Contents - Previous - Next

Basic physiology of lactation

Before discussing the possible relationships between BMI and lactation performance it is necessary to establish an understanding of some of the underlying factors which influence human milk production.

Firstly, although nutritionists apply the label 'at risk' to lactating women, the marginal energetic stress of human lactation is actually among the lowest measured for any mammalian species (Prentice & Whitehead, 1987). This results from the very slow growth rate of the human infant. The incremental energy needs for a woman breastfeeding a single infant amount to about 25% of her non-lactating requirement, whereas for a rat dam feeding a litter of eight pups or a ewe feeding two fast-growing lambs the incremental needs may be as high as 300%. Furthermore, even at moderately low BMI, women tend to have larger fat stores (expressed as a proportion of body weight) than many other species. The combination of these factors provides a greater flexibility in meeting the costs of milk synthesis, and helps to explain why human lactation is less sensitive to maternal nutritional status than might be expected.

Table 2. Maternal and infant factors influencing lactation performance

Controlling factors

Baby's size

birth weight



Baby's appetite


Maternal characteristics




Suckling behaviour


weaning foods

fluid intake

Limiting factors

Secretory capacity of milk cells

Substrate supply to breasts

Secondly, it is important to make the distinction between factors which control and those which may limit lactation performance (see Table 2). It is necessary for the mother-infant pair to maintain a delicate balance between milk supply and demand. This is achieved through sensitive feedback mechanisms which control milk flow. In most circumstances it is these controlling factors, involving both, maternal and infant variables, which dominate. The most important of these is the infant's current size which tends to yield the highest correlation with milk intake; on average larger babies demand and receive more milk (Prentice et al., 1986). The best illustration of this effect is the fact that, even in the developing world, mothers who breastfeed twins tend to produce about twice as much milk as those who feed singletons (Prentice et al., 1986). Since infant weight is correlated with birth weight, and since both are correlated with maternal weight, there are a number of confounding influences which complicate the interpretation of possible relationships between maternal BMI and lactation performance. These should be borne in mind throughout this paper.

A further complication is illustrated in Table 3. It lists correlations between milk volume and maternal anthropometry at three age ranges from our own Gambian studies (Prentice et al., 1986) and shows that significant, albeit weak, associations between indices of maternal body size (weight and height) can disappear when the two positive associations are included as the numerator and denominator in calculating BMI since they cancel each other out. This problem should also be borne in mind and may be pertinent with reference to other papers within this Supplement.

Table 3. Correlations between maternal anthropometry and milk volume in rural Gambian women

Stage of lactation (months)























Data from Prentice et al. (1986). *P < 0.05, **P < 0.01.

Is BMI a useful discriminator of lactation performance in terms of breast-milk quantity?

Based on an earlier analysis of the available world literature we have demonstrated elsewhere that women in developing countries appear to produce very similar amounts of milk to those in affluent countries (Prentice et al., 1986). This immediately suggests a lack of any clear association between maternal nutritional status and milk production. However, such an analysis requires refinement in the present context since many groups of women in the developing world have quite high BMIs, and the small amount of data on women with very low BMIs might be swamped.

Figure 1 shows a scatter plot of milk volume at peak lactation against maternal BMI for all of the studies listed in Table 1. The data from Zaire stand out as being much lower than from all other populations. This could be due to a number of factors. The first relates to methodology since many of the measurements appear to have been performed in a metabolic ward setting to which the mothers were transported for 24 h observation during which their normal demand-feeding was replaced by a timed feeding schedule. This disruption of the usual habit and lifestyle, and the stress of unusual surroundings, can have a profound inhibitory effect on milk output which may have been particularly severe for rural mothers. Alternatively it has been suggested that a primary protein deficiency may be the cause of impaired lactation in these women (Hennart & Vis, 1980). For the purposes of the present analysis the true explanation is somewhat irrelevant since the mean BMI is in any case rather high, and since within the Zaire data there is no association between maternal BMI and milk volume in either the urban or rural groups even in BMI categories <18.5 (see Fig. 2). On these two counts it seems reasonable to exclude the Zaire data points from Fig. 1. The resulting Fig. 3 reveals that there is no perceptible association between BMI and the volume of milk even at BMI <18.5 (suggested as a possible cut-off prior to this meeting). Even in this range milk volume is remarkably high, particularly when it is considered that such women have a lower body weight and height than the groups with higher BMIs (see Table 1), and since we have shown above that maternal size can have an effect on milk volume which may be independent of BMI.

Fig. 2. Milk volume plotted according to maternal BMI in urban and rural women in Zaire. Data from WHO Collaborative Study (1985). Error bars = SE. - Urban

Fig. 2. Milk volume plotted according to maternal BMI in urban and rural women in Zaire. Data from WHO Collaborative Study (1985). Error bars = SE. - Rural

Fig. 1. Relationship between average milk volume and average maternal BMI in published studies of lactation performance. Data-sets are listed in Table 1.

