately 25% of the total energy intake in the maternal diet for women exclusively breastfeeding single infants, and up to 50% for those breastfeeding twins [1]. In addition, milk is a very complex secretion, consisting of cells (leucocytes, macrophages, and epithelial cells), lipids (triacylglycerols, free fatty acids, phospholipids, sterols, hydrocarbons, and fat-soluble vitamins), carbohydrates (lactose, oligosaccharides, galactose, glucose, and glycoproteins), proteins (caseins, a-lactalbumin, lactoferrin, secretory IgA and other immunoglobulins, Iysozyme, enzymes, hormones, and growth factors), non-protein nitrogenous compounds (urea, creatine, creatinine, uric acid, amino acids including glutamine, nucleic acids, nucleotides, and polyamines), water-soluble vitamins, macronutrient elements, and trace elements [2,3]. Furthermore, it is now clear that the proportions of these constituents in breastmilk are uniquely appropriate for the human neonate at a time when growth and development are occurring at near-maximal rates, yet many of the infant's systems (such as the digestive, hepatic, immune, neural, renal, and skeletal systems) are functionally immature. In addition, breastfeeding is an integral component of the complex psychological and metabolic dependencies of the infant on its mother, with single physical functions, such as nursing, providing the stimuli of touch, balance, smell, hearing, and vision, and each having specific effects on the infant [4]. Nevertheless, the magnitude of these benefits of human lactation, by and large, has not been afforded appropriate prominence in either health-professional education or the specialist's medical literature (for example, ref. 5).

Successful lactation requires the development of fully functional mammary glands. Whereas other major organs are morphologically and functionally relatively mature at birth, the mammary gland undergoes very limited structural development in utero, with the most dramatic changes in women occurring during puberty, pregnancy, lactation, and weaning [6]. The latter three stages (pregnancy, lactation, and involution during and after weaning) form the phases of the lactation cycle (fig. 1). This cycle can vary greatly in length from a little over nine months to many years, and it may be repeated many times, depending on female fecundity.

TABLE 1. Functions attributed to breastfeeding and breastmilk

Breastfeeding and the infant
  • a unique food
  • augments the infant's digestive enzymes
  • stimulates growth and development
  • may modulate behaviour optimizes infant intelligence
  • protects against illness

Breastfeeding and the mother

  • reduces risk of breast cancer
  • may reduce body weight
  • suppresses maternal fertility

Foetal development

A comprehensive investigation of the foetal development of the human mammary gland was carried out by Dawson in 1934[7]. It was observed that a mammary band (milk streak) appeared as a raised portion of ectoderm on either side of the midline, extending from the axilla to the groin of the human embryo by about the fourth week of intrauterine life. This band contains a narrow ribbon of raised epithelial cells known as the milk line (mammary ridge). Whereas the mammary band generally disappears, the milk line diminishes in length from its caudal end, and the cranial extremity thickens into a small nodule of ectodermal cells in the thoracic region at about 6 to 7 weeks of age. The nodule composed of epidermal cells then sinks into the underlying mesenchymal tissue to form a mammary bud. By 12 to 16 weeks, the overlying skin no longer protrudes, allowing the formation of an indentation that ultimately forms the areola and nipple. From this time to birth, the mammary gland anlage buds to form a number of solid cords, which traverse the underlying mesenchymal and subcutaneous tissue. These cords branch and become canalized to become ducts at 20 to 24 weeks after conception. At the end of gestation, Russo and Russo [6] observed the development of very primitive lobular structures composed of ducts ending in short ductules, lined by one layer of epithelial cells and one layer of myoepithelial cells. The epithelial cells had fine cytoplasmic vacuolization containing lipid droplets and apocrine secretion, which was not confined to the primitive alveolar structures.

Although the endocrine control of foetal mammary gland development is poorly understood, the higher levels of androgen in male than in female foetal rats have been associated with the suppression of mammary development in the male foetus, and the administration of steroid hormones to pregnant mice and rats can induce abnormal mammary development in their offspring (see ref. 8). These findings strongly suggest that more attention should be given to the foetal environment as a potential cause of subsequent lactation failure.

FIG. 1. Breast development and the lactation cycle

Pre-pubertal development

"Witches' milk" is one of the very few pre-scientific terms still in current use and refers to colostrum-like fluid [9,10] that is secreted from the nipples of newborn infants. It was thought that infants secreting this "milk" were possessed by witches, and these infants were not favoured. It is now clear, however, that this is a normal transient event, as secretion can be expressed from the nipples of most infants by seven days after birth, and involution of the neonatal mammary gland is complete by eight weeks postpartum [10]. Throughout childhood, only isometric growth of the breast occurs, with limited elongation and branching of the ducts [7].

Pubertal development

In humans, unlike other mammals, extensive positive allometric growth of the breast occurs at puberty. From 10 to 12 years of age, girls enter puberty and, over time, develop to sexual maturity. Three phases of puberty have been identified: thelarche (commencement of breast development), pubarche (growth of pubic hair), and menarche (start of menstruation).

At thelarche, the ovaries of the immature female start to secrete oestrogen, which initiates positive allometric growth of the primitive mammary structures. The mammary ducts elongate and extend their epithelial lining, branching dichotomously, resulting in the formation of a branched, treelike structure that extends from the nipple into the mammary connective tissue. In addition, terminal buds form on the ducts in preparation for the development of alveolar and lobular tissue structures.

During thelarche, mammary elasticity, vascularization, connective tissue volume, and fat deposition increase [6], leading to the development of the characteristic shape of the mature human breast [11]. Compared with other mammals, mammary growth in pubescent girls is far in excess of the development required for subsequent successful lactation, and therefore it has been suggested that this growth is not related to lactation but rather provides an indicator of sexual maturity.

During puberty there is accelerated growth of the nipple and development of subareolar tissue leading to the elevation of the areola and nipple. About 15 to 25 lactiferous ducts lead to each nipple from discrete lobes within the mammary parenchymatous tissue. These lactiferous ducts may merge within the nipple so that the number of ductal openings is less than the number of lactiferous ducts. The lactiferous ducts dilate at the base of the nipple to form milk sinuses, which, during lactation, accumulate milk drained from the lobes. Unlike other mammals, the lobes of the human mammary gland are separated by deposits of adipose tissue, and the proportion of adipose to secretory tissue varies between individuals. Each lobe is subdivided into lobules that, in turn, are composed of 10 to 100 alveoli. The alveoli are lined with a single layer of lactocytes (mammary secretory epithelial cells) surrounded by starlike myoepithelial cells, a basement membrane, and a network of blood capillaries [9,12].

Under the cyclic influence of ovarian oestrogen and corpus luteal progesterone, as well as the presence of other metabolic and growth-promoting hormones [8, 12], the mammary glands are stimulated to grow. From three to four days before the onset of menstruation, women may experience swelling, tension, fullness, tightness, heaviness, and pain in their breasts [9]. Fluid retention in the connective tissue and enhanced ductular and lobulo-alveolar tissue growth increase breast volume by 30 ml [6, 9] to 100 ml [13]. Increases up to a doubling of breast size have been observed in some women. After menstruation the mammary gland is characterized by some apoptosis, with minimum breast volume occurring five to seven days post-menstrum [6]. Mammary involution following menstruation never completely returns the mammary gland back to the previous premenstrual morphology and, hence, allows the mammary parenchyma to develop, albeit gradually, during successive menstrual cycles [6] until women reach about 30 years of age. In spite of the functional responsiveness of the breast to stimulation [12], little information is available on breast development in relation to the commencement of sexual activity.

Lactation cycle

The lactation cycle begins at conception (fig. 1). During the lactation cycle, there is further growth of the breast (mammogenesis), the initiation of milk synthesis and secretion (lactogenesis 1 and lactogenesis 2), lactation (galactopoiesis), regression of the breast during and after weaning (involution), and relative quiescence during subsequent menstrual cycles.


The mammary gland develops the histologic and biochemical capacity to synthesize and secrete milk during pregnancy. Histologic studies have separated mammary development during pregnancy into two distinct phases: mammogenesis and lactogenesis 1. Mammogenesis occurs from early pregnancy and is characterized by proliferation of the distal elements of the ductal tree, creating multiple alveoli (acini) of variable size and shape [14]. Lactogenesis 1 occurs in the later stages of pregnancy and is characterized by the differentiation of resting mammary cells into lactocytes, with the potential to secrete the unique fats, carbohydrates, and proteins characteristic of milk [15].

Initially, mammary development during pregnancy appears to be an acceleration of the parenchymal hypertrophy associated with the menstrual cycle. Indeed, an increase in the sensitivity and tenderness of the breast, and nipple sensitivity in particular, is often one of the first indications of pregnancy, and this can occur within a few days of conception and before the due date of the next menstrual period. Thus, the factors initiating mammogenesis at this time must be closely related to those responsible for the mother's recognition of her pregnancy. Subsequently, the subcutaneous veins become enlarged and visible through the skin, and the areola usually enlarges and becomes more darkly pigmented [14]. Extensive lobulo-alveolar growth occurs during the first half of pregnancy, and in the third trimester there is a further increase in lobular size associated with hypertrophy of the lactocytes and the accumulation of secretion in the lumina of the alveoli [6].

Although there is little precise information on the hormonal control of mammary development in women during pregnancy, the changes in the patterns of circulating hormones are now well established. The implantation of the blastocyst in the uterine wall is associated with the secretion of human chorionic gonadotrophin (hCG), which maintains and increases the steroidogenic activity of the corpus luteum, until hCG secretion decreases at about 8 to 10 weeks of gestation. In the later stages of pregnancy, the maternal serum concentrations of progesterone and oestrogens are increased by de novo synthesis in the placenta [16].

In classical studies on ovariectomized-hypophysectomized-adrenalectomized rats and mice, Lyons [17] and others have shown that ductal mammogenesis is promoted by oestrogens, growth hormone, and corticosteroids. In addition, lobulo-alveolar development occurred at maximal rates in the presence of oestrogen, progesterone, prolactin, growth hormone, corticosteroids, and placental lactogen. Studies on isolated human mammary tissue in culture suggest that insulin, cortisol, growth hormone, prolactin, oestrogens, progesterone, and epidermal growth factor (EGF) are involved in the proliferation and differentiation of human mammary cells in tissue culture [18-22]. Relaxin has been implicated in the proliferation of porcine mammary parenchyma [23]. The action of oestrogen and progesterone on mammary parenchyma leads to the secretion of EGF and transforming growth factor-a (TGF-a), both of which are potent mammary mitogens [22]. In contrast, TGF-b1 inhibits mammary growth [22, 24, 25]. These studies imply that the development of the human mammary gland during pregnancy is controlled by a complex sequence of stimuli and inhibition similar to the hormonal mechanisms that control the growth and development of the mammary glands of common laboratory animals.

Hytten [26] used a water-displacement technique to measure the volume of the left breast of 11 women at three months of gestation and seven days postpartum. In 10 women breast volume increased by 60 to 480 ml, and in one woman the volume decreased by 20 ml. The relationship, however, between breast volume at the end of pregnancy and the volume of the "empty" breast on the seventh postpartum day is unknown. We [27] have developed a computerized breast measurement (CBM) system that uses video images of structured light stripes projected onto the breast to quantitate both long-term (throughout the lactation cycle) and short-term (during the day) changes in breast volume (figs. 1 and 2). Preliminary findings for breast development (fig. 3) from preconception until just before delivery [28] have demonstrated that significant growth can occur during the first trimester and that this growth can either continue throughout pregnancy or reach a plateau during the second trimester. The six mothers studied to date have had a successful lactation outcome irrespective of their pattern of breast growth during pregnancy.

Lactogenesis 1

The timing of lactogenesis 1 (development of potentially functional lactocytes) in women has not been precisely defined. In mid-pregnancy the true lobulo-alveolar system develops, the proliferative changes are reduced, and there is increasing cellular differentiation with the accumulation of cellular organelles and secretory products [6, 15, 29]. Lactose is the most osmotically active component of the colostrum, and hence excess lactose synthesis during pregnancy could lead to breast distension. Nevertheless, the tight junctions between the lactocytes are open during pregnancy, and thus lactose can escape across the mammary epithelium into the bloodstream and then be excreted in the urine. The concentration of lactose in the blood of pregnant women increases during mid-pregnancy [30], and this coincides with an increase in the excretion of lactose in the urine [28, 31]. These findings suggest that in women lactogenesis 1 occurs approximately halfway through pregnancy. Studies using an increase in the concentration of lactose in the blood or mammary secretion as an indicator of the occurrence of lactogenesis 1 have suggested that lactogenesis 1 occurs at different stages of pregnancy in different species. Whereas lactogenesis 1 occurs in late pregnancy in rats [32], ewes [33], and sows [34, 35], it occurs earlier in pregnancy in cows [36], goats [37], and women [30]. These species differences in the timing of lactogenesis 1 during pregnancy make it difficult to speculate on the control of the development of the functional lactocyte. A greater knowledge of the control of mammogenesis and lactogenesis 1 in women is important not only for the understanding of normal lactation, but also because early pregnancy [13] and lactation [38] have been associated with a reduced risk of breast cancer.

Lactogenesis 2

Whereas the umbilical cord couples the developing foetus to continuous life support from its mother's placenta, nursing after birth provides the growing infant with comparable but intermittent life support from its mother's breasts. Thus, it is essential for the breast to develop its unique synthetic capacity during pregnancy, so that the initiation of an adequate supply of milk accompanies the birth of the infant. The occurrence of lactogenesis 1 halfway through pregnancy in women permits lactogenesis 2 to occur, even if the infant is delivered prematurely, although the milk composition is different from that of full term mothers, possibly because of incomplete hypertrophy of the mammary gland or incomplete exposure to prolactin and other hormones [39].

FIG. 2. Computerized breast measurement system assessment of function for right and left breasts of two mothers (A and B) over 24 hour periods. The degree of fullness is the volume of the breast at the end of a breastfeeding minus the minimum volume of the breast during the 24-hour period, divided by the storage capacity of the breast, expressed as a percentage. The rate of milk synthesis is calculated from the difference between the breast volume at the end of a breastfeeding and the breast volume just before the beginning of the next breastfeeding, divided by the time between the measurements [60]

In women the control of lactogenesis 2 (the initiation of copious milk secretion) appears to be under endocrine regulation similar to that in other mammals. The pioneering work of Kuhn [40] established that progesterone withdrawal was the trigger for lactogenesis 2 in the rat, and subsequently Nicholas and Hartmann [32] established the temporal relationship between the withdrawal of progesterone approximately 24 hours before parturition and the increased rate of lactose synthesis (a measure of lactogenesis 2) coinciding with birth. This pattern appears to be consistent for most mammals studied to date.

In women, however, progesterone withdrawal is delayed until after the delivery of the placenta; thus, there is a frame shift to the post-partum period in the dose temporal relationship between the fall in progesterone and the increase in the concentration of lactose in the colostrum [41].

Lactogenesis 2 is delayed by more than a day to between 30 and 40 hours post-partum (fig. 4). Although this delay may seem inconsistent with the perceived energy requirements of the human infant, it is consistent with the acquisition of mucous membrane protection from maternal colostrum, as well as the remarkable resilience of the human neonate to nutritional abuse and the unusual weight loss m human infants after birth.

FIG. 3. The development of the left breast of one mother depicted by models generated using the computerized breast measurement system

Lactogenesis 2 also occurs 30 to 40 hours after delivery in mothers who have had a Caesarean section [42], since the timing of placenta removal in relation to birth is the same as in those who delivered normally. This is fortunate for human lactation, as Caesarean section delivery in other species, such as ewes, results in a delay in lactogenesis 2 of more than a day when compared with normal delivery [33].

The involvement of the placenta in the initiation of lactation has been further substantiated by the finding that if a fragment of placenta is retained after delivery, lactogenesis 2 occurs only after its removal [43]. These observations also are consistent with progesterone withdrawal acting as the trigger for lactogenesis 2 in women, as it does in other mammals.

