The authors concluded that short-term energy supplementation was ineffective, probably because the women included in the study were not at nutritional risk. (Their reported energy intakes of 1,570 to 1,617 kcal/day were suggestive of undernutrition, but their body mass index at 4 weeks post partum averaged 19, which is considered to be at the low end of the normal range.) Mean milk intakes of the exclusively breastfed infants in this population (686 to 830 g/day) were very similar to those of infants in industrialized countries.

In theory, the energy stored as fat deposits during pregnancy is used to support milk production post partum, but there are few data with which to evaluate this relationship. Studies in animals have provided no consistent evidence of poor lactation performance resulting from poor prenatal nutrition alone (Kliewer and Rasmussen, 1987). Compared with controls, there was no statistically significant decrease in milk yield or litter weight at day 14 among rats fed 50% of ad libitum intakes before and during pregnancy but allowed to eat ad libitum during lactation. In cross-sectional studies of relatively well-nourished women in the United States, correlations between infant milk intake during the first 5 months post partum and maternal prepregnancy weight or pregnancy weight gain were not statistically significant (Butte et al., 1984a; Dewey et al., 1986). Although the range in prepregnancy weight for height was relatively wide (76 to 153% of desirable body weight) in these studies, very few of the women gained less than 11 kg (24 lb) during pregnancy.

In Indonesia, prepregnancy maternal body mass index was positively associated with milk intake of breastfed infants at 18 to 22 weeks post partum but not at 2 to 6, 10 to 14, or 26 to 30 weeks (controlling for infant birth weight did not change this association) (van Steenbergen et al., 1989). Few data have been reported in developing countries on the potential influence of pregnancy weight gain. Because maternal nutritional status (as indicated by weight for height) during pregnancy is strongly associated with nutritional status post partum, it is not possible to examine separately the influence of prenatal maternal body composition.

Mixed results have been obtained from observational studies of associations between indices of relative weight (such as body mass index) during lactation and milk volume in human populations. In industrialized countries, milk volume and maternal anthropometric variables have not been associated (Butte et al., 1984a; Dewey et al., 1986; Prentice et al., 1986). In developing countries, the situation is more complex. In an apparently undernourished population in Bangladesh, Brown et al. (1986b) found no statistically significant association between infant milk intake and maternal anthropometric variables, but they did report that milk energy output (the total energy value of the milk produced) was higher in women with greater arm circumference and triceps skinfold