Fig. 3. Figure I redrawn after exclusion of data from Zaire.

A few of the studies included in Fig. 3 have investigated possible associations between BMI (or reasonable proxies for BMI) and milk volume within the populations being studied.

Figure 4 illustrates data from Kenya divided according to maternal weight-for-height (WH) into WH-plus and WH-minus groups (Steenbergen et al., 1983). Milk volume was slightly lower in the WH-minus group (BMI = 18.7) than in the WH-plus group (BMI = 23.5) for most stages of lactation. Analysis of covariance demonstrated that the average deficit of 80 g/day was significant at P = 0.028, but infant weight was not included as a covariable. Separate correlation analysis showed infant weight to be important and it may be confounding the observation of a difference. In any case the milk volumes of both groups were high and well within the range observed in affluent populations.

Fig. 4. Milk volume at different stages of lactation in 'undernourished' and 'adequately nourished' Kenyan women. Women were divided into WH-minus (weight-for-height <90%) or WH-plus (weight-for-height >90%). Data from Steenbergen et al. (1983). Error bars = SE.

Fig. 5. Milk volume plotted according to maternal BMI in women from Myanmar. Data from Khin-Maung-Naing et al. (1980). Error bars = SE.

Figure 5 shows data from the Myanmar study (Khin-Maung-Naing et al., 1980). In this study there was a significant positive relationship between BMI and milk volume, but it was caused by an enhanced yield at high BMIs rather than by a reduced milk yield at low BMIs. Furthermore all of the groups produced remarkably high milk volumes given the small size of the mothers. Even the lowest BMI subgroup (<16.8) had an average milk yield of 767 g/day.

Fig. 6. Milk volume plotted according to maternal BMI and maternal weight gain during lactation in women from Bangladesh. Data from Brown et al. (1986). Error bars = SE.

Fig. 7. Comparison of peak milk volume in women from Bangladesh and USA. Data from Brown et al. (1986) and Nommsen et al. (1991). Error bars = SE.

Figure 6 shows data from the Bangladesh study of Brown et al. (1986) divided both by BMI in early lactation and by the rate of maternal weight gain in lactation. It demonstrates that those mothers who managed to gain weight during lactation, perhaps indicating a greater current energy intake, produced a greater volume of milk. However, BMI did not significantly discriminate between high and low producers in either weight gain category although there was a tendency towards a higher milk yield in the higher BMI mothers within the low weight gain group. Once again the remarkable feature of the Bangladesh data is the high milk yields of all subgroups. This is emphasized in Fig. 7 which compares their average milk volumes with those reported from a group of well-nourished American women (Nommsen et al., 1991) and shows that they are virtually identical at peak lactation.

Table 4. Databases with significant correlations between maternal fatness and breast-milk fat






All parities




Parity >4





3 months p.p.




6 months p.p.




9 months p.p.




12 months p.p.





4 + 6 months p.p.





0-9 months p.p.

Significant positive associations

Data from references Prentice et al. (1981), Nommsen et al. (1991), Villalpando et al. (1991) and Brown et al. (1986).
p.p. = post partum.

Both the inter- and intra-population analyses presented above force us to the conclusion that, at the levels of nutritional status studied so far, there is no detectable relationship between maternal BMI and milk volume. We must add the caveat that, of course, it is inevitable that lactation performance will eventually be compromised if BMI falls sufficiently far. However, it appears from the data available at present that this threshold must be at a very low BMI.

Is BMI a useful discriminator of lactation performance in terms of breast-milk quality?

Although it is difficult to refer to a single publication which summarizes the pertinent data there is a wide consensus that breast-milk quality is largely unrelated to maternal nutritional status (and hence BMI) with two possible exceptions. The first is that micronutrient content may be reduced particularly in the case of some water soluble vitamins, for which milk levels seem to parallel maternal plasma levels, and for certain minerals such as calcium. However, this is not an issue which is directly related to BMI, and associations would only occur when low BMI was acting as a proxy for general undernutrition. The second exception is milk fat (and hence milk energy) for which there are a number of reports of associations with maternal BMI.

Table 4 lists studies in which positive associations have been found between measures of maternal fatness (but not BMI specifically) and milk fat in The Gambia (Prentice, Prentice & Whitehead, 1981), Mexico (Villalpando et al., 1991; Butte et al., 1992), Bangladesh (Brown et al., 1986) and USA (Nommsen et al., 1991). The correlations are not very strong but they nonetheless appear convincing. In particular the results from Bangladesh shown in Fig. 8 appear to be indisputable, and although they are not related directly to BMI it is highly likely that BMI would be closely correlated with arm circumference and triceps skinfold implying that low BMI would be associated with low milk fat and energy content.