Although lactogenesis 2 in women does not require either the suckling stimulus or milk removal [44], it does require the presence of adequate concentrations of lactogenic hormones. The concentration of prolactin in a woman's blood is high at parturition, and suppression of prolactin secretion by the administration of bromocriptine results in the inhibition of lactogenesis 2 [44]. Nevertheless, undesirable side-effects and the potential role of prolactin in facilitating maternal behaviour have cautioned against the use of this drug for the suppression of lactation in women who choose not to breastfeed.

Close assessment of the initiation of lactation in mothers with type I diabetes shows that the increase in the concentration of lactose in the colostrum is delayed by about 24 hours compared with normal mothers [45-47]. Since this delay coincides with the reduction in insulin therapy after birth and the reestablishment of control of glucose homeostasis, permissive amounts of insulin also may be required for lactogenesis 2. Notwithstanding this delay, diabetic women with perseverance can establish successful lactation.

Milk "coming in," which is sensed by mothers as a sudden enlargement of their breasts with milk, is a post-lactogenesis 2 event occurring two to three days after delivery [42]. The physiological significance of the sensation of milk coming in is not clear, but it may represent the time when the mother's capacity to synthesize milk first exceeds the infant's appetite. That is, it may represent the transition from an endocrine-promoted lactogenesis 2 to another form of control during established lactation. Despite these considerations, severe engorgement can be associated with milk coming in, and this is a very painful event of short duration but long remembered by mothers. The risk of severe engorgement is reduced if mothers avoid timed schedules for breasfeeding and fully breastfeed their babies to appetite, day and night, from birth. The implications for lactogenesis 2 and milk coming in are important when considering the design of programmes for the early discharge of mothers from maternity hospitals [48].

FIG. 4. Peripartum milk lactose expressed as a percentage of the maximum lactose concentration for the cow, rat, and woman


The maintenance of galactopoiesis appears to be under a combination of endocrine, autocrine (local), and metabolic control, which varies according to the species and the stage of lactation. The relative importance of these mechanisms, however, depends on whether the species has been selected for dairy production. Nevertheless, the removal of milk from the mammary gland is of the utmost importance for the maintenance of milk secretion in all mammals. In women, frequent suckling not only stimulates the release of oxytocin and thereby elicits milk ejection, but also stimulates the release of prolactin [49] and results in the removal, inactivation, or both of local inhibitory factors [50].

The finding in the early 1970s that the sucking stimulus evoked the release of prolactin [49] provided a potential explanation for the regulation of milk synthesis. We have not, however, found any relation between the suckling-evoked release of prolactin and milk yield. Although the prolactin response decreases in later stages of lactation, there is no decrease in milk yield [28]. Therefore, the concept that demand feeding regulates the rate of milk synthesis by evoking the release of prolactin must be seriously questioned.

About 20 years ago, it was generally accepted that the nutrition of the mother was the most important determinant of milk synthesis. This assumption was based on numerous studies demonstrating the dependence of milk production on the level of nutrition in dairy cows and goats. As a result, it was concluded that feeding the malnourished mother would thereby feed the child [51]. Nevertheless, subsequent studies by Prentice et al. [52] have shown that improving the nutritional status of malnourished mothers does not increase milk production. This, together with extensive studies on milk production in women from both developing and developed countries, has led to the conclusion that maternal nutrition is not an important determinant of milk production in women (see ref. 53).

The demand-fed infant consumes irregular quantities of milk at irregular intervals during the day [53]. These feeding patterns also are consistent with mothers' awareness [54] and with recent studies that indicate that it is the baby's appetite that determines milk yield [53]. Indeed, a comprehensive consideration of the literature leads to the conclusion that ensuring that the infant has sufficient access to the breast to satisfy its appetite for milk optimizes milk production. The basic question, however, remains: How does the breast regulate milk synthesis to meet this unpredictable external demand for milk by the infant?

In 1984 Neil Matterson [55] provided a practical answer to this question: "The more he cries the less milk he drinks, so less milk is produced, so there's less for him to drink, so he cries because he didn't get a drink. Do you understand that?" Perhaps this confusing explanation of the control of milk synthesis underlies epidemiologic results showing that many mothers give up breastfeeding in the belief that they cannot produce enough milk for their babies [56].

A prerequisite for investigating the control of milk synthesis during galactopoiesis in any suckling mammal is to measure the rate of milk synthesis accurately. The classical means of determining the daily rate of milk synthesis in women has been the "weigh-suckle-weigh" or "test-weighing" method [57], which conventionally records the combined milk output from both breasts and sums the milk consumed during all breastfeedings over a period of 24 hours. If correction is made for evaporative water loss during each breastfeeding, test weighing is a very accurate method of determining milk transfer between the mother and her infant [58]. But neither test weighing nor alternative methods of measuring milk production [59] measure the short-term (between breastfeedings) rates of milk synthesis in individual breasts.

Our CBM system measures the short-term rates of milk synthesis in women by determining the rate of increase in breast volume between breastfeedings [27]. We installed the CBM system in the homes of seven mothers so that we could measure their milk production and changes in breast volume while they breastfed on demand in their normal environment. In addition to the short-term rate of milk synthesis, we were able to calculate two new parameters for the assessment of breast function from the progressive changes in breast volume over the 24-hour period. First, the storage capacity of the breast was calculated, that is, the demonstrated capacity of the breast to store milk that was available to the infant. This was calculated as maximum breast volume minus minimum breast volume observed over the 24-hour period [60]. Second, the degree of fullness of the breast was calculated as the volume of the breast at the end of a breastfeeding minus the minimum volume of the breast during the 24-hour period, divided by the storage capacity of the breast. Thus, the degree of fullness varied from one when the breast was full to zero when the breast was empty.

To illustrate our findings using the CBM system, the results for two of the seven mothers who participated in this study [60] are presented (table 2 and fig. 2). Mothers A and B were fully breastfeeding their babies and had similar levels of milk production: 912 and 950 g/24 hours, respectively. Nevertheless, as with most other women we have studied, this production was not divided equally between the mother's breasts, emphasizing the importance of investigating the regulation of milk synthesis in individual breasts rather than in individual women.

TABLE 2. Computerized breast measurement system assessment of breast function for two mothers

Measurement Mother A Mother B
Stage of lactation (mo) 4 5
Milk yield (g/24 h) 912 950
Storage capacity (ml) 780 190
Relative breast size (%)
minimum 69 92
maximum 71 96
Individual breasts
Measurement Left Right Left Right
Milk yield (g/24 h) 108 804 338 612
Storage capacity (ml) 180 600 80 110
Feedings/24 h 3 4 7 7
Rate of synthesis (ml/h)
minimum 3 17 6 23
maximum 15 49 18 31

Source: ref. 60.

A detailed review of the results of our investigations of human lactation over the 24-hour period using the CBM system is given by Daly and Hartmann [61], and a summary of our observations and conclusions is given in table 3. These studies clearly demonstrated that the infant's appetite determined the milk intake at a particular breastfeeding. The breast storage capacity varied greatly between women (table 2), and this factor significantly influenced the breastfeeding frequency required to maintain an adequate milk supply. In addition to these observations, we were unable to show a relation between the increase in blood prolactin at a breastfeeding and the subsequent rate of milk synthesis between breastfeedings [62]. Furthermore, the short-term rates of milk synthesis of a woman's right and left breasts often responded independently from breastfeeding to breastfeeding. For example, after the first feeding, the milk synthesis rate of the right breast could be higher than that of the left breast, but after the next feeding, the reverse could be true, with the left breast now showing the higher rate of milk synthesis.

The negative relationship between the degree of breast fullness and the short-term rate of milk synthesis (fig. 2) was of particular interest. In this respect, the right breasts of mothers A and B represent two extremes. Mother A, with a large storage capacity, had high rates of milk synthesis when her breast contained the least milk (fullness approximately zero) and low rates of milk synthesis when her breast was nearly full (fullness approximately one). In contrast, mother B. with a small breast storage capacity, achieved a high level of milk production by more frequent breastfeedings with relatively consistent short-term rates of milk synthesis, as her degree of fullness was low after each breastfeeding.

TABLE 3. Observations and conclusions from computerized breast measurements of the circadian changes in breast volume in lactating women

Observation Conclusion
The infant did not consume all available milk at a breastfeeding The infant's appetite determines milk intake at a breastfeeding
Mothers with small breast storage capacity breastfed more frequently The storage capacity of the mother's breasts dictates the flexibility in frequency of breast feeding
Rates of milk synthesis varied greatly within and between breast-feedings The short-term rate of milk synthesis is under autocrine (local) control
Rate of milk synthesis was highest when the breast contained the least milk Autocrine control responds to the fullness of the breast
Right and left breasts differed in both milk production and size Autocrine mechanisms may affect the number of lactocytes in the breast

Our results for women are consistent with the autocrine theory of milk synthesis control during established lactation, which has been recently proposed for dairy goats by Peaker and Wilde [50]. These workers have isolated a protein called feedback inhibitor of lactation (FIL) that appears to suppress milk synthesis as milk accumulates in the mammary gland between breastfeedings by reversibly inhibiting the transfer of newly synthesized protein from the endoplasmic reticulum to the Golgi vesicles.

We also have shown that the difference in the fat content of breastmilk between the beginning and the end of a breastfeeding is related to the degree of fullness of the breast, rather than to whether it is either fore or hind milk [63]. Furthermore, Heesom et al. [64] have demonstrated that medium-chain-length fatty acids inhibit glucose metabolism and lipid synthesis in isolated mammary acini of rats. The presence of lipases in breastmilk and the nature of the accumulation of triacylglycerol in the alveolus suggest that autocrine control of milk fat synthesis by free fatty acids also is mechanistically plausible in lactating women [65]. In vitro studies have shown that free fatty acids may act as messenger and modulator molecules, mediating responses of the cell to extracellular signals [66].

A further interesting development arises from studies by Molenaar et al. [67]. They used 35S-labelled cRNA probes to localize the sites of a-lactalbumin, a-S1-casein, and lactoferrin mRNA synthesis in sheep. Early in lactation, mammary gland expression of a-lactalbumin and a-S1-casein was high in some alveoli but not in others. Those alveoli with high expression of a-lactalbumin and a-S1-casein contained few fat globules in their cells and lumina, whereas those in which expression of these proteins was absent contained abundant fat globules. These latter alveoli also almost exclusively expressed lactoferrin. These findings suggest that milk secretion either is heterogeneous across lobules or occurs sequentially with time in the alveolus as newly secreted milk accumulates. The latter concept seems more plausible.

It is tempting to speculate that the FIL and free fatty acids may act locally to sequentially regulate the short-term rates of milk constituent synthesis, either within a mammary gland or, more likely, within lobules according to the degree of emptying of the lobules in each breast.


Studies of lactation in village women and the few remaining hunter-gatherer societies suggest that the normal duration of lactation in women is three to four years. In these societies, and more recently in a number of developed countries, weaning is prolonged and gradual (that is, child-led weaning). The cessation of sucking results in distension of the gland with milk and atrophy of the epithelial structures. Finally, milk secretion is greatly suppressed, and the lactocytes disintegrate and desquamate. Phagocytosis of the degenerated alveoli reduces the lobulo-alveolar structures, and ductal systems become predominant. It seems that the involution of the human breast occurs much more gradually than that of the mammary glands of other mammals [68], and this may explain the relative ease of relactation in women.


We would like to thank the mothers who kindly volunteered for our studies and the Nursing Mothers' Association of Australia. The work from our laboratory was supported by the National Health and Medical Research Council of Australia.


Discussion of paper by Hartmann et al.

Discussion of paper by Hartmann et al.

Dr. McLaren

The two women you showed us both produced more milk from the right breast than the left. Is this usual, and is it related to which breast the baby is put to first at a feeding?

Dr. Hartmann

I think that perhaps two-thirds of women are uneven in the milk production between their left and right breasts, but it doesn't necessarily mean they are consistently higher in the right breast. We found that some women changed from one lactation to the next, so most milk production in one lactation is in the right breast, and in the next it's in the left breast, so there is a fairly random distribution between breasts. Some people think it has to do with the handedness of mothers. We haven't looked at that, but certainly it can change during subsequent lactations. It seems to me to relate to a behavioural characteristic in animals, which is teat order. There is evidence from Sweden, I believe, that indicates that when babies are put on their mother's abdomen, they make their way up to the breast. They also tend to associate with the breast that they attach to first. This observation would be consistent with observations of the establishment of teat order in other mammals.

Dr. McLaren

I was particularly interested in your observation that fat content increases as the breast is emptied. Does that mean it is better for a woman to empty one breast before transferring the baby to the other?

Dr. Hartmann

This is an interesting situation, and one has to be very careful in responding, because how it is applied in practice can be quite misleading. The breast has regular mechanisms that can cope with most patterns of feeding. So far, our results indicate that it doesn't much matter how the breast is emptied; the baby will get the same amount of energy over a 24-hour period. One-sided feeding has only been recommended for women with oversupply problems. For normal feeding it is reasonable to finish one breast before offering the other.

Dr. McNeilly

Back to milk production. Presumably, in the case of twins, both breasts can operate to full capacity if they are emptied at the same rate?

Dr. Hartmann

That is correct. Our studies suggest there is additional capacity. I think mothers in Western societies have to down-regulate their physiological capacity to produce milk to their babies' requirements. We can look at the endocrine system as setting the upper limits of the mother's milk production capacity and the autocrine mechanism as then cutting it down to the actual needs of the baby. With twins the mother sustains a higher level of production that perhaps approximates the breasts' maximum capacity. We studied a woman who fully breastfed triplets for two and a half months, so the breast has enormous capacity.

Dr. McNeilly

You didn't mention the relationship of oxytocin to storage capacity. It seems that much of the milk won't come out without oxytocin release. Is that true? There is very little information on that.

Dr. Hartmann

Residual milk, in my opinion, has done more to confuse our understanding of lactation than just about any other concept. This is because it is based on the very unnatural situation of the dairy cow. The suckling situation is totally different. Oxytocin is obviously required for milk ejection, but the amount of milk that is taken depends on the baby's appetite, not what the mother has available. So in mothers with high storage capacity, there is enormous variation between the storage capacity and the amount of milk that the baby might take at any one time.

Dr. McNeilly

But if the mother didn't release oxytocin at all, would that dramatically reduce the amount of milk the baby could get?

Dr. Hartmann

Yes, it certainly can be a problem. We don't notice it particularly in women who are breastfeeding. It is often a problem for mothers who are expressing milk. If they don't get a let-down, they can have a great decrease in milk production.

It is very important to look at milk production from individual breasts, which relates to Dr. McLaren's first question. We did an initial study in Oxford on six mothers to determine if the interval between feedings was correlated with the amount of milk the baby took. If you combine the production of both breasts, there's no relationship. If you split the intake from the left and the right breasts, there is a very tight coupling. That is, the gap between feeds determines how much the baby takes from a breast. For the other breast, there is no such relationship. It was clear women showed "breastedness" in the same way they show handedness, but it wasn't linked to handedness, because we had five "left breasted" mothers and one "right-breasted" mother, but they were all right-handed.

So it's not clear what the relationship is. It was almost as if the control mechanisms were located in one side, as in true handedness. When we moved to Bristol, we studied 20 more mothers and found no such relationship. We found the relationship was with the first breast to be offered, which confused us somewhat. Therefore, we didn't publish the results.

Constituents of human milk

Constituents of human milk

Ann Prentice


Breastmilk contains all the nutrients required by the newborn baby. It also contains non-nutritional components that may promote infant health, growth, and development, such as antimicrobial factors, digestive enzymes, hormones, trophic factors, and growth modulators. In some situations, breastmilk may also contain harmful components, such as pollutants, drugs, allergens, and viruses. Human milk has a unique composition, which differs from that of other mammals in its ingredients and their concentrations.

