Discussion

Atwater factors indicate the average amount of energy yielded by one gram of ingested carbohydrate, fat or protein; they are used in the calculation of the metabolizable energy content of foods, for instance in food composition tables and in infant formulas. Atwater (as well as Durnin and Southgate after him) derived them from the heat of combustion, corrected for energy losses in the form of unabsorbed nutrients in feces and urine of adults. The question was raised whether the same factors were also applicable to infants. The answer to this question does not affect energy requirements per se but becomes important in a discussion of recommended dietary intakes. Several factors may influence the metabolizable energy derived from food: (1) the chemical form of the macronutrient in the food, (2) the coefficient of digestibility; (3) the extent to which the nutrients are not completely oxidized, but stored in the body; (4) gut maturation and (5) age. In growing infants nitrogen retention will be higher. Preterm infants absorb less fat than term infants, and fat is generally less well absorbed by newborn infants than by older infants. Fat digestibility is also highly dependent upon the fat source and its processing, e.g. butterfat is poorly absorbed, whereas a mixture of vegetable oils is absorbed nearly to the same extent as human milk. In a study of 10 breast-fed infants fed unpasteurized milk, Southgate found that metabolizable energy averaged 92%. Application of the Atwater factors to human milk components indicated 96% metabolizable energy. Using Atwater factors in normal infants, therefore, does not seem to entail great errors. Application of the Atwater factors in preterm or sick infants may overestimate energy availability.

In young infants the energy content of human milk is of particular importance. Since it is very variable throughout days and feeds and there is no generally agreed upon, standard method for obtaining representative milk samples and for estimating their energy density, published figures vary considerably. Butte et al, using different methods, obtained values between 0.65 and 0.67, whereas values from Sweden (0.72) and a WHO study in Hungary are considerably higher (Waterlow). In the first two figures of her paper, Butte used energy intakes as reported. Dewey pointed out that differences in fat secretion in breast milk between groups of women had been observed, even when exactly the same methods were used. Maternal body fat can affect milk fat (Prentice), as can fat intake in lean women (Dewey). Since pasteurization alters the fat, it is important to note whether pasteurized or non pasteurized milk is used. In the end, the prevailing opinion was that Dewey and Butte had made the most rigorous assessments and that their values should therefore be relied upon primarily.

Several participants were intrigued by the low level of the first two data points in the line representing energy requirements derived from TEE and growth in Butte's figures 7 and 8. Most likely this is an artifact due to an underestimate of the cost of growth in these first two time periods.

Should recommendations be the same or different for breast- and bottle-fed infants? Reeds argued that requirements and intakes should not be confused. Requirements are to be seen as a function of the organism and not of the diet, whereas recommended dietary allowances are a function of the diet and the degree to which it meets requirements. Dewey pointed out that in practice the picture was less clear and the feeding mode seemed to affect physiology. Energy expenditure is lower in breast-fed infants or, in other words, formula-fed infants appear to require more energy than breast-fed ones. These differences are most marked between 3 and 6 months of age; then they gradually disappear, probably as a consequence of the phasing out of pure breast-feeding. Butte tried to derive energy requirements from data of a mixed group of infants, 50% breast- and 50% formula-fed. Dewey advocated separate recommendations for the two feeding groups in order to avoid the :impression that breast-fed infants do not get enough energy and ought to be supplemented or the risk that formula fed infants will not get enough energy to cover their needs. Giving a wide range of requirements does not appear to be a satisfactory solution either.

Butte et al tried to determine how much of a difference in diet-induced thermogenesis (DIT) there was between breast- and formula-fed infants. During the first 4h after the meal, DIT appeared slightly lower in breast-fed infants, but the difference was not statistically significant.

Waterlow queried the validity of 42% for the energetic efficiency of protein synthesis (Table 5, footnote d), and suggested that a figure of 75% would be more in accordance with the evidence.

Do infants growing up in the more stressful environment of developing countries or urban slums have the same or higher energy requirements than infants in industrialized countries? The little information that exists on this issue shows smaller differences than expected. Total energy expenditure (TEE), expressed as kcal/kg, was for instance very similar in infants from The Gambia and the UK (Prentice). Butte compared TEE of small groups (n = 20) of 4-month-old infants from Mexico and Houston. In. Mexico it was 74 kcal/kg, in Houston 64 and 73 kcal/kg for breast- and bottle-fed infants, respectively. Several participants felt that more information was needed to decide the extent to which frequent infections and desirable catch-up growth add to energy requirements in poor environments.