Fig. 8. Breast-milk fat and energy content plotted according to maternal arm circumference and triceps skinfold thickness in women from Bangladesh. Data from Brown et al. (1986). Values as g/day were derived by multiplying concentration by daily milk volume. Error bars = SE.

There are, however, other studies which fail to show any relationship between milk fat and BMI, and there are others which actually indicate a negative association. Fig. 9 shows milk fat according to BMI in the data from Myanmar (Khin-Muang-Naing et al., 1980) which is particularly valuable insofar as the database contains some women with very low BMIs. There is no perceptible association between BMI and milk fat even at the lowest BMI category. Fig. 10 illustrates the two cases of reverse associations. The first is from Brazil (Marie Spring et al., 1985) and the second from

Fig. 9. Breast-milk fat concentration plotted according to maternal BMI in women from Myanmar Data from Khin-Maung-Naing et al. (1980). Error bars = SE.

Kenya (Steenbergen et al., 1983). Once again the analysis was not performed directly against BMI. In Brazil the malnourished women were classified as those <90% weight-for-height. In Kenya it can be calculated that the BMIs in the WH-plus and WH-minus groups were 18.7 and 23.5 respectively. Such paradoxical associations are difficult to explain and could be related to sampling procedures which are notoriously difficult to design because of the large changes in fat concentration within feeds and throughout the day.

Fig. 10. Examples of inverse associations between maternal nutritional status and breast-milk fat. Data from Marin Spring et al. (1985) and Steenbergen et al. (1983).

Fig. 11. Relationship between average milk energy content and average maternal BMI in published studies of lactation performance. Data-sets are listed in Table 1.

These sampling difficulties make it somewhat dangerous to make comparisons across different studies so the cross-country analysis of milk energy content derived from Table 1 and plotted in Fig. 11 should be interpreted with some caution. In spite of this it appears to reveal a similar message to the analogous scatterplot for milk volume in Fig. 3. There seems to be no evidence that milk energy content is compromised even at BMIs <18.5.


The foregoing analysis of the available world literature indicates that lactation performance is surprisingly independent of maternal BMI, and that BMI would not therefore provide a useful index of functional impairment in the breastfeeding mother as judged by the quantity or quality of milk production. Of course, there could be other functional consequences for a thin mother who has to meet the additional stress of lactation. Loss of body weight might be anticipated as she subsidises her daily energy budget from her own adipose tissue stores. However, there is little evidence of substantial weight loss in thin lactating women (Prentice & Prentice, 1990) except in particular circumstances such as the Gambian hungry season during which all adults lose weight. In fact the data are once again remarkable for the lack of weight loss [cf. Bangladesh (Brown et al., 1986), Gambia (Prentice et al., 1986) and Myanmar (Khin-Muang-Naing et al., 1980) studies cited above]. The only logical conclusion that can be derived from these findings is that, often contrary to appearances, these thin women must be consuming sufficient energy on a day-to-day basis to support their milk production. Under conditions where daily intake does not match requirements a fatter mother would be expected to be able to support lactation for longer, and BMI might then emerge as a useful indicator of risk. Such conditions have not yet been studied, and this suggests that they are rather extreme and may occur only under famine or near famine conditions.


Brown KH, Robertson AD, Akhtar NA & Ahmed MG (1986): Lactational capacity of marginally nourished mothers: relationships between maternal nutritional status and quantity and proximate composition of milk. Pediatrics 78, 909-919.

Butte NF, Wong WW, Patterson BW, Garza C & Klein PD (1988): Human milk intake measured by administration of deuterium oxide to the mother: a comparison with the test-weighing technique. Am. J. Clin. Nutr 47, 815-821.

Butte NF, Wong WW, Klein PD & Garza C (1991): Measurement of milk intake: tracer to infant deuterium dilution method. Br J. Nutr 65, 3-14.

Butte NF, Villalpando S. Wong WW, Flores-Huerta S. Hernandez-Beltran M de J. O' Brian Smith E & Garza C (1992): Human milk intake and growth faltering of rural Mesoamerindian infants. Am. J. Clin. Nutr 55, 1109-1116.

Frigerio C, Schutz Y. Prentice AM, Whitehead RG & Jequier E (1991): Is human lactation a particularly efficient process? Eur J. Clin. Nutr 45, 459-462.

Goldberg GR, Prentice AM, Coward WA, Davies HL, Murgatroyd PR, Sawyer MB, Ashford J & Black AE (1991): Longitudinal assessment of the components of energy balance in well-nourished lactating women. Am. J. Clin. Nutr. 54, 788-798.