Breastmilk composition is not constant and varies with stage of lactation, breastfeeding pattern, season, and parity. It also differs among individuals and among communities, for reasons that are not well understood. Maternal nutrition is an obvious candidate, but the situation is complex. Supplementation studies and cross-cultural comparisons have demonstrated that the total concentrations of fat, protein, and lactose are relatively insensitive to current dietary intake and nutritional status, whereas the fatty acid profile and the concentrations of several micronutrients, particularly water-soluble vitamins, are responsive to maternal diet.

For many infants, nutrient intake from breastmilk becomes insufficient after four to six months, and other foods should be added to the diet. Nevertheless, breastmilk can continue to be a valuable nutrient source and to provide non-nutritional factors even for older children. Consequently, breastfeeding for one to two years as part of a mixed diet has many advantages, particularly for children living in impoverished circumstances.


Breastmilk is a complex fluid, rich in nutrients and in non-nutritional bioactive components. Knowledge of the composition of human milk and the factors that influence it has increased considerably over the past two decades. This paper briefly summarizes current understanding, drawing largely on our cross-cultural studies in the Gambia and Britain as examples. More comprehensive reviews have recently been published [1-4].

Nutritional and non-nutritional components

Breastmilk contains all of the nutrients needed by the newborn baby during the first weeks of life. These include the metabolic fuels (fat, protein, carbohydrate), water, and the raw materials for tissue growth and development, such as fatty acids, amino acids, minerals, vitamins, and trace elements.

More than 98% of the fat in breastmilk is in the form of triglycerides, constructed within the mammary epithelial cell from medium- and long-chain fatty acids derived either from the maternal circulation (carbon chain lengths £ 16) or manufactured locally (carbon chain lengths ³ 16) [5]. Short-chain fatty acids (carbon chain length £ 8) are only present in trace amounts. Oleic acid (18:1) and palmitic acid (16:0) are the most abundant fatty acids in breastmilk triglycerides, with comparatively high proportions of the essential fatty acids, linoleic acid (18: 2w 6) and linolenic acid (18: 3w 3). Comparatively high proportions of other long-chain polyunsaturated fatty acids, such as arachidonic acid (20: 4w 6) and docosahexaenoic acid (22: 6w 3), are also present [5]. These long-chain fatty acids are constituents of brain and neural tissue and are needed in early life for mental and visual development [6]. At least half of the triglyceride molecules in breastmilk contain palmitic acid attached to the central carbon of the glycerol component, a property that increases digestibility, absorption, and mineral balance [5, 7]. The lipid component of breastmilk is the transport vehicle for fat-soluble micronutrients such as prostaglandins and vitamins A, D, E, and K.

Proteins account for approximately 75 % of the nitrogen-containing compounds in breastmilk. Non-protein nitrogen substances include urea, nucleotides, peptides, free amino acids, and DNA. The proteins of breastmilk can be divided into two categories: micellar caseins and aqueous whey proteins, present in the ratio of about 40:60 [8]. The predominant casein of human milk is b-casein, which forms micelles of relatively small volume and produces a soft, flocculent curd in the infant's stomach. The major whey proteins are a-lactalbumin, lactoferrin, secretory IgA, and serum albumin [8], with a large number of other proteins present in smaller amounts. Secretory IgA is the principal immunoglobulin of breastmilk. It is synthesized in the mammary epithelial cell by the coupling of two IgA molecules, produced locally by lymphocytes resident in the breast tissue, with two proteins, J-chain and secretory component [8]. The specificity of breastmilk secretory IgA antibodies reflects the mother's exposure to mucosal infection and is independent of the specificity profile of blood-borne IgA [9]. Many of the proteins in breastmilk have a multitude of potential functions. Lactoferrin, for example, transports and promotes the absorption of iron, is bacteriostatic to a range of organisms, and acts as a nutritional protein, producing amino acids for absorption on digestion [8, 10].

The principal carbohydrate of human milk is lactose, a b-disaccharide manufactured in the mammary epithelial cell from glucose by a reaction involving a-lactalbumin [11]. In addition, breastmilk contains significant quantities of oligosaccharides, predominantly lactose-N-tetraose and its monofucosylated derivatives, representing about 10% of total milk carbohydrate. The oligosaccharide composition reflects the Lewis blood group and secretor status of the mother [12].

In addition to the nutritional components, breastmilk contains a wealth of bioactive components that may have beneficial non-nutritional functions [8, 9, 13, 14]. These include a wide range of specific and non-specific antimicrobial factors; cytokines and anti-inflammatory substances; and hormones, growth modulators, and digestive enzymes (table 1), many of which have multiple activities. These components may be of particular importance for young infants because of the immaturity of the host defence and digestive systems early in life. The physiological significance of many of these substances has yet to be determined, and some may be present merely as "spillover" or excretory products from metabolic processes occurring within the mammary epithelial cell. For those with established significance, the site of action may be within the mother's breast, within the infant's alimentary canal, or, after absorption, within the infant's body. Some antimicrobial components, for example, are active both within the breast, minimizing the risk of breast infection and mastitis [15], and within the baby's gastrointestinal and respiratory tracts, protecting the mucosal surfaces from infection by bacteria, viruses, and parasites [9]. By contrast, the site of action of the peptide feedback inhibitor of lacation (FIL) is within the breast, its function being the autocrine regulation of milk production [16]. On the other hand, casomorphins, opioid-like substances that may affect infant behaviour and mood in addition to a range of other functions, are produced in the baby's intestines by the degradation of breastmilk casein [17]. Many bioactive substances are also valuable nutrient sources and ultimately are digested and absorbed in the normal way. Protease inhibitors in breastmilk may afford a degree of protection from digestion for some breastmilk components [8]. A sufficient proportion of antimicrobial proteins, for example, escape digestion and emerge in the faeces, suggesting that antimicrobial activity continues throughout the length of the infant's gastrointestinal tract [10].

Breastmilk has also been shown to be an excretory route for a range of substances that might be harmful to the baby (table 1). These include viruses, such as human immunodeficiency virus (HIV) [18]; environmental and occupational pollutants, such as DDT, PCBs, and dioxins [19]; components of the mother's diet that might be toxic or allergenic, such as trans-fatty acids, aflatoxins, and cow's milk protein [2022]; commonly used stimulants, such as nicotine, caffeine, and theobromine [23, 24]; and various drugs and radioactive compounds [25-27]. Where exposure to xenobiotics jeopardizes infant health, difficult and often controversial decisions have to be made about whether the risks outweigh the benefits of breastfeeding [26, 28].

Comparison with animal milks

Breastmilk contains a unique combination of ingredients, differing from the milks of other mammals in both the concentration and the nature of its many components. In common with the milk of other primates, human milk has low energy and nutrient density compared with the milks of most other mammals, except for a high density of carbohydrates [4]. In addition, the daily output of the major nutrients in milk relative to the size of the mother is lower in humans than in other mammals, especially dairy and laboratory species [29].

The composition of cow's milk, the basis of most breastmilk substitutes over the centuries, is compared with that of human milk in table 2. In addition to the obvious concentration differences, the milks differ considerably in the structure of many of the milk fractions [30]. For example, in cow's milk the major proteins are a-casein and , b-lactoglobulin; the ratio of casein to whey protein is 80:20; the casein micellar volume is double that of human milk, and the curd formed is hard; the principal milk immunoglobulin is IgG; and lactoferrin and Iysozyme are present only in small amounts [30]. Cow's milk triglycerides contain a higher proportion of short chain fatty acids and a lower proportion of long chain and polyunsaturated fatty acids; furthermore, the positional distribution of fatty acids on the glycerol molecule is different [30]. In addition, many of the non-nutritional factors found in human milk are absent from cow's milk or are present only in trace amounts. For the human baby, these differences affect the digestibility and absorption of nutrients, the bioavailability of micronutrients, and the potential benefits from non-nutritional factors.

TABLE 1. Examples of the non-nutritional components of breastmilk

Antimicrobial factors Growth factors
secretory IgA, IgM, IgG
complement C3
bifidus factor
lipids and fatty acids
antiviral mucins, GAGs
epidermal (EGF)
nerve (NGF)
insulin-like (IGF)
transforming (TGF)
Cytokines and anti-inflammatory factors Digestive enzymes
tumour necrosis factor
platelet-activating factor: acetyl hydrolase
bile acid-stimulating esterase
bile acid-stimulating lipases
lipoprotein lipase
Hormones Transporters
feedback inhibitor of lactation (FIL)
thyroid hormones
corticosteroids, ACTH
parathyroid hormone
lactoferrin (Fe)
folate binder
cobalamin binder
IgF binder
thyroxine binder
corticosteroid binder
Potentially harmful substances Others
viruses (e.g., HIV)
trans-fatty acids
nicotine, caffeine
food allergens
PCBs, DDT, dioxins
d -sleep peptides

Data compiled from references given in the text.

Variations in breastmilk composition

The composition of breastmilk is not uniform, and the concentrations of many of its constituents change during the lactation period and differ between individual mothers. As variations in concentration are not necessarily inversely related to breastmilk volume, differences in breastmilk composition affect the daily intakes of milk components by the breastfed child. There are several factors that are known to influence the concentration of breastmilk constituents in predictable ways [31]. These include stage of lactation; breastfeeding routine; parity, age, and other maternal characteristics; regional differences; and, in some situations, season of the year and maternal diet. These are discussed in detail below.

Stage of lactation

Human lactation can be divided into four phases that differ in the composition and volume of milk produced: colostral, transitional, mature, and involutional. Colostrum is secreted for the first three to five days after delivery, transitional milk until the end of the second week, mature milk during full lactation, and involutional milk at the end of lactation. These definitions are arbitrary; the timing varies from one mother to another, and composition does not change abruptly. Typical concentrations of selected milk constituents are shown in table 3 [29, 32-35]. Notably, colostrum is rich in secretory IgA, lactoferrin, vitamin A, and sodium compared with mature milk but has relatively low concentrations of fat, lactose, and vitamin B1. Involutional milk is characterized by low lactose content and high concentrations of protein, fat, and sodium [11, 36]. Because milk volume is low during the colostral phase, rising slowly during the first week to the higher levels of established lactation [37-39], the daily intake of most milk components by breastfed babies increases after birth, reaching a peak after several weeks (table 4). The exception is secretory IgA and, hence, total protein intake, which is maximal in the first week (table 4).

Mature breastmilk composition also changes during the course of lactation, although not as markedly as in the early weeks [31, 36, 40]. Many nutrients show a gradual decrease in concentration of around 10% to 30% during the first year of lactation, often reaching a low plateau thereafter. A greater decrease occurs for some components, such as zinc [41]. Some components show little change, especially those involved in osmoregulation, including lactose and sodium, whereas a few, notably Iysozyme, increase.

Breastfeeding routine

Breastmilk composition can vary during the day and from the beginning to the end of a feeding. This is most pronounced for fat and fat-soluble components such as vitamin A and zinc [5, 42]. The fat content of breastmilk can change by as much as fivefold during the course of a feeding [42]. The fat concentration is influenced by the breastfeeding routine of the mother, and short-term variations are related to the volume of milk produced per feeding and the time interval between feedings [42]. Differences in breastfeeding routine can affect the diurnal variation in fat concentration. In the Gambia, for example, where mothers feed on demand and sleep with their infants who suckle during the night, the highest breastmilk fat concentration tends to occur in the early morning, whereas in Western societies, with different feeding schedules, the early morning is associated with the lowest fat concentration of the day [42]. Other constituents, such as protein, may show small but consistent changes from the beginning to the end of a feeding and during the day [36, 43], whereas others, such as calcium, are unaffected [44].

TABLE 2. Typical contents of human and cow's milk

Contents Human milk Cow's milk
total (g/100 ml) 4.2 3.8
fatty acids £ 8C (% ) trace 6
polyunsaturated fatty acids (%) 14 3
Protein (g/100 ml)
total 1.1 3.3
casein 0.4 0.3 2.5
a-lactalbumin 0.3 0.1
lactoferrin 0.2 trace
IgA 0.1 0.003
IgG 0.001 0.06
lysozyme 0.05 trace
serum albumin 0.05 0.03
ß-lactoglobulin - 0.3
Carbohydrate (g/100 ml)
lactose 7.0 4.8
oligosaccharides 0.5 0.005
Minerals (g/100 ml)
calcium 0.030 0.125
phosphorus 0.014 0.093
sodium 0.015 0.047
potassium 0.055 0.155
chlorine 0.043 0.103

Data compiled from references given in the text.

TABLE 3. Typical contents of 100 ml breastmilk produced on days 1-5 (colostrum) and more than 15 days (mature milk) post-partum

Contents Colostrum Mature milk C: M%a
energy (kcal) 55 67 82
Fat (g) 2.9 4.2 69
Lactose (g) 5.3 7.0 76
Protein (g)
total 2.0 1.1 182b
secretory IgA 0.5b 0.1 500b
lactoferrin 0.5 0.2 250
casein 0.5 0.4 125
Calcium (mg) 28 30 93
Sodium (mg) 48 15 320
Vitamin A
(µg retinol equivalents) 151 75 201
Vitamin B1 (µg) 2 14 14
Vitamin B2 (µg) 30 40 75
Vitamin C (µg) 6 5 120

Data compiled from references given in the text.

a. Percentage ratio of concentrations in colostrum and mature milk.
b. Considerably higher on days 1-3.

Maternal parity and age

Breastmilk composition may be influenced by the parity and age of the mother. In the Gambia young, primiparous mothers have higher concentrations of several constituents, especially fat, total protein, and immunoproteins, whereas older mothers of very high parity (nine or more children) tend to produce milk with reduced quality [45-48]. Other components, such as calcium, do not change with parity [44]. Similar observations have been made in some studies elsewhere [36], but not in others, and the mechanisms involved remain elusive.


The season can influence breastmilk composition. In subSahelian Africa, where food availability, infection rates, farm work, and child-care patterns vary between seasons, variations in the concentrations of some constituents, such as fat, immunoproteins, and watersoluble vitamins, have been observed [45, 49, 50]. The changes may be related, in pan, to alterations in the mother's diet or breastfeeding behaviour. Milk ascorbate level, for example, closely parallels maternal plasma ascorbate concentration and vitamin C intake, and is high during the season when mangoes are plentiful but low for the rest of the year [50].


Differences in breastmilk composition have been reported between urban and rural populations, and between different socio-economic, geographic, and ethnic groups. Failure to consider differences in duration of lactation, breastfeeding practices, maternal parity and age, sampling protocols, and assay techniques may have contributed to the impression of marked regional differences. A recent evaluation of the available data suggests that the similarities between regions are more striking than the differences, particularly with respect to the major nutrients [31]. Nevertheless, some distinct regional differences are evident, particularly in the concentrations of certain protein components, minerals, vitamins, and trace elements [31, 44, 51]. The reasons are largely unknown but may be related, in part, to the maternal diet and the local environment.

Maternal diet and nutritional status

In the past it was commonly believed that poorly nourished mothers had reduced lactational performance, in both the amount and the quality of breastmilk produced. This view has now been shown to be largely incorrect [4]. A recent examination of the world literature could not demonstrate any convincing relationships between maternal nutritional status, as indicated by body mass index (BMI), defined as weight/height2, and either breastmilk output or energy content [52], even in very thin mothers (BMI < 18.5 kg/m2).

TABLE 4. Typical daily intakes (g) of selected breastmilk constituents

Age Protein Fat Lactose Calcium
Total Secretory IgA Other
Day 1 5.0 4.8 0.2 <1 1 0.01
Day 3 12.0 5.0 7.0 5 12 0.06
Day 8 9.0 1.8 7.2 22 40 0.19
3 mo 7.5 0.7 6.8 29 52 0.22

Calculated intakes assume the following daily milk volumes: day 1 (0-24 h), 40 ml; day 3 (48-72 h), 200 ml; day 8, 600 ml; 3 months, 750 ml [1, 38, 52]. concentration data from table 3 and other references in the text.

