Contents - Previous - Next



Energy requirements of pregnancy have either been estimated by summing up energy costs (tissues of the conceptus + tissue accretion in the mother + incremental maintenance costs) or inferred from energy intakes of pregnant women with successful outcome of pregnancy. Frequently the estimates obtained by the first method are substantially higher than the ones obtained by the second one. A large part of the discussion was devoted to attempts at explaining this difference.

There is a wide range in the proportion of the energy cost of pregnancy that can be spared by the mother if she substantially reduces her energy expenditure for physical activity. A considerable reduction of energy expenditure can result from relatively small changes in types and intensities of activities, which may be quite difficult to observe and record, if they all consistently go in the same direction (Durnin). In many instances this seems to actually happen and to be the main explanation for the discrepancy between energy cost estimates obtained by the two different methods. In such populations it appears reasonable to take this into account in calculations of requirements, but it should not be assumed that all pregnant women can substantially reduce their physical activity. The full cost of pregnancy should therefore remain the basis for recommendations and reductions only made where actually applicable.

Energy requirements of non-pregnant, non-lactating individuals are now generally expressed as BMR × PAL, the latter reflecting the habitual or desirable activity level, and there seems to be good agreement in the group that applying the same system to pregnant and lactating women would be a good way of accounting explicitly for differences in energy requirements for physical activity.

Earlier estimates assumed the deposition of 3 kg of fat in the mother. In their paper, Prentice et al argue for a reduction of this amount by 20%. Butte emphasized the great variability in weight gain during pregnancy and is not convinced that we have enough solid evidence to reject the earlier estimate. There is some discussion on the efficiency of fat deposition. The prevailing view seems to be that it is high since most of the fat that is deposited is dietary fat and liponeogenesis plays only a minor role (Millward & Reeds). Prentice remarked that the situation be different in countries like The Gambia, where only 16-28% of the dietary energy come from fat. If obligatory fat oxidation exceeds fat intakes and liponeogenesis becomes more important (which has not been conclusively shown yet), more energy would be needed for fat deposition. This energy would not be included in BMR, which is measured postprandially (Millward).

An intriguing phenomenon is the wide variation in the maintenance cost of pregnancy. Women with a high fat mass and BMI at conception show large increases in BMR from the beginning of pregnancy (Allen), whereas in three studies (one of them involving whole body calorimetry) the BMR of very lean women has been observed to decrease in the initial phases of pregnancy (Prentice). The physiological mechanisms underlying this variability are not clear yet, but the phenomenon raises the question whether prepregnancy BMIs or fat mass should be taken into consideration in the calculation of energy requirements. The Subcommittee on Nutritional Status and Weight Gain During Pregnancy of the National Academy of Sciences in the US (1990) recommended different weight gains during pregnancy for women with different prepregnancy BMI. A positive relationship between energy intake and weight gain during pregnancy has been observed in underweight, but not in normal-weight and overweight American women (Prentice). The idea of prescribing different energy intakes for women with different prepregnancy BMIs was considered but finally rejected because of lack of enough information on the subject.

The validity of assessments of energy intake was discussed at considerable length. Observed energy intake increments are usually considerably smaller than expected, also in women in industrialized countries, who can be assumed to eat to appetite. Since these increments tend to be smaller in longitudinal than in cross-sectional studies, measurement fatigue and underreporting were considered possible, but the evidence and the arguments presented in the discussion did not go beyond what is contained in the position paper of Prentice et al.

Increments in energy intakes and BMR throughout pregnancy are represented by an integral under a curve, and the size of this surface depends enormously on baseline values. Butte doubted if baseline data were always obtained from non-pregnant, non-lactating women.

Waterlow expressed disappointment about the fact that the report did not attempt to relate energy requirements to pregnancy outcome in terms of birthweight. Women in developing countries are usually able to maintain their body weight or even increase it through many pregnancies, whereas relatively large percentages of infants are born with a low birthweight.


The incremental energy costs of lactation can be calculated as breast milk volume × energy density × conversion efficiency and modified by changes in maternal fat stores and activity. As in pregnancy, calculated costs of lactation are often considerably higher than the observed increments in energy intakes. Most of the difference must be attributable to a mobilization of maternal fat stores and a reduction in physical activity, but the group agreed with the authors of the position paper that, as in pregnancy, the full cost of lactation should form the basis for recommendations, i.e., loss of maternal fat and reduced physical activity of the mother should not be considered obligatory, but only taken into account where appropriate.

Milk production is remarkably robust (unaffected by BMI down to about 17 (Prentice et al, 1994) and by infection of the mother (Zavaleta et al, 1995)) and mainly driven by demand: milk donors, wet-nurses and mothers of more than one infant can increase their milk output substantially, whereas an increased energy intake does not necessarily result in a higher milk output.

The energy content of breast milk has usually been calculated by applying modified Atwater factors to proximate constituents of milk samples. This represents the metabolisable energy available to the child. The mother must supply gross energy which is about 5% higher. The fat content in the milk is highly variable and representative sampling therefore difficult. The group agrees with an estimate of about 2.8 kJ/g. Various approaches suggest that 80% is a reasonable estimate of conversion efficiency, and calculations show that changes in this assumption have relatively small effects; assuming an efficiency of 90 instead of 80% would make a difference of only about 200 kJ (Reeds).

Incremental energy intakes during lactation vary widely around an average of about 1.5 MJ. In Mexico (Allen) and The Gambia (Prentice) they are relatively high, and women gain weight during lactation; in other countries (e.g. Scotland) they are lower, probably because mothers try to lose some of the fat deposited during pregnancy.


Prentice AM, Goldberg GR, Prentice A (1994): Body mass index and lactation performance. Eur. J. Clin. Nutr. 48 (suppl 3), S78-S89,

Subcommittee on Nutritional Status and Weight Gain During Pregnancy (1990): Nutrition during pregnancy. Washington DC: National Academy Press.

Zavaleta N. Lanata CF, Butron B. Peerson J. Brown KH, Lönnerdal B (1995): Effect of acute infection on quantity and composition of breast milk. Am. J. Clin. Nutr. 62, 559-563.

Contents - Previous - Next