Hennart P & Vis HL (1980): Breast-feeding and post partum amenorrhea in Central Africa. I Milk production in rural areas. J. Trop. Pediatr. 26, 177-183.

Hennart P. Ruchababisha M & Vis HL (1983): Breast-feeding and post partum amenorrhea in Central Africa. 3 Milk production in an urban area. J. Trop. Pediatr. 29, 185-189.

Hennart PF, Brasseur DJ, Delogne-Desnoeck JB, Dramaix MM & Robyn CE (1991): Lysozyme, lactoferrin, and secretory immunoglobulin A content in breast milk: influence of duration of lactation, nutrition status, prolactin status, and parity of mother. Am. J. Clin. Nutr. 53, 32-39.

Infante C, Lara W & Vio F (1985): Isotope dilution measurement of breast-milk production in Chilean urban mothers. Hum. Nutr.: Clin. Nutr. 39C, 379-386.

Khin-Muang-Naing, Tin-Tin-Oo, Kywe-Thein & New-New-Hlaing (1980): Study on lactation performance of Burmese mothers. Am. J. Clin. Nutr 33, 2665-2668.

Lovelady CA, Lonnerdal B & Dewey KG (1990): Lactation performance of exercising women. Am. J. Clin. Nutr 52, 103-109.

Madhavapeddi R & Rao BSN (1992): Energy balance in lactating undernourished Indian women. Eur J. Clin. Nutr 46, 349-354.

Marin Spring PC, Amancio OMS, Nobriga F. Araujo G. Koppel SM & Dodge JA (1985): Fat and energy content of breast milk of malnourished and well nourished women, Brazil 1982. Ann. Trop. Paediatr 5, 83-87.

Motil KJ, Montandon CM, Hachey DL, Boutton TW, Klein PD & Garza C (1989): Relationships among lactation performance, maternal diet, and body protein metabolism in humans. Eur J. Clin. Nutr 43, 681-691.

Nommsen LA, Lovelady CA, Heinig MJ, Lonnerdal B & Dewey KG (1991): Determinants of energy, protein, lipid and lactose concentrations in human milk during the first 12 months of lactation; the DARLING Study. Am. J. Clin. Nutr 53, 457-465.

Orr-Ewing AK, Heywood PF & Coward WA (1986): Longitudinal measurements of breast milk output by a 2H2O tracer technique in rural Papua New Guinean women. Hum. Nutr.: Clin. Nutr 40C, 451-467.

Prentice AM & Prentice A (1990): maternal energy requirements to support lactation. In Breastfeeding, nutrition, infection and infant growth in developed and emerging countries, eds SA Atkinson, LA Hanson & RK Chandra, pp. 67-86. St John's, Newfoundland: ARTS Biomedical Publishers and Distributors.

Prentice AM & Whitehead RG (1987): The energetics of human reproduction. In Reproductive energetics in mammals, eds ASI Loudon & PA Racey, Symp. Zool. Soc. Lond. 57, 275-304.

Prentice A, Prentice AM & Whitehead RG (1981): Breast-milk fat concentrations of rural African women. 2. Long-term variations within a community. Br J. Nutr 45, 495-503.

Prentice AM, Paul AA, Prentice A, Black AK, Cole TJ & Whitehead RG (1986): Cross-cultural differences in lactational performance. In Human lactation 2, eds M Hamosh & AS Goldman, pp. 13-44. New York: Plenum Publishing Corp.

Raaij JMA van, Schonk CM, Vermaat-Miedema SH, Peek MEM & Hautvast JGA (1991): Energy cost of lactation, and energy balances of well-nourished Dutch lactating women; reappraisal of the extra energy requirements of lactation. Am. J. Clin. Nutr 53, 612-619.

Sadurskis A, Kabir N. Wager J & Forsum E (1988): Energy metabolism, body composition, and milk production in healthy Swedish women during lactation. Am. J. Clin. Nutr 4X, 44-49.

Steenbergen WM van, Kusin JA, de With C, Lacko E & Jansen AAJ (1983): Lactation performance of mothers with contrasting nutritional status in rural Kenya. Acta Paediatr Scand 72, 805-810.

Steenbergen WM van, Kusin JA, Kardjati S & de With C (1989): Energy supplementation in the last trimester of pregnancy in East Java, Indonesia: effect on breast-milk output. Am. J. Clin. Nutr 50, 274-279.

Villalpando S. Butte NF, Wong WW, Flores-Huerta S. Hernandez-Beltran M de J. O'Brian Smith E & Garza C (1991): Lactation performance of rural Mesoamerindians. Eur J. Clin. Nutr 46, 337-348.

WHO (1985): Report on the WHO Collaborative Study on Breast-feeding: the quantity and quality of breast-milk. Geneva: WHO.

Contents - Previous - Next