Direct dietary supplementation studies mostly support this view. In a Gambian study, where poorly nourished lactating mothers were given a high-energy, nutritionally balanced supplement that provided a net energy gain of 3 MJ/day, there was no impact on breastmilk volume [4, 37]. Breastmilk fat and protein concentrations were increased slightly by the supplement, but lactose levels fell, resulting in only a marginal effect on total breastmilk energy. A review of other intervention studies concluded that there was no persuasive evidence for the positive effects of diet on breastmilk energy output [53]. Although breastmilk fat concentration has been correlated with various aspects of maternal fatness in a number of studies [4, 36], including those in the Gambia [45], this observation is not universal; in some populations negative relationships have been reported [4].

Lactation, therefore, appears to be relatively robust in the face of poor nutrition. Maternal diet can, however, affect the breastmilk concentrations of many minor constituents, particularly long-chain polyunsaturated fatty acids, some vitamins, zinc, selenium, iodine, and fluorine [51]. The profile of fatty acids in the mother's diet and adipose tissue stores is reflected in the fatty acids of breastmilk [5, 47]. The concentrations of two water-soluble vitamins, riboflavin (vitamin B2) and ascorbic acid (vitamin C), show rapid, dose-related responses to maternal supplementation [4, 50]. The fat-soluble vitamins A, D, E, and K are less responsive to diet because of the buffering action of maternal stores and carrier proteins, but large supplements can result in increased breastmilk concentrations, occasionally to potentially toxic levels [51]. Maternal zinc supplementation may slow the decline in breastmilk zinc concentration during lactation, although the magnitude of this effect and its significance for the breastfed child are still uncertain [41, 54].

Worldwide variations in breastmilk composition have suggested that poor maternal calcium intake may be a factor in determining breastmilk calcium concentration [55]. Mothers in the Gambia, for example, where the diet contains little calcium, have an average breastmilk calcium concentration more than 20% lower than that of British women [44, 56]. A recent supplementation study, however, which tripled the calcium intake of Gambian women for the first year of lactation, had no impact on breastmilk calcium concentration [56]. Intriguingly, a possible link with calcium intake during the preceding pregnancy emerged during this study, an observation that deserves further investigation [57].

Differences between mothers

Even when these various influences are taken into consideration, breastmilk composition varies considerably from one mother to another in the same population. Typically, major constituents such as fat, protein, and calcium can differ by two- to threefold between mothers at the same stage of lactation [44], and the concentrations and activities of some of the minor constituents can be highly variable [48]. At the same stage of lactation, breastmilk volume also varies between mothers [37]. In general, between-mother differences in composition and volume are maintained throughout lactation and are not necessarily related to each other [36, 37, 45, 48, 53, 58]. As a consequence, the intakes of breastmilk components are also highly variable, and some infants consistently receive substantially more or less from breastmilk than others of the same age. It is, therefore, important, when assessing the adequacy of breastfeeding for individual children, that measurements be made of both breastmilk intake and breastmilk composition.

Contribution of breastmilk to infant nutrition

Breastfeeding will normally provide all the nutrient requirements of a baby for the first four to six months of life [59]. After this time some nutrients may become limiting, including not only dietary energy, but also essential minerals such as zinc and iron. Current recommendations are that solid foods are unlikely to be necessary before four months, but a mixed diet should be offered by six months [59].

Moreover, breastmilk can continue to make substantial contributions to infant nutrition for many months after the introduction of weaning foods [40, 60]. This is especially true in communities where weaning and adult foods are of poor quality. In the Gambia, for example, older, partially breastfed children obtain more than two-thirds of their vitamin A and fat from breastmilk [60]. When breastfeeding stops, the intake of these nutrients drops dramatically. A similar effect is seen in Dutch children weaned onto macrobiotic diets [61]. The contribution of breastmilk to vitamin A nutrition may be the reason why prolonged breastfeeding protects against xerophthalmia and eye disorders in regions where vitamin A deficiency is common [60]. In addition, significant quantities of many of the non-nutritional factors continue to be ingested by the partially breastfed older child, which may be especially important for children living in impoverished and unsanitary environments [60]. In affluent societies, where high-quality weaning foods and health care are read fly available, the benefits of partial breastfeeding beyond six months are less apparent [60]. For children in poorer societies, however, the many beneficial effects of prolonged breastfeeding on nutrition, health, birth spacing, and the family economy [61)] strongly suggest that breastfeeding for one to two years should be encouraged as part of a mixed diet.


Discussion of paper by Prentice

Discussion of paper by Prentice

Dr. Garza

Dr. Prentice, I also am intrigued by the calcium data. If you look at casein levels, are they comparable between the Gambia and Britain, or is the casein just not saturated? (I know that is not an appropriate term.)

Dr. Prentice

Clemens Kunz of the University of Dortmund in Germany measured the casein levels for us not long ago, and they seem to be very similar in American and Gambian mothers. In human milk, however, the main vehicle for calcium is citrate, not casein, and we are concentrating on the citrate levels first before going for the casein, as we are not expecting to see any major differences there.

Dr. McNeilly

Dr. Prentice, I have learned a lot. Could you tell me if nutrition might affect the rate of milk synthesis, or is that not shiftable?

Dr. Prentice

Dr. Hartmann could probably answer that question better than I. The only information we have is from our Gambian supplementation study, in which both unsupplemented and supplemented women produced the same amount of milk per day, but the supplemented women fed fewer times per day. So we think that it quite possibly has something to do with the mother's nutritional state, not necessarily with substrate supply affecting the storage capacity, as Dr. Hartmann said. Perhaps the unsupplemented women had smaller storage capacity and had to feed much more frequently during the day to produce the same volume. This is speculation, because we have not done any work on it. Have you any comment, Dr. Hartmann?

Dr. Hartmann

Yes, I think that when substrate supply is limited, the babies have to feed very frequently, regardless of the storage capacity, to keep the rate of synthesis at the maximum. Providing the mothers with more energy from the diet somewhat increases their ability to produce milk. Then the baby might not suckle as frequently to get the volume of milk that it requires, so I don't think it is inconsistent.

Dr. Rasmussen

I would like to contribute the results from a supplementation trial in Guatemala that have been presented at meetings but haven't yet been published. We selected women at the end of pregnancy and the beginning of lactation based on mid-calf circumference, not on body-mass index (BMI), because BMI has the properties that you mentioned. In that supplementation trial, we gave mothers either a low-energy or a high-energy supplement; we didn't have a no-treatment group, because we didn't consider it ethical, given the state of the population. The women in the high-energy group produced more milk, and-of particular importance to this meeting-they exclusively breastfed their babies for a longer period. We were not in the field long enough to ask if there was an effect on fertility in these women (in other words, on the length of their amenorrhoea), but I think there were clearly changes in the way the women breastfed as a result of changes in their milk volume. So I think there may be important effects of supplementation that we may not have seen in the trials that we have had today.

Dr. Prentice

I don't want to detract from that at all, but I still feel that we may be talking about two different things: the biological limits on lactation and the whole area of the mother's well-being-the social factors that affect the way she and the baby interact to increase milk volume. It seems from the work we have been doing that most women can produce reasonable volumes of milk, and if they don't, there are other factors that are probably affecting it other than the capacity of the breast in most instances. I wanted to ask you if you had any macronutrient content data from the breastmilk-that was my brief this morning. Does it affect the composition of the milk?

Dr. Rasmussen

The only component we have looked at is energy. There are compensatory changes in energy, such that women who produce more milk have less energy-dense milk The effects on infant growth are expected to be minimal. But our statistical power was not calculated on measuring infant growth.

Dr. Prentice

I think I would have predicted that. In our supplementation studies we found that if we gave women essentially a higherfat diet, there was more fat in the breastmilk, but lactose was reduced. This has been seen in quite a lot of other studies as well. Thus, the mother's breastmilk fat concentrations may well be responsive either to her own diet or to her adipose tissue stores. We have seen it, other people have seen it, but some other studies have not, and so it is still relatively controversial. But whatever happens, total milk energy seems to stay remarkably similar.

Dr. Hartmann

Could I add that we looked at some immunologic factors in the same material, and lactoferrin remained the same in both groups.

Dr. Garza

What was the content of fat in the Gambian diet?

Dr. Prentice

It's about 20% to 25%.

Dr. Garza

In studies done with Salvador Villalpando and Nancy Butte in Mexican women consuming diets of lower fat content, we found much lower levels of fat in human milk that were not fully compensated for by changes in lactose.

Dr. Prentice

That were nor compensated for?

Dr. Garza

That's right. We saw the same directional changes you did, but the reduction in fat was much greater than the increase in lactose, so that in populations with very low levels of fat (in the 10% to 12% range), the breastmilk fat content may be influenced dramatically.

Dr. Prentice

I certainly could believe that fat intakes might be limiting at that very low level. Did you find a difference in the shift between the manufactured fatty acids and the dietary fatty acids?

Dr. Garza

We didn't look at fat composition. Dr. Villalpando, has fat composition been examined since those earlier results were available?

Dr. Villalpando

We are looking at those data now. We see an increase in the medium-chain-length fatty acids, but the milk energy increase was not that impressive.

Dr. Prentice

Mike Crawford long ago suggested the same thing from Tanzanian studies, but there the total fat tended to stay the same but with larger amounts of medium-chain-length fatty acids.

Professor Howie

I apologize that this question is a little lateral to what you have been talking about, but I want to ask about the concept that breastfeeding mothers may not have to increase their own diet in proportion to the amount of milk they are giving the baby. In other words, if a mother is breastfeeding, the total energy cost to the community is less than if she were feeding herself and her baby by artificial means. That concept was discussed quite a bit a year or two ago, and I wonder if you have any comments on that and on its potential importance for nutrient provision.

Dr. Prentice

I think I would still say the same thing. There do appear to be women who are not able to increase their food intake, that is, their energy intake, and there do seem to be relatively simple compensations that can occur: a reduction in activity, for example. We are still not really sure that in lactation there is much in the way of changes in costs like basal metabolic rate so on, although there may well be in pregnancy, but fairly minor and adjustments may be satisfactory to enhance efficiency. These include accessing energy from fat stores; many, but not all, women in that situation will lose some of their fat stores while they are lactating. Certainly, in the Gambia some women put on weight during lactation at the time of year when food in good supply. But they are obviously eating enough is for both weight gain and milk production. So, yes, I think we would still be saying the same, that there are compensations for the energy needs of lactation.

Problems of establishing lactation

Problems of establishing lactation

Michael W. Woolridge


The problems associated with the establishment of lactation can be viewed primarily from the mother's perspective but should also be viewed from that of the caregiver supporting her and the maternity facility providing her with care.

Factors intrinsic to the mother

The first time a mother puts her baby to her breast, she is doing something for which nothing will have adequately prepared her. Childbirth and breastfeeding are comparable in this respect; unfortunately, anticipation of the former may prevent her from giving adequate consideration to the latter, particularly to practical issues. Clinical experience suggests that ensuring that practical issues are correct can be vital to a successful outcome for breastfeeding, yet it rarely proves possible to address them adequately antenatally. The result is that the mother and her newborn, both novices to the process, go through an extremely rapid learning curve that can run completely smoothly or may be fraught with difficulties. Much will depend upon whether the mother encounters problems, and her perseverance and determination in overcoming them. Cultural factors during her upbringing and the expectations of society are likely to influence her persistence, but for individual women little direct influence can probably be exerted over these.

Factors extrinsic to the mother

The pathway to success is not solely determined by factors intrinsic to the mother (or her baby) but also by the quality of care she receives both antenatally in preparation for breastfeeding and post-nasally. Several extrinsic factors can therefore be proposed that are likely to have a beneficial impact on the individual mother's chances of success:

  • good antenatal education and preparation
  • optimal post-natal support, delivered by appropriately trained health-care professionals
  • continued community support from a social network (partner, peers, and relatives, lay helpers or counselors, and health-care workers)

Despite the potential impact of individual-specific factors, it is imperative for the health-care system to offer wellstructured support, ensuring the mother's access to welltrained and motivated staff, to offer all women the best chance of success.

For convenience, the factors leading either to a successful or to an unsuccessful establishment of lactation will be addressed under three broad headings:

  • cultural influences on breastfeeding
  • perinatal practices in hospital
  • individual-specific influences

Before addressing these issues, I would like to focus on the physiological constraints on the establishment of lactation, looking at the factors that regulate both milk synthesis and effective milk transfer from mother to baby.

How long does it take to establish breastfeeding?

This key question can be addressed by reconsideration of classical physiology. Technically, the new mother must acquire the practical skills for handling her baby and positioning and fixing her baby on the breast. This may take from several days to a few weeks. The more fundamental physiological question is "How long does it take for a mother's milk supply to match her baby's needs?" This question has largely been overlooked, although the evidence has been available for some time.

What triggers the onset of milk production?

At one time the view was current that early and frequent breastfeeding was instrumental in the early establishment of milk supply [1]. The data, however, only support a benefit of frequent suckling: early initiation, per se, has been shown to be without impact [1, 2]. This is because separation and delivery of the placenta, not early suckling, acts as the primary trigger for lactogenesis (the onset of milk production). During pregnancy the breasts have been primed for milk production by a "cocktail" of maternal hormones, some intrinsic to the mother (prolactin, growth hormone, insulin, thyroid hormone) and others originating from the placenta (oestrogen, progesterone, placental lactogen, chorionic gonadotrophin). At birth the placenta is delivered, thus removing a significant source of the steroid hormones (oestrogen and progesterone) that have blocked the action of circulating prolactin upon breast tissue. Lactogenesis then begins, and measurable increases in milk output are detected between 30 and 72 hours after delivery [3, 4]. It is important to appreciate this fact, as retained placental products can be one of the few physiological constraints on the establishment of milk production. Any remaining placental tissue can secrete sufficient steroid hormones and thus continue to suppress the onset of milk synthesis [5].

Variability in lactogenesis

There are dramatic differences between women in both the initial changes in the rate of milk synthesis and the initial level of milk production. There can be a fivefold difference in the initial level of milk output, with infant intakes of between 200 and 900 g/24 hours on the fifth day post-partum [6]. This initial level is unrelated to birthweight; only at four to six weeks of age does infant weight predict intake [79]. Thus, production rates more likely reflect intrinsic variability between women. Physiological factors likely to affect this include variation in the amount of secretory tissue, in circulating hormone levels, and in the sensitivity of response to these hormones, a function of the hormone receptors associated with the secretory tissue.

One implication, however, is that milk production (triggered by placental delivery) is established without clear reference to the baby's size; it is only as a result of the interplay between the mother and her infant that the broad limits on milk output become refined to match the baby's needs. This, in effect, is a "calibration" process, which will involve upregulation of milk supply in the majority of women, whose initial milk output is low, but occasionally down-regulation of supply in women for whom it is initially excessive.

The crucial point is that breastmilk production (supply), which has been initiated "blind," must be progressively fine-tuned (calibrated) to meet the baby's needs (demand). If during this time the baby were offered calories from another source (for example, artificial formula), the blunting of appetite so caused might lead the mother's breasts to underestimate her baby's true needs, with the resultant danger that milk output might be calibrated at an inappropriately low level.

There is clinical evidence (personal observation, author's clinic) that in some women this process of down-regulation may be irreversible. So, if milk output fails to be optimized in the early weeks because of poor management, or if appetite is depressed because of competing calories from an alternative source, then once milk output has been set below the baby's overall needs, it may not be possible to reverse the process. For this reason, the avoidance of complementary milk feedings or "topup" bottles of formula is essential, as they are highly likely to interfere with the establishment of an adequate milk supply.

They can have a further non-physiological effect by undermining the mother's confidence in her milk supply, causing her to feed less frequently. In addition, an artificial bottle teat can represent a gratifying oral stimulus, in simple tactile terms, causing the infant to shift its preference from the breast to the bottle teat (this is discussed at greater length below). Pacifiers and nipple shields can pose a similar threat to the successful establishment of breastfeeding.

How long does this calibration process take?

Published data indicate that milk production continues to rise from birth to around four to six weeks, at which point it tends to reach a plateau [3]. A recent study has placed the population average for peak milk output at just short of five weeks [10], although this underplays the considerable variation that exists between women (five days and six months were the most discrepant points in a small sample of 30 women [6].

Nonetheless, evidence from a compilation of data (cross-sectional) from 16 studies of exclusively breastfed infants [3] indicates that milk output rises sharply until four to six weeks, then levels off at an average of 750 g/day for a singleton. (If a mother has twins, her milk output will continue to climb beyond 750 g/day to around 1,500 g/day, indicating that there is no absolute constraint on an individual woman's milk output, simply that 750 g/day represents the normal level at which milk output stabilizes for a singleton.) When the introduction of weaning foods is delayed, milk output is held at this level for weeks or months [11], until the introduction of solids causes fewer breastfeedings to be offered and milk supply to decline. In many cultures, however, weaning foods are introduced early. In Thailand, for instance, they are invariably introduced by one month of age. In such circumstances, there is a suggestion that they compete calorically with breastmilk, driving milk output down, rather than truly supplementing intake [10].

Factors affecting the establishment of lactation

Cultural Influences on breastfeeding

Breastfeeding fails far more commonly for cultural reasons than for biological reasons. There are few physiological constraints on a successful outcome for breastfeeding. In contrast, the wealth of factors impacting upon the individual woman from family, friends, health-care workers, and the world at large exerts an enormous influence on her. Just as these factors are acknowledged to influence the probability of her initiating breastfeeding, so, too, they are likely to exert an influence over her tendency to continue once she has started.

It is axiomatic to a successful outcome that women should want to breastfeed; women who encounter and overcome breastfeeding problems usually attribute their success to their own perseverance and determination. How do health-care workers inspire in women the desire to breastfeed, and how do they influence their motivation to succeed?

Women's attitudes and beliefs are shaped by their own experience but are also fashioned by cultural pressures, and it may be difficult for health-care workers to exert much influence over these. Nonetheless, epidemiologic research continues to identify the widespread health benefits of breastfeeding to both the mother and her infant [12-14]. If women are to make a truly informed choice on how to feed their infants, then health-care workers have an obligation to provide mothers antenatally with all relevant information about these benefits and about the hazards associated with artificial feeding, irrespective of any pre-existing choice the mother may have.

The role of health-care workers

In the past, a supportive environment for the new mother would have been provided by the extended family, whose members would have undertaken most of the routine tasks that would otherwise fall to the mother. In traditional rural communities of the developing world, cultural taboos on undertaking household tasks may apply for four to six weeks after delivery [15], the time taken for the milk supply to become established. The absence of the extended family in industrialized cultures suggests that motherhood generally is undervalued. In many cultures, particularly those in transition, family ties have been eroded and the extended family has become so fragmented that new parents may have little contact with relatives who would traditionally have provided help and support. Furthermore, in those cultures where bottle-feeding has predominated over the past decades, a vast cultural expertise on breastfeeding has been lost. Under such circumstances, one would hope that health-care workers (midwives and the primary health-care team in the community) would be in a position to make up these deficiencies by providing a collective professional expertise. In practice, however, such an idealized picture does not exist. Little emphasis is given to the management of breastfeeding in professional training, and health-care workers commonly lack both the breadth of knowledge and the core practical skills to enable them to manage lactation effectively. This situation should be rectified as soon as is practicably achievable, by including lactation management in the core training curriculum of every health-care worker whose designated role includes management of the nursing couple (midwives, for example). The emphasis should be on the acquisition of key skills through the mentor-apprentice system, and the combination of knowledge and skills is essential if women are to receive the expert support they need from health care staff.

Perinatal practices that affect the establishment of breastfeeding

It is legitimate to view the factors that either help or hinder the establishment of breastfeeding from the perspective of the maternity facility providing health care. Although largely unpredictable events can shape an individual mother's chances of establishing breastfeeding, there is enormous scope for hospital practices and staff attitudes to impact upon breastfeeding success.

  • In general, these factors can be grouped as follows:
  • antenatal parent education
  • quality of professional training in knowledge and skills
  • hospital policy on key practices
  • extended post-natal support, including both routine community support and specialist clinical support

The quality of support in each of these areas can be set through policy formulation and implementation by the maternity facility.

The impact of mafernity care practices

The expectation is of a supportive health-care system, but the reality may be even worse than indicated above. Too often, medical practices in hospital have impeded or hindered the successful establishment of lactation. Historically, separation of the mother and baby immediately after birth was rationalized without empirical justification [16]; such a policy necessitated nurseries for the care of the newborn, which further encouraged the routine giving of supplementary fluids, including formula. Abolishing nurseries as a concept and encouraging the mother to be the primary caretaker with responsibility for her infant through routine "rooming-in" is fundamental to reversing these outmoded practices.

A global campaign, the Baby-Friendly Hospital Initiative, sponsored by international health agencies (WHO and UNICEF) and supported by national governments, is now in place to reverse this situation, by ensuring that hospital practices are designed to protect and promote breastfeeding.

TABLE 1. Ten steps to successful breastfeeding

1. Have a written breastfeeding policy that is routinely communicated to all health-care staff

2. Train all health-care staff in skills necessary to implement this policy

3. Inform all pregnant women about the benefits and management of breastfeeding

4. Help mothers initiate breastfeeding within half an hour of birth

5. Show mothers how to breastfeed and how to maintain lactation even if they should be separated from their infants

6. Give newborn infants no food or drink other than breastmilk unless medically indicated

7. Practise "rooming-in"-allow mother and baby to remain together 24 hours a day

8. Encourage breastfeeding on demand

9. Give no artificial teats or pacifiers (also called dummies or soothers) to breastfeeding infants

10. Foster the establishment of breastfeeding support groups and refer mothers to them on discharge from the hospital or clinic

Source: ref. 7.

Under this initiative, hospitals are encouraged to match standards of post-natal care set out in the WHO/UNICEF "Ten Steps to Successful Breastfeeding" (table 1) and to seek external assessment against these criteria. If they are successful in demonstrating compliance with the global criteria, they can be designated as a "Baby-Friendly Hospital," in effect, a global charter mark in recognition that they have made a policy commitment to making the mother and her baby the focus of a supportive, caring environment in which breastfeeding can flourish. The core practices of the Ten Steps are as follows:

Policy formulation.. To minimize the tendency for new mothers to be given conflicting advice, it is vital for all health-care staff to adhere to the same set of practices. The best way to do this is for the maternity facility to establish a clear policy for the management of the new mother and her infant, training all health-care staff to a level where they understand the need to comply with this policy and are able to implement it with confidence. Monitoring staff compliance, breastfeeding intention, and outcome should then become a routine part of the audit cycle. Policies can be refined only if they are strictly implemented, followed by routine monitoring of specific clinical outcomes. Only then can the policy be modified to reduce any adverse outcomes.

Professional training. This requires a commitment to increasing the knowledge and skill base of hospital staff. Eighteen hours of training in lactation management, including at least three hours of supervised clinical practice, is urged by WHO/UNICEF as the minimum necessary for any member of the health care staff (previously untrained in lactation management) who has contact with the breastfeeding mother and her infant. Nonetheless, if routine training (a midwife, for example, might receive prequalification) combined with post-qualification skill development enable nursing staff to overcome routine problems of breastfeeding, these can be deemed both adequate and appropriate, whatever the duration.

Antenatal education. The quality of antenatal education, including whether the proposed method of feeding receives adequate discussion, is highly influential in determining whether women make a truly informed choice on how to feed their infants. It is also to be hoped that all potentially breastfeeding women receive sufficient preparation to avoid or overcome several common initial breastfeeding problems.

Early contact between mother and baby. Early initiation of skin-to-skin contact between the mother and infant, leading spontaneously to the first breastfeeding, is likely to be of fundamental importance in establishing close affectionate ties between them. In the immediate postpartum period, the mother is regarded as passing through a "sensitive period" [18], and the baby shows raised alertness at this time [19]. Although early skin-to-skin contact with the mother will undoubtedly help to reinforce the newborn's ability to seek and locate the breast [19, 20] and to facilitate recognition of the newborn [21, 22], there is little functional need for the altricial human newborn to "imprint" on its mother. For the human infant, it is more important to secure the mother's sustained emotional attachment, so that she continues to provide vital warmth, nourishment, comfort, and protection.

Therefore, early mother-infant contact serves the infant primarily by securing maternal attachment [23] and, as such, it is a biological imperative. Ensuring prolonged highquality contact in the immediate post-partum period, therefore, deserves to be a core practice of intrapartum care, irrespective of how the mother subsequently chooses to feed her infant.

Although there is some empirical support for a maternal sensitive period, there remains the question of whether giving high-quality skin-to-skin contact at a later stage may not also be capable of facilitating the development of emotional attachment, or, conversely, whether subsequent unrewarding or fractured breastfeedings may undo any initial benefits. Of the many studies conducted to evaluate the impact of enhanced early access of the mother to her infant, only one [18] has considered whether there are any deficits if the enhanced contact (45 minutes private, skin-toskin) is provided at a later stage (12 hours after delivery) compared with immediately after delivery. It showed a nonsignificant reduction in affectionate behaviour. A significant graded reduction in affectionate behaviour was observed, compared with a control group of mothers permitted only visual contact with their babies after delivery. Therefore, the evidence is rather slight to contest the possibility that prolonged, high-quality, skin-to-skin contact offered at any subsequent time might be capable of matching some, if not all, of the benefits of early mother-infant contact. If the mother has been deprived of early contact because of intrapartum medication or surgical procedures, high-quality contact should be encouraged with as much enthusiasm at a later stage. Furthermore, professional care should always be available during early breastfeedings and should strive to ensure they are as rewarding and problem-free as possible.

Further key practices. Although practices such as rooming-in and demand feeding may be regarded as essential to breastfeeding success, at a more general level they may be regarded simply as extending the newborn infant the courtesy of being regarded as an individual in its own right, rather than as an imperfectly formed automaton to be directed as we see fit. This change in philosophy in the way we view the human newborn is essential if we are to recognize that the newborn has much right to choice as the mother, although there is an obvious practical constraint on this choice being "informed." If the infants were in a position to make an informed choice on how they would be fed, on health grounds they would undoubtedly choose to be breastfed. It is difficult to see what cultural objections they might raise! So, we must learn to set their rights alongside the mother's and the father's in reaching a decision about the method of feeding. Although it is possible to constrain the infant's pattern of feeding to conform to a Western "idealized" cultural norm- as Pinilla and Birch [24] have shown for night feedings-there is no evidence that this benefits long term habit formation in the child, whereas it may prove detrimental to the infant's emotional development [25].

Avoiding bottles and teats. We have already dealt with the physiological reasons for avoiding unnecessary supplementation and the use of artificial teats and pacifiers. Each of the WHO/UNICEF Ten Steps is research-based, and although this step (step 9) is perhaps the least well supported, it can be defended by arguing both from first principles and from clinical experience. It is important to reemphasize at this point that we are considering events during the critical period of the establishment of lactation. It is acknowledged that different rules may apply during established lactation, when we accept that the occasional offering of bottles for social reasons (preferably of the mother's own expressed breastmilk) is unlikely to disrupt breastfeeding success.

Evidence from a study in Brazil indicates that the use of pacifiers is associated with a shorter duration of breastfeeding [26], as predicted. Clinical experience also suggests that pacifiers can impede the successful establishment of breastfeeding. Although the mechanism is not clear, certain lines of evidence can be considered.

First, ultrasound studies [27, 28] indicate that during suckling the mother's breast distorts and conforms to the internal geometry of the baby's mouth. Milk removal is achieved by the baby's compressing the base of the teatlike shape formed from the breast and nipple between his gums, and expressing milk from the sinuses lying within the "teat bulb" with a peristaltic wavelike action of the tongue. On a bottle teat, the infant attempts to achieve the same action, but because the artificial teat is more rigid, there is. instead, a tendency for the soft tissues of the baby's mouth to deform to accommodate to the geometry of the teat. Expressing milk from the teat bulb may be resisted by the less compliant material from which the artificial teat is constructed, with the result that the baby adapts by a shift towards extracting milk more by suction than by peristaltic expression. If the baby develops a reliance on one method of milk extraction over the other, he may not adapt well when offered both breast and bottle or when shifting between the two. To date, no specific carryover effect has been scientifically demonstrated; it has only been inferred from clinical observation.

Additionally, artificial teats may constitute a "supernormal" sign stimulus [29, 30], causing greater tactile stimulation of the baby's mouth than the more compliant breast. Whereas the breast naturally retracts elastically, the artificial teat remains in place, requiring little effort from the baby to hold it in his mouth. The supernormal stimulus, represented by the artificial teat, can compete for attention with the natural stimulus of the breast, and the baby, if exposed to it at an early stage, may be unable to ignore it. The outcome is that the baby may develop a preference for the artificial over the natural (nipple shields can be comparable in this respect). The term nipple confusion has been coined for the behaviour of the baby who refuses the breast in favour of the bottle or pacifier, but such a term may be insufficiently descriptive. The term acquired teat preference may be more accurate and would still embrace the established breastfeeder who refuses the bottle.

Another possibility is that bottle-feeding bypasses the triggering of the baby's natural adaptive reflexes for feeding and the maturation of the natural repertoire of responses necessitated by breastfeeding (rooting, gaping, drawing elastic breast tissue into the mouth, and suckling) [31]. There may also be a critical window during which the infant's natural response repertoire matures, so that time spent sucking on an artificial teat may represent both a lost opportunity and enhancement of an aberrant response repertoire.

Finally, the excessive use of soothers can mean that the baby is put to the breast either too infrequently or for too short a time, with the result that the baby does not secure sufficient nutrition. Not only can this adversely affect the infant's nutritional status, it will also disrupt the mechanism of supply and demand. At older ages, we have seen infants who are growing inadequately or are failing to thrive simply because pacifiers have been introduced. This appears to be due largely to the net reduction in suckling time and its effect on both immediate intake and future supply, which is only reversed when the mother makes active efforts to withhold the pacifier.

The use of pacifiers can also cause the earlier return of fertility, resulting in closer child spacing. In many cultures this may have an adverse impact on maternal as well as infant health [14, 32].

Cup-feeding. One way around the problem of giving additional fluids, where they are medically indicated, is to give them by cup or spoon [33]. In many centres, cup-feeding has been successfully reembraced and appears to be associated with a greater success rate for establishing breastfeeding.

A multicentre trial has been initiated in the United Kingdom to formally test such a view.

Individual-specific influences during the establishment of lactation

Practices that can either promote or impede breastfeeding success for the individual mother and her newborn can be broadly grouped into the following categories:

  • obstetric procedures intrapartum and use of analgesia
  • paediatric medical interventions necessitating admission to the neonatal intensive-care unit or step-down care nursery
  • maternal pathophysiology of milk production
  • routine post-natal practices affecting mother and infant care
  • quality of post-natal care and support

For the individual mother, there may be unexpected predisposing factors that militate against the successful establishment of breastfeeding.

In rare circumstances, the baby may be born with an oral anomaly that will require a specialist's attention, or the nature of the baby's birth (for example, pre-term, small for gestational age, or congenital abnormality) may require separation from the mother for medical care. Although medical services should be able to tackle many of the clinical problems with relative ease, the role of the mother at this time can often be overlooked. The one significant thing the mother may be able to do for her infant under these circumstances is to express her breastmilk, yet too few centres recognize the need to attach equal significance to this non-acute aspect of clinical care.

Excluding such medical crises, however, the baby may be sleepy or lethargic as a result of intrapartum analgesia (as from pethidine) [20] or may be irritable following an instrumental delivery. The resulting behaviour, either a lack of response by the baby or signs of discomfort when the baby is handled, may make it very difficult for the novice mother to handle her baby positively and with confidence, and she may feel her efforts to breastfeed are being rejected. Concern should also be raised over increasing reports that modern epidurals and gastric suctioning, either alone or in combination, adversely affect the initiation of breastfeeding; their impact should be evaluated epidemiologically.

The foetus may have had access to its thumb, fingers, or hand in the womb (often observed during antenatal ultrasound examination) and may have had the opportunity to acquire an artificial sucking style. If the foetus has become used to sucking on a discrete, relatively rigid object for several weeks, the infant may find the elastic, retractile tissue of the nipple and areola less attractive in tactile terms and thus may be reluctant to adapt to sucking at the breast. This theory could be readily evaluated by independent assessment of ultrasound scans and early sucking behaviour.

The main reason for raising this theoretical possibility is that it represents a potential individual-specific predisposing factor. We cannot yet predict which babies will or will not find the offering of artificial teats counter-productive to the establishment of breastfeeding, but simple adherence to the policy recommendation in step 9 of the Ten Steps should avert such a problem. It is explicit that this policy relates to the management of the normal full-term infant and that there are permissible clinical circumstances for departing from it.

Caution should be exercised, however, to protect against two harmful inferences: that because a baby has been observed on ultrasound to be thumbsucking in utero, problems will automatically arise with breastfeeding; or that difficulties of fixing and attachment in the early days are due to antenatal thumbsucking, so that no remedial efforts will be made. Neither of these inferences is defensible. Rather, the issue is raised because awareness of the possibility may help caregivers provide appropriate support rather than leaving the mother on her own to come to terms with an unrewarding infant who refuses to feed from the breast. It may be necessary for this baby to relearn how to suck correctly in the context of breastfeeding, in which case the offering of bottles, pacifiers, or finger foods will only perpetuate the artificial oral preference. A baby's early tendency to refuse the breast will undermine the learning experience for the new mother, whereas acknowledgement of her predicament and the offering of appropriate help and support will be encouraging and empowering.

Successful achievement of the technical aspects of breastfeeding will also be militated against if the mother is given inexpert help in attaching her baby at the breast in the early days [34]. Forcing her baby to the breast, or holding him there against his will, is likely to frustrate the woman's own efforts and can lead to many common complaints. The baby may also find these efforts aversive and develop a behavioural coping strategy that presents itself as breast refusal or rejection; once developed, this behaviour can be very distressing to both mother and infant and is often difficult to reverse.

Finally, research conducted with professional referrals to a clinical support service for breastfeeding women [6] has identified several classes of problems, all of which may be perceived as breastmilk insufficiency. While 98% of professional referrals to the clinical service were for reliable symptoms (for example, inadequate weight gain, unsettled infant behaviour, and sensations of reduced supply), up to 85% of them could be reversed by straightforward practical steps that included:

  • technical improvements to physical aspects of the breastfeeding process
  • optimizing feeding management (feeding frequency and duration, pattern of breast usage)
  • providing support and encouragement

Four demonstrable causes of insufficiency were identified. Genuine pathophysiologies of milk production represented only 2% of clinical referrals. Another category (5% of referrals) represented physiologically low milk output, while the remaining two (8% of referrals) would normally be regarded as lactation failure of maternal origin, but were in fact acquired conditions owing to suboptimal management of the nursing couple (for more detailed discussion, see ref. 6).


The factors most likely to contribute to a mother's initial success with breastfeeding are support and encouragement provided by health-care workers who are skilled and knowledgeable in managing lactation. These professionals must work within a system that is similarly caring and supportive, the policy of which recognizes the intrinsic rights of both the mother and the baby, allowing them unrestricted access to each other (as would be the case if the mother delivered at home). More proactively, the health care system and its staff should positively encourage those practices that have been shown to have a beneficial impact on the establishment of breastfeeding, while eradicating all practical hindrances. Finally, it is to be hoped that the mother will have a relatively trouble-free breastfeeding experience, or at least that her motivation and persistence with breastfeeding will be sufficient for her to overcome any problems she may encounter. A sensitive and supportive attitude from health-care staff will be highly valued by women and can contribute significantly to their success by boosting their confidence in their ability to breastfeed.


1. Salariya EM, Easton PM, Cater Jl. Duration of breastfeeding after early initiation and frequent feeding. Lancet 1978;ii:11413.

2. Woolridge MW, Greasley V, Silpisornkosol S. The initiation of lactation: the effect of early versus delayed contact for suckling on milk intake in the first week

post-partum. A study in Chiang Mai, Northern Thailand. Early Hum Dev 1985;12:269-78.

3. Neville MC, Keller R. Seacat J. Lutes V, Neifert M, Casey C, Allen J. Archer P. Studies in human lactation; milk volumes in lactating women during the onset of lactation and full lactation. Am J Clin Nutr 1988;48:1375-86.

4. Hartmann PE, Owens RA, Cox DB, Kent JC. Breast development and control of milk synthesis. Food Nutr Bull 1996;17:292-304.

5. Neifert MR, McDonough SL, Neville MC. Failure of lactogenesis associated with placental retention. Am J Obstet Gynecol 1981;140:477-8.

6. Woolridge MW. Breast-feeding: physiology into practice. London: Royal College of Physicians, 1995:30-2.

7. Pao EM, Himes JM, Roche AF. Milk intakes and feeding patterns of breast-fed infants. J Am Diet Assoc 1980;77:540-5.

8. Butte NF, Garza C, Smith EO, Nichols BL. Human milk intake and growth in exclusively breast-fed infants. J Pediatr 1984;104:187-95.

9. Imong SM, Jackson DA, Wongsawasdii L, Ruckphaophunt S. Tansuhaj A, Chiowanich P. Woolridge MW, Drewett RF, Baum JD, Amatayakul K. Predictors of breast milk intake in rural Northern Thailand. J Pediatr Gastroenterol Nutr 1989;8:359-70.

10. Drewett R. Amatayakul K, Wongsawasdii L, Mangklabruks A, Ruckpaopunt S. Ruangyuttikarn C, Baum D, Imong S. Jackson D, Woolridge M. Nursing frequency and the energy intake from breast milk and supplementary food in a rural Thai population: a longitudinal study. Eur J Clin Nutr 1993;47:88091.

11. Howie PW. Protective effect of breastmilk against infection. Food Nutr Bull 1996;17:373-83.

12. Dewey KG, Finley DA, Lönnerdal B. Breast milk volume and composition during late lactation (7-20 months). J Pediatr Gastroenterol Nutr 1984;3:713-20.

13. Victora CG. Infection and disease: the impact of early weaning. Food Nutr Bull 1996;17:390-96.

14. Rasmussen KM, McGuire MK. Effects of breastfeeding on maternal health and well-being. Food Nutr Bull 1996;17:36472.

15. Schmitt BD. Prevention of sleep problems and colic. Pediatr Clin North Am 1986;33:763-74.

16. Thomson M, Westreich R. Restriction of mother-infant contact in the immediate postnatal period. In: Chalmers 1, Enkin MW, Keirse M, eds. Effective care in pregnancy and childbirth. London: Oxford University Press, 1989:1322-30.

17. Protecting, promoting, and supporting breastfeeding: the special role of maternity services. SCN News 1989:4.

18. Hales DJ, Lozoff B. Sosa R. Kennel JH. Defining the limits of the maternal sensitive period. Dev Med Child Neurol 1977;19:45461.

19. Widstrom AM, Ransjo-Arvidson AB, Christensson K, Matthiesen AS, Winberg J, Uvnas-Moberg K. Suction in healthy newborn infants. Acta Paediatr 1987;76:56672.

20. Righard L, Alade MO. Effect of delivery room routines on success of first breast-feed. Lancet 1990;336:1105-7.

21. Macfarlane A. Olfaction in the development of social preference in the neonate. In: Porter R. O'Connor M, eds. Parent-infant interaction. Ciba Foundation Symposium 33. New York: Elsevier, 1975:103-17.

22. Varendi H. Porter RH, Winberg J. Does the newborn baby find the nipple by smell? Lancet 1994; 334:989-90.

23. Gubernick DJ. Parent and infant attachment in mammals. In: Gubernick DJ, Klopfer PH, eds. Parental care in mammals. New York: Plenum Press, 1981:243305.

24. Pinilla T. Birch LL. Help me make it through the night: behavioural entrainment of breast-fed infants' sleep patterns. Pediatrics 1993;91:436-44.

25. Bowlby J. Attachment and loss. Vol. 1. Attachment. London: Hogarth Press, 1969.

26. Victora CG, Tomasi E, Olinto MTA, Barros FC. Use of pacifiers and breastfeeding duration. Lancet 1993; 341 :404-6.

27. Weber F. Woolridge MW, Baum JD. An ultrasonographic study of the organisation of sucking and swallowing by newborn infants. Dev Med Child Neurol 1986;28:19-24.

28. Bu'Lock F. Woolridge MW, Baum JD. The development of the coordination of sucking, swallowing and breathing: an ultrasonographic analysis of term and preterm infants. Dev Med Child Neurol 1990;32:66978.

29. Tinbergen N. The study of instinct. London: Oxford University Press, 1951:13-31.

30. Gunther M. Instinct and the nursing couple. Lancet 1955;i:5758.

31. Hopkinson J. Garza C. Management of breastfeeding. In: Nutrition during infancy. St Louis, Mo, USA: CV Mosby, 1988:298-313.

32. McNeilly AS. Breastfeeding and the suppression of fertility. Food Nutr Bull 1996;17:340-8.

33. Lang S, Lawrence CJ, Orme RL. Cup feeding: an alternative method of infant feeding. Arch Dis Child 1994;71:365-9.

34. Royal College of Midwives. Successful breastfeeding. 2nd ed. Edinburgh: Churchill Livingstone, 1991.

Discussion of paper by Woolridge

Discussion of paper by Woolridge

Dr. Valdés

I have a question regarding the group of women with an irreversible decline in milk output. You can use lactation in adopted mothers as a comparison. It takes longer, and the production doesn't increase as in the first month. It steadily increases throughout the first six months. Why doesn't that happen in a biological mother?

Dr. Woolridge

Our finding is disconcerting, especially when one hears about grandmothers in Africa who are able to re-establish lactation simply through suckling. Some of our interventions to boost these mothers' supplies have been quite aggressive, to see if there was any way of improving output. We would not diagnose the mothers who responded either to increased suckling or to pharmacotherapy. Where there has been substantial down-regulation, it is often difficult to shift back up. Where there has been up-regulation from birth, the potential to carry on up-regulation is retained. Downregulation, if it is excessive, can be a harmful process. Clearly there are other factors, particularly psychosocial factors in women in industrialized countries, that seem to exert a block over and above the motivation to feed. But in physiological terms, this down-regulation could be a critical factor.

Dr. Rasmussen

I am wondering if any of the babies in what you call the behavioural category, contented failure to thrive, are the ones who were rejecting the breast. Is there any indication that they are responding to the taste of the milk? That is, are mothers eating things that the infants are allergic to or just don't prefer?

Dr. Woolridge

There is no evidence of anything long-term. Certainly there is evidence that the baby will refuse or reject the breast for a brief period, either with the return of the menstrual cycle or from a particular food that the mother includes in her diet. But we haven't seen any long-term cases of breast refusal. I will cite two anecdotes, however. One is a mother who had a severe case of thrush, and she had been given a whole set of pharmacological preparations by her general practitioner. She was putting Timodene, which is for athlete's foot, on her nipples, and the baby was going to the breast but pulling back. She also smoked, and we weren't sure if it was the taste of the nicotine or tar products in her milk. While chatting with her, I squeezed some of this Timodene out onto my finger and tasted it, and it was quite revolting. Clearly, the Timodene was responsible for this baby's reluctance to go to the breast, and that had been going on for eight weeks.

The next example is also quite disconcerting. Recently we saw two babies whose first breastfeedings after delivery had been interrupted to give oral vitamin K. The babies grimaced at being given the vitamin and then refused to breastfeed for the next five days. Now, if you want to teach a baby what is harmful and bad, you give it a feeding and then follow it with a distasteful flavour, a perfect case of one-trial learning. You have to be very careful about when oral preparations are given to the baby, and not give them in association with the first or subsequent breastfeedings.

Dr. Prentice

I wish to continue with the discussion about irreversibility, because certainly in our Gambian experience it is difficult to see how that fits in. Certainly, we see women whose breastmilk volume goes way down in the wet season, possibly because they have become very anorexic with various infections. Then their breastmilk volume will come back up again, will double, and they will be well past the six weeks of lactation at that stage.

Dr. Woolridge

If one accepts that the calibration process takes four to six weeks, the results of down-regulation may differ if the weaning process invades that critical period or comes after it.

Dr. Prentice

What we see is at a later stage. This makes me wonder about the underlying physiology: for example, the role of the number of lactation cells, and what affects their longevity. Are we observing down-regulation? Is the process truly irreversible? What proportion of your subjects producing less than 450 ml of milk were in the irreversible group? Could those you were calling lazy children get used to being down-regulated? Would those mothers respond by increasing milk volume?

Dr. Woolridge

There is some overlap between the groups in about 4% of the clinical population. So I am not suggesting that we should suddenly recalibrate our views, but it is a possibility to be aware of; that is, the process of down-regulation can create clinical dilemmas.

Dr. Victora

I have a question about one of the Ten Steps to Successful Breastfeeding, which relates to the use of pacifiers or dummies. We looked into the literature a couple of years ago and found virtually nothing. Since then, we have done a couple of studies in Brazil that showed that children who had a pacifier introduced in the first month of life were three times more likely to be weaned in the subsequent three to four months than children who were not using a pacifier. The magnitude of this effect was about as large as that of introducing a bottle with artificial milk, so we were quite impressed by our finding. Obviously, we can't know whether this is a cause-effect association, or if pacifier introduction is a maternal strategy for weaning the child. I wonder if you have any data on that, and also if you could let us know if you have had any success in the United Kingdom in preventing mothers from introducing the pacifier?

Dr. Woolridge

We did try to discourage the use of pacifiers up to the time Ed Mitchell found that pacifiers might be protective against sudden infant death syndrome. We are still quite anxious to find out what the explanation for that particular finding is. It is an area where there is huge reluctance, as many women in the United Kingdom use pacifiers quite extensively. Here is one scenario of how these behavioural problems come about. Let's say you have a baby who has had access to his thumb or fist in utero and has become used to sucking on a small, rigid, discrete object. Suddenly the object is no longer available. The baby is encouraged to go the breast, but it doesn't find the breast as physically discrete in terms of its tactile qualities. The baby initially refuses the breast, but the midwife is rather heavy-handed and insists that the baby go to the breast. The baby starts to fight and refuses the breast, and after about three days of this, someone eventually gives in and gives the baby a bottle. The baby thinks, "Wonderful, this is what I have been missing out on," as this is closer to the thing that the baby has been used to. There is some evidence in some babies that, if you like, some antenatal preconditioning changes their expectation of what they should get afterwards. I will show you one last slide. There is much talk about "nipple confusion," which is an expression I don't like. Having trained as a zoologist, I thought I would bring you a piece of Dutch ethology. This is a slide showing how you can fool an oystercatcher into incubating the wrong clutch of eggs. If you give it a clutch of eggs that is larger than its normal clutch, it will incubate those in preference to its own clutch. If you give it a football that is painted to look like its eggs, it will try to incubate that. This is what is called a supernormal stimulus. It's so good at switching on the animal that it can't be refused. Nature didn't expect scientists to come along and paint footballs the colour of eggs. I do think that bottle teats and pacifiers represent supernormal stimuli for the baby. If you like, they stimulate the mouth with a greater tactile strength. For some babies, if breastfeeding is unsatisfactory, they will actually refuse the breast in favour of this more tactile object.

Professor Hanson

You made a comparison between lactating mothers in the United Kingdom and in Thailand. I would like to ask if you could qualify that a bit. Does this mean you are comparing British mothers who are exclusively breastfeeding with mothers in a traditional society, where I am convinced they rarely breastfeed exclusively? For instance, on the Indian subcontinent, a number of traditional foods that are introduced early to the infant are heavily contaminated and may have severe consequences for the infant's future life. A number of severe infections originate from this feeding practice. Once they start other foods, exclusive breastfeeding is rare. Could you speak about this?

Dr. Woolridge

I think the boundaries are very blurred. In Thailand some women initiate weaning foods purely on an experimental basis, introducing very low levels of supplement from, say, one month to three months of age. It's only at about four months that they start to give those weaning foods in substantial amounts, but I agree about the potential contamination. What is intriguing in Thailand is that many of the traditional methods of preparing weaning foods make the food sterile. They will steam rice in a vine leaf, which removes the pathogen load. In the United Kingdom very few mothers actually exclusively breastfeed. The specific recommendation of the Department of Health is to delay the onset of weaning until three months, but when you survey women, you find that most of them have offered their babies something by 10 weeks of age. So in fact there is not quite the discrepancy you might expect between these two cultures.

Professor Hanson

I will not be using the term weaning food in the same way that you do, because from the start of breastfeeding, the whole family-for instance, in Pakistani societies-will participate in feeding the baby and will put all kinds of food in the baby's mouth. This certainly brings the risk of infection and doesn't bring much energy or useful food to them. It is a risk that I think is often disregarded. You would need someone to stay with the family to find this out. What has been called exclusive breastfeeding has often not been.

Dr. Hartmann

I would like to go back to the relationship between socioeconomic class and breastfeeding. The same sort of thing happens in Perth, Australia. In the high socio-economic class about 80% of mothers breastfeed to six months of age. But this is a very recent behaviour. I think the problem is that we haven't successfully reached the lower socioeconomic classes with the breastfeeding message, and that is why breastfeeding rates have reached a plateau.

Dr. Woolridge

I agree, and unfortunately that group is often also very difficult to target in many other respects, such as in relation to the discontinuation of smoking and the way in which they use their financial resources.

Dr. Rasmussen

I want to come back to the comment on the pacifiers and offer you an alternative theory from the maternal perspective. A hungry breastfeeding infant is a vigorous young person who may make the mother's nipples quite sore. The introduction of a pacifier can have what I call the "nipple-saving effect" on breastfeeding continuation in women who might otherwise discontinue. Now in that case you may have women whose breastfeeding is already very adequate, and the pacifier has no effect whatsoever. It may be superior to giving the baby other foods, or it might not be.

Dr. Woolridge

Clinically, our objective when we are counselling women with distressed, unsettled infants is based on the premise that the infant has an appetite-control mechanism and that the most likely regulator of that control is fat intake. We try strategies that maximize the efficiency of milk delivery, so that she can help her baby achieve a state of satiety. You are quite correct that, even when she has an ample supply of milk, if the baby is unsettled, she is likely to resort to a pacifier to find some other way of settling the baby, rather than have her nipple traumatized in the process. Clinically, we focus on the primary target of ensuring delivery of synthesized milk, and then we address colic and other concerns as secondary matters.

Dr. Valdés

One thing I was surprised to see was the differences between economic classes. In the developing countries it is just the other way around: our lower-economic-level mothers nurse much more than those at higher levels. At the same time, we feel that mothers in the lower economic groups are easier to reach. We are much more successful in addressing and changing their behaviours.

Dr. Woolridge

Certainly, there are ways in which it is quite inappropriate to compare an industrial country with Thailand. In the United Kingdom in 1940, the picture is reversed. What is surprising is how difficult we are finding it to target women in the promotion of such a recently acquired practice. Artificial formula was quite appropriately developed as an emergency backup option, but the trouble is that it has become the first choice of too many people.

Dr. Valdés

I have some comments related to pacifier use. I can't understand why the mother would have sore nipples.

Dr. Rasmussen

Poor positioning would do it.

Dr. Valdés

If the mother correctly positions the infant when she breastfeeds, for whatever time, she won't have sore nipples. But what we observe, in waiting rooms for example, are mothers who naturally put the baby to the breast when the baby cries. There, you have short feedings. The mother who uses a pacifier instead of putting the baby to the breast loses an opportunity to feed.

Dr. Pollitt

Just a short comment in connection with this issue of "nipple confusion." We have to remember that sucking on a pacifier is different from sucking on a breast. With the pacifier there is no negative pressure, and the activity of the tongue differs.

Dr. Woolridge

There is also no fluid delivered, so there is no need for phase-locked sucking, swallowing, and breathing, whereas on the breast these different activities have to be phase-locked. On a pacifier breathing is carried out independently from sucking.

Professor Hanson

I'd like to come back to the point of motivating women in poor countries to breastfeed exclusively. We were impressed to find that a motivation campaign in very traditional areas-villages, subslums, and city slums-results in substantial increases in this practice. Even traditional societies can be approached very successfully.

Social and biological determinants of lactation

Social and biological determinants of lactation

Salvador Villalpando, Samuel Flores-Huerta, Mardia López-Alarcón, and Ignacia Cisneros-Silva


During the last decades breastfeeding rates have fallen to such a level that at least one

million infant deaths annually have been attributed to the replacement of breastfeeding by formula-feeding. After the recent World Summit for Children, unprecedented efforts were targeted to reverse this trend. The evolution of breastfeeding patterns and related practices in Mexico is presented as a case study. Before the Baby-Friendly Hospital initiative, the rate of breastfeeding initiation and the average breastfeeding duration in Mexico were among the lowest in developing countries. The use of formula in hospitals was common. Knowledge about breastfeeding and awareness of its advantages were low among the public and health professionals. The International Code related to breastmilk substitutes became law, and the Baby-Friendly Hospital Initiative was included in the National Plan for Health in 1991. Formula manufacturers decided voluntarily to stop advertising their products. After five years, several surveys show noticeable changes in hospital policies and practices regarding breastfeeding promotion and improvement in the rate of breastfeeding initiation and exclusivity. Total duration of breastfeeding, public awareness, and mothers' knowledge about the basics of breastfeeding are increasing. Despite continued inadequate post-partum support and counselling, and an inconsistent effectiveness in conveying to the public the benefits of breastfeeding, progress has been made in the last five years and will be reflected in improved infant health in the near future.


Humans are the only mammals whose offspring survives without natural lactation. This has been made possible on a worldwide scale by technological developments for modifying and preserving milk from other species [1]. As a result, the breastfeeding rate in developed countries has fallen dramatically. In the mid-1940s, about 50% of infants were breastfed up to three months in England and Australia; by 1971, only 25% of American babies were breastfed. Concerns regarding the negative effects of artificial feeding of infants resulted in the active promotion of breastfeeding as a healthier feeding mode. Currently, up to 80% or more of mothers breastfeed up to six months in some countries, although rates are highly variable [2].

A similar decline in the rate of breastfeeding has occurred in developing countries. It started 20 to 30 years later than in their affluent counterparts. The deadly consequences of such a decline are eloquently summarized by Stephen [3] in his report as a representative of the United States Government to the 34th World Health Assembly. He pointed out that at least one million infant deaths in developing countries are due to diarrhoea and malnutrition associated with formula-feeding. As part of the multinational agreement signed during the World Summit for Children, important national and international efforts were directed towards curbing and reversing the prevalence of formula-feeding in developing countries.

Mexico: A case study

To address the social and biological determinants of lactation establishment, I will present the case study of Mexico as an example, because Mexico lags behind many Latin American countries in improving the frequency and duration of breastfeeding [4]. The social determinants will be classified as (1) the general framework represented by the economy and the pertinent national policies, (2) the organizations acting as health-care purveyors, (3) the training and commitment of health-care professionals, and (4) the individuals participating in the process of lactation.

The magnitude of the problem in Mexico

Before the Baby-Friendly Hospital Initiative started its operation in Mexico in 1991, the frequency of breastfeeding was highly variable in rural, urban, and metropolitan communities. Recently, Pérez-Escamilla reviewed the studies reporting the prevalence of breastfeeding in Mexico from 1953 to 1988 [5]. In these studies the percentage of women initiating breastfeeding ranged between 84% and 99% in rural and urban populations. For the purpose of this article, the data originally categorized as urban by PérezEscamilla were separated into urban and metropolitan, defining metropolitan as cities of more than one million inhabitants during the year the studies were conducted. This distinction is relevant because of differences in rural-tourban migration rates and in the availability and accessibility of health care services. As seen in table 1, breastfeeding prevalences at birth varied widely among rural (73% to 99%), urban (79% to 94%), and metropolitan (59% to 86%) populations. Urban and metropolitan populations tended to have lower prevalences of breastfeeding. The prevalence was progressively lower at three and six months postpartum in the three categories. The most notable reduction occurred in the urban and metropolitan groups. The wide variability in the prevalence might be attributed to differences in the operational definitions of breastfeeding, differences in sampling procedures, and the extreme differences in the sample sizes of the referred series (median, 105 subjects; range, 50 to 5,011).

A more reliable picture might be obtained from three nationwide studies [6-8]. The proportion of never-breastfed infants in the 1986 National Health Survey in the 32 Mexican states [6] ranged from 5% to 34% (median, 21%). Another study, representing all geographic regions and both urban and rural populations, found an overall prevalence of breastfeeding in the first months post-partum varying from 60% among the urban elite to 89% in rural populations [7]. These figures dropped to 18% and 54%, respectively, at six months. The National Nutrition Survey of 1988, based on a probabilistic sample, reported that 13.8% of infants younger than four months never received breastmilk, while 60% were fed breastmilk for at least four months, but only 11.5% were exclusively breastfed for this period. They found significant regional differences in the prevalence of breastfeeding and determined in a hazard analysis that better living conditions (a proxy for socio-economic situation), higher levels of maternal education, introduction of other milks, and living in urban municipalities were detrimental to breastfeeding. Data about exclusive breastfeeding, defined as being fed breastmilk but no other milks or liquids, were available in only two series [7-9]. The prevalence of these feeding categories at one month postpartum was 19% (7) and 23% (8, 9), and at three months, 0% and 3%.

Surveys conducted after the Baby-Friendly Hospital Initiative was launched indicate that the prevalences of breastfeeding and exclusive breastfeeding are increasing. In metropolitan Mexico City, the initiation of breastfeeding was reported as universal [10]. Within the first month postpartum, breastfeeding prevalence was 76%, with an 11% rate of exclusive breastfeeding; at four months post-partum, the rates fell to 57% and 4%, respectively. Preliminary data from a larger study conducted in 36 public and private hospitals in 1994 and 1995 showed that before hospital discharge, the overall prevalence of breastfeeding was 87%, and the prevalence of exclusive breastfeeding was 70% (E. Rios, personal communication, 1995). Another study combining data from mother-infant pairs participating in a breastfeeding promotion programme and data from controls reported a 93% prevalence of breastfeeding at one month post-partum, falling to 74% at four months post-partum. No data on exclusive breastfeeding were available [11].

TABLE 1. Median (range) prevalence of breastfeeding according to residence and age of child in the series reviewed by Pérez-Escamilla and Dewey [5]a

Age Rural Urban Metropolitan
No. series % breastfeeding No. series % breastfeeding No. series % breastfeeding
Birth 13 91
(73 -99)
9 85
5 66
3 mo 2 73.5
6 55
6 54
6 mo 9 77
4 52
6 36

a. The series represent samples or subsamples from individual geographic localities defined by the authors in each article.

Economics and national policies of breastfeeding

The manufacturers of infant formulas are frequently international enterprises. Their market interests go beyond formulas and even beyond specialized products for infants and children. In Mexico one manufacturer controls more than 75% of the infant formula produced within the country and more than 97% of the powdered formula. This almost monopolistic control gives manufacturers a great influence on market regulation [12, 13].

The International Code for Marketing Breastmilk Substitutes was adopted by the Mexican health authorities in 1989. Two years later, the Mexican Congress enacted the code into law. The Ministry of Health included the BabyFriendly Hospital Initiative in the National Plan for Health [14]. As a result, by the end of 1994, 15% of the 532 qualified hospitals in the country were awarded the title of Baby-Friendly Hospitals [15]. Before the Baby-Friendly Hospital Initiative was implemented in Mexico, formula and baby foods were advertised extensively and promoted heavily by hospitals, paediatricians, and nurses. A very high percentage of hospitals received free or reduced-price formula from manufacturers. Only 7 of 59 hospitals purchased all their formula, and only 2 paid full price [12]. Usually, hospitals were supplied by more than one manufacturer. In the majority of cases, hospital officers stated that such donations were unsolicited. In addition, manufacturers were allowed to distribute samples of formula to mothers at the time of hospital discharge. Mothers were approached outside the hospital premises in most cases, thus relieving hospitals from legal responsibility.

In 1993 all companies selling infant formulas in Mexico signed a voluntary landmark agreement committing themselves to stopping these practices. As a result, the report of the US Agency for International Development/Latin American and Caribbean Health and Nutrition Sustainability (USAID/ LACHNS) [16] found that distribution of free samples or gifts to mothers by formula companies almost vanished from the two hospitals surveyed. In a subsequent study of 36 hospitals conducted by Rios et al. (E. Rios, personal communication, 1995), officers from seven hospitals acknowledged accepting free or low-priced formula supplies.

Health-care Institutions

Perhaps the most contradictory aspect of Mexican healthcare institution policy related to maternal and infant health is the mandate by labour laws to provide mothers of infants a free six-month supply of a breastmilk substitute. In the past, conflicts within hospitals between the declared norm and the practice were frequent. In a survey of 59 hospitals in 1991, administrators claimed to have installed facilities and to have adopted routines favourable for breastfeeding. Nevertheless, investigators observed significant deviations from their claims [17]. For instance, none of the hospitals practiced early mother-infant contact. Although most claimed to have facilities for rooming-in, only 5% practiced it systematically. Most hospitals kept newborns in nurseries and provided mothers with portable cribs for the infants during the daily visits of the infants to the mothers' rooms. Infants were fed formula in the nurseries, having very few occasions to suckle at the breast. Only 4% of the hospitals allowed exclusive breastfeeding during the day; almost 16% did so during the night. The hospital schedules for breastfeeding were related more to staff shortages than to intentional policies. About half of the hospitals surveyed provided new mothers with written information about how to feed their babies at discharge. Although material encouraging breastfeeding predominated, it was astonishing to find that 20% of the hospitals provided information encouraging bottle feeding [17].

Preliminary data obtained by the follow-up study in 36 hospitals by the Nestle Infant Formula Audit Commission conducted by Rios from 1994 to 1995 (E. Rios, personal communication, 1995) provide evidence of a substantial improvement in those policies and practices. About half the mothers suckled their babies in their own beds, though only a small percentage were allowed to suckle within the first 30 minutes post-partum. About 80% of breastfeeding women had proper facilities in the hospital that enabled them to feed their infants throughout the day. At discharge 40% of the mothers reported they had participated in educational activities about breastfeeding, and 4% had received information exclusively about formula-feeding. The USAID/LACHNS report [16] found that elimination of infant formula from hospitals was a widespread practice, probably because breastfeeding is less time-consuming and produces substantial savings. In contrast, the implementation of other policies had less success. In two hospitals with full rooming-in, 40% to 60% of the infants were fed before the initiation of breastfeeding, most of them with oral glucose solutions. The funding for educational activities and counselling was low, and the intended educational messages were not conveyed efficiently.

A recent survey compared the performances of 14 public hospitals in Mexico City. Half were certified as Baby-Friendly Hospitals [18]. The evaluation was based on interviews with mothers on discharge from the hospital. The questionnaires explored the hospitals' policies related to the "Ten Steps to Successful Breastfeeding" [19]. Certified hospitals scored significantly better than noncertified hospitals in many areas: pre- and postpartum instructions on infant feeding, the numbers of infants suckled within the first 30 minutes of life, rooming-in, exclusive breastfeeding, non-use of feeding bottles, and availability of support groups. Encouragement from the health team to breastfeed was similar in both types of hospitals. Formula was prescribed for a negligible proportion of mothers. It must be underlined that despite the differences described, the non-certified hospitals have had a good level of compliance with the policies of the Baby-Friendly Hospital Initiative. The effect of rooming-in on the initiation of breastfeeding was compared in a cohort of 15,574 mother-infant pairs [20]. Sixty-eight per cent of the infants in the rooming-in programme breastfed within six hours after birth, and all were breastfeeding at discharge. None of the infants in the traditional wards did so. In spite of the shortcomings described, active breastfeeding programmes have a clear and positive impact on breastfeeding prevalence.

Another problem is the high proportion of Caesarean deliveries, which might interfere with the decision of women to breastfeed. A longitudinal study by our group [11] found that almost 30% of the women in a cohort delivered by Caesarean section. These figures are consistent with many other reports. Although women giving birth vaginally breastfed more frequently than those delivering by Caesarean section, breastfeeding rate differences vanished by 60 days post-partum, making it evident that rooming-in stimulates the initiation but does not lengthen the duration of breastfeeding.

Health professionals

There is a general notion that physicians and allied professionals in Mexico lack adequate skills to advise lactating women about initiating and sustaining lactation. Such a notion is based on the evidence that their knowledge of nutrition is very superficial [21]. Nutrition is not included in the curricula of medical schools as an individual discipline. The principles of infant feeding are taught in courses of paediatrics. The most popular textbooks of paediatrics have 3 to 6 pages of information on human lactation and 9 to 25 pages of information on artificial feeding, and comparable amounts of classroom time are devoted to these topics.

In a recent study more than 500 medical students and more than 170 residents from different Mexican universities answered a questionnaire exploring their knowledge of the basics of nutrition [22]. The critical grade for an appropriate level of knowledge was 4.3, which was higher than the average grade of the medical students and residents. In 1992 the National Centre for Maternal Lactation launched a programme to train health professionals in infant nutrition and breastfeeding. The Centre produced educational materials to support the future educational activities of those graduating from the Centre. There has been no published evaluation of this programme.

Individual beliefs

The multiple social factors that are important to the establishment of lactation vary among cultures. In Mexico regional differences often encompass significant cultural diversity. Such differences will be illustrated by data gathered from two independent cohorts of 346 urban and 216 rural mother infant pairs.

Communities studied

The rural setting is the Otomi Indian village of San Mateo Capulhuac north-west of Mexico City (n = 216). The 5,000 inhabitants live mostly from subsistence agriculture, with maize the main staple. There is little consumption of animal products. The diet of lactating women is marginal in energy and protein and low in fat. The level of physical activity is high [23].

The urban setting is represented by two neighbourhoods of Mexico City: Iztapalapa in the southeast (n = 170) and Tizapan in the south-west (n = 176). Both are densely populated by families of low-income workers living in small apartments. The diet of urban lactating women is closer to the Recommended Dietary Allowances [24] in energy and protein than that of rural women. Their lifestyle tends to be sedentary.

Besides the geographic differences, the two groups have significant differences in level of poverty, education, and accessibility to a paying job. There are also differences in sanitary facilities in the home. Although piped water was available in most rural households, it was not chlorinated, and fecal contamination was the norm (table 2).

In the rural community, 99% of the infants were predominantly breastfed in the first two weeks (other liquids could be provided except for non human milk). These high rates prevailed up to six months (table 3). None of the mothers failed to breastfeed their previous children, and all mothers reported being breastfed as infants. Formula was not commercially available in the village at the time of the study, although feeding bottles were present. The price of one can of formula in the nearest town was equivalent to one day's salary. The high prevalence of breastfeeding seems to have very strong cultural support, while formula-feeding was not practised for economic reasons. The breastfeeding rate was also high (93%) immediately after delivery in the urban group. It declined progressively in the post-partum period. By the third month fewer than threequarters of the women were providing any breastmilk to their infants, and only 40% of the mothers were predominantly breastfeeding. At the end of the sixth month, 52% of the mothers were feeding their infants any breastmilk, and only 25% were feeding breastmilk without additional nonhuman milk (table 3).

TABLE 2. Socio-economic characteristics of the communities

Characteristic Rural
(n = 216)
(n = 346)
Poverty level of the communitya 4 2
Household characteristics
piped water 93% 92%
flush toilet 6% 87%
dirt floor 64% 11%
refrigerator 5% 60%
crowdingb 91 % 80%
Maternal characteristics
education, median years (range) 3 (0-6) 9 (2-14)
living with partner 98% 87%
working outside home 2% 32%

a. 1 = mildest, 5 = worst (see ref. 24 for explanation of scale).
b. Three or more persons per room.


Discussion of paper by Villalpando et al.

Discussion of paper by Villalpando et al.


Professor Hanson

May I just comment on history? The first described case of artificial feeding was by one of the Assyrian kings. Also, in archaeological investigations of cultures along the Indus, many spoons and feeding utensils that were obviously used for artificial feeding have been found. So all through history this has been common. In Sweden in 70 and 150 A.D. there were areas where there was no breastfeeding at all; nobody knows why. And there was a law in Norway from 1040 that mothers should breastfeed, obviously, because of the fact that many didn't, so this is an old thing.

Dr. Prentice

I would like to explore a little more with you and Dr. Woolridge the issue of women feeling that they have insufficient milk. This seems to come out of every culture that we've heard about today, and, certainly, my experience in Africa is similar. Women will say that they are worried that they don't have enough milk. Where formula or other foods are available, they'll move on to either. What is it that women expect of a baby that makes them feel that they don't have enough milk? Dr. Woolridge showed that a lot of perceived milk insufficiency is found in women with milk volume in the normal range. From my own experience, having breastfed in the United Kingdom and in Africa, British women expect children to stay asleep for long periods at night. The family expects it even if the mother doesn't, which puts a lot of pressure on women to somehow make the child not feed at such frequent intervals. I just wondered if similar sorts of cultural problems would be underlying your women's responses as well?

Dr. Villalpando

I think the infant's behaviour often raises the question of whether milk production is sufficient. It is not the same in rural areas, where the infants probably cry as much as urban infants, but they don't have any other option.

Dr. Valdés

Our rural mothers may breastfeed 20 times a day. In the waiting room, they feed once, twice, or three times. When you ask them how many times they have breastfed today, they say maybe one, two, or five times, but they don't have this notion that taking the baby to the breast often is a bother. But for mothers from higher socio-economic groups, if the baby is not quiet for four hours, which was what they were used to when the babies were bottle-fed, they feel they don't have milk.

Dr. Woolridge

I think there is a total difference in outlook between a mother in the United Kingdom and a mother in Thailand. As we have just heard, the mothers in Thailand have no expectation of failing. They just assume it is the only option, and they will do it successfully. Intriguingly, they still report perceptions of insufficient milk, but it doesn't cause them to discontinue breastfeeding. It is a common cultural concern, even in that sort of society. I agree that in the United Kingdom there is an expectation that a baby should sleep for long stretches to fit in with the adult lifestyle. It's an inappropriate expectation, and much of it has been derived from formula-feeding.

Dr. McNellly

Mary Houston investigated this question of insufficient milk. She concluded that for many women it was a legitimate reason for giving up breastfeeding. If a woman says she doesn't have enough milk, she shouldn't carry on, because it's going to damage the child. Some women were certainly using that as a legitimate excuse to stop breastfeeding.

Dr. Villalpando

I think that is true for our urban group. As you could see, the individual who most often bought the first can of formula without any advice was the mother. She was really legitimizing her decision to stop breastfeeding by saying that she had insufficient milk production.

Dr. Van Eskrik

If I may comment on the Thai example, different methods produce very different pictures. I was doing ethnographic work in another part of Thailand, and the discussions about insufficient milk were quite different. Of course, they are not phrased that way, but I came across two things quite regularly, not in questioning but from observation. First, women were becoming concerned that their breasts were too small, that there was something wrong with their shape. There are pictures on calendars in every tiny hut, and a woman would be very strongly influenced to think about whether or not her breasts were adequate to this task. That's something that I think is changing quite rapidly. The second is a fundamental approach, strongly linked with Theravada Buddhism, that the capacity to nurture well is a result of merit stored from former lives. This means that women talk about themselves as being someone who is either a good nurser or not a good nurser, almost as if we would say that this is a characteristic that they would expect to see in themselves. It is also very interesting to think about how women classify themselves as both mothers and women. I know it doesn't fit with the kind of work you were doing, but at least it suggests perhaps another way to look at cultural factors.

Professor Howie

May I make a point about almost a war of communication that seems to be a common factor in developed and developing countries? I think one of your slides made the point how clever formula manufacturers are at getting their message across. In fact, they target health professionals even more than they target the patients themselves, and how successful that campaign is! We have seen literature distributed by the manufacturers who now have an obligation to state the advantages of breastfeeding. They give one page on the benefits of breastfeeding and maybe five pages about cracked nipples, tiredness, and breast abscess. In fact, the section that is supposed to encourage breastfeeding mothers is targeted so cleverly that it has exactly the opposite effect. What do you feel about making a systematic attempt to get better education that is not dominated by the formula manufacturers, so that some misunderstandings about breastfeeding can be overcome? We recently asked mothers who were breastfeeding if they knew that the baby would lose weight in the first few days after birth, and the number who knew this was about 5%. We are simply not getting basic information over to our mothers, and I suspect from what you are saying that there is a failure in Mexico, the United Kingdom, and many other countries.

Dr. Woolridge

There are two sides to this problem. One is to attack the problem of medical education by working with assistant professors of paediatrics who are lecturing to students, inviting them to seminars about lactation and so forth. This is a huge task. The other is the lack of a strong network of people promoting breastfeeding awareness to hospitals, physicians, and the general public to counteract the propaganda from formula manufacturers, because most literature handed to mothers is prepared by those companies.

Dr. Perez

What is more comfortable for a working mother three months after the child's birth, breastfeeding or bottlefeeding, and why?

Unidentified speaker

I would say breastfeeding. The situation, though, is that we put so many obstacles against mothers breastfeeding successfully that we have this notion that the mother and baby have to be changed to fit the circumstances of work and so forth. We shouldn't be changing human behaviour, we should be changing working conditions to fit breastfeeding.

Dr. Villalpando

Many years ago the labour laws in Mexico supported long post-partum leaves of absence intended to aid breastfeeding. However, fewer than half of the mothers breastfed for such a long period, so the outcome was the reverse of what many expected.

Dr. Rasmussen

Let me put it another way, that convenience is in the eye of the beholder. If you do a survey in the United States, uppereconomic-class women will tell you that breastfeeding is by far the most convenient, for all of the reasons we know. If you ask lower-economic-class women the same question, you don't get the same answer. Their lives are chaotic, they're on the bus, they're embarrassed to breastfeed on the bus. The last thing in the world they consider breastfeeding to be is convenient for them. The answer to Dr. Perez's question depends on a woman's life circumstances, and whether breastfeeding works for her. It's wonderful to have labour laws to say you're entitled to a maternity leave, and if you have a job or are protected by these laws, that may be wonderful in other ways. The United States is in the embarrassing position of not even having such laws, so I can hardly cast any stones. What about women who work in the informal sector for whom there is no protection whatsoever? Their labour is just as valuable to their families, the interference of their work conditions with breastfeeding may be as great or greater, yet they have no protection at all. In their case, convenience, again, may be an issue.

Dr. Valdés

We are researching that area. We have a study on working women, and one aspect we are concerned about is women's satisfaction with breastfeeding. We were surprised that the mothers in the control group who did not breastfeed were very satisfied, and the mothers who exclusively breastfed or human-milk-fed for six months, even while working, were equally satisfied.

Professor Hanson

I use the Swedish example, where it is rare for mothers not to breastfeed. They are entitled to 18 months off, most of that period with full pay, so it works.

Dr. Valdés

To go back to the issue of insufficient milk and mothers' perceptions, I think many women don't understand lactation well. They are encouraged to think that their milk will come in and their breasts will feel very full. But at about four to six weeks the breast physiologically adapts, so that they have less fullness and enlargement. Some women interpret the change in sensation to mean that they have less milk. They really need to understand that lactation is an adaptive process, and even though the breasts are quite soft, they still produce milk.

Professor Hanson

One issue I don't think is covered much in connection with the insufficient-milk syndrome is the contribution that the infant makes to the mother's perception about insufficient milk. We have to remember that the infant regulates its intake by many different aspects of its sucking behaviour. To a large extent, this is accomplished by differences in sucking amplitude or frequency. The infant has some degree of control over what actually takes place. We seldom look at the contribution the infant makes to the mother's decisions regarding feeding.

This also extends to another comment I want to make in connection with the determinants of milk production. There is an enormous variability in birthweight in different populations in the world. In Sweden, for example, the average is approximately 3,500 g. In Santa Maria Cauque, Guatemala, the average is approximately 2,600 g, a 900-g difference. There is likely a relation between an infant's birthweight and its sucking behaviour. The degree of breastfeeding success in different populations is also determined, in part, by other infant characteristics. Yet in general, there tends to be a greater emphasis placed on the mother as the one who will make the final decisions on what is going to happen, and not enough attention is given to the infant's roles.

Dr. Garza

Let me go back to the role the infant plays, because I think it is an important one. In studies we did about 10 years ago, we were interested in understanding whether infants were hungry whenever they demonstrated behaviours that the mother interpreted as hunger. We followed women longitudinally for four months beginning at about the third month postpartum. Women who were planning to breastfeed for 6 to 12 months were recruited. At the point they introduced solid foods, we asked why they were doing this. I think 100% of them said it was because the child was either fussy or not sleeping. They were interpreting this as evidence that the child was hungry and, therefore, that they were unable to sustain or increase milk production sufficiently to satisfy the infant.

The remarkable thing, however, was that when we measured the energy consumption of infants before and after the introduction of solid foods, we found no difference. Thus, I think children give certain cues that women interpret as hunger but that are, in fact, driven behaviourally by other needs. The child likely has reached the point where it wants to put something else in its mouth, and mothers tend to interpret those cues as hunger. Because the infant's behavioural response to other foods is often positive, their mistaken interpretations are reinforced. In fact, those responses appear to have nothing to do with physiological hunger defined in the usual terms. There are inappropriate periods in early infancy to introduce other foods. The physiological timing of complementary foods determined by behaviours other than hunger is dependent on a number of conditions that relate to the infant's surroundings.

Dr. Villalpando

Studies have investigated infant intake by 24-hour test weighings. Fewer than 50% of suckling episodes result in the intake of more than 50 g of milk. The remainder result in the intake of as little as 5 or 10 g of milk. This suggests that the infant is not really hungry. It's just sucking for the joy of sucking.

Unidentified speaker

I would like to return to the determinants of a woman's preference, because I cannot understand why generally in Latin America, fewer than 20% of women are fully breastfeeding at six months post-partum? We need to find an answer to that.

Unidentified speaker

I haven't heard anybody talk about the same woman with different babies in successive births. Is past behaviour predictive of future behaviour?

Dr. Victora

We did research on that. Some mothers who tended to breastfeed one child for a long time were likely to do the same with the next child. If mothers breastfed a very short time, the results were about the same in subsequent pregnancies. There was, however, a gray area when mothers breastfed from one to six or seven months. They were very unpredictable. One could not predict what would happen with the next baby.

I would like to go back to a point raised earlier. I think babies have very strong personalities. When we did our pacifier study, the main factor influencing pacifier use was the infant. Some babies just refuse them, and those are the babies who tend to breastfeed longest.

Dr. Prentice

To add a caveat to that, in our African experience, where 100% of women were breastfeeding for 18 months or so, women in successive lactations tended to show a characteristic breastmilk volume and breastmilk composition. Thus, it's a mixture of culture, maternal feelings, and the way a mother's breasts work.