Contents - Previous - Next

4. Results and discussion

This section will consist of a summary of the enormous mass of longitudinal data which has been obtained in the five centres. In very general terms, most of the data agree very well in the general implications. The most important of these implications would be that, in four of the five countries (Thailand excepted), it appears that energy intake in the diet of these pregnant women has increased by only minimal amounts - probably less than 100 kcal/d - although the mothers added the expected quantities of weight during pregnancy, including a store of adipose tissue, and the babies were born of normal weight, relative to the body mass of the mother.

These data are the result of studies done in the five countries over a period of several years on pregnant women living under very different circumstances in a variety of environmental situations.

The populations have been very diverse and perhaps some attention needs to be paid to a comparative analysis of their physical characteristics (Table 1). There are many apparent differences between the groups but, as a proportion of the initial body mass, many of these differences become much reduced. The environmental and family circumstances, and way of life might also confuse the interpretation of the data. For example, how representative are these groups of women of the wider populations in The Gambia, Thailand, and the Philippines? How important are environmental influences such as economic state, type of dwelling, number of children, productivity of the land, climate, whether or not the husband is working away from home for extended periods, and so on?

In the case of the Scottish and Dutch women, the problem is perhaps less complicated. But, especially in the case of the Gambian women, there may be considerable environmental and perhaps genetic stimuli exerting an effect on the results.

Table 1. Initial characteristics of the women (means and SD)


Study centre

Scotland 88

Holland 57

The Gambia 52

Thailand 44

Philippines 51

age (yrs)












height (m)






weight* (kg)






sum of 4 skin folds* (mm)






fat mass. (kg)






fat mass (as % body weight)






fat-free mass* (kg)






fat-free mass (as % body weight)






energy intake* ( kcal/d)






BMR. (kcal/d)






(kcal/kg body weight/d)






(kcal/kg fat free mass/d)






birth weight (g)






placental weight (9)






* Measured at or near 10 weeks gestation. Fat mass and fat-free mass have been estimated from the sum of 4 skinfolds and body weight.

The Gambian women seem to have been the beneficiaires of a remarkable physiological adjustment whereby the women unsupplemented with extra food, by becoming pregnant, actually save so much energy in basal metabolism that they end with a positive energy balance over the whole of pregnancy of about 11,000 kcal (46 MJ). Becoming pregnant, far from requiring extra energy, is a benefit to their state of energy balance (Figure 1).

The weight gain (Table 2) of the different groups suggests superficially that there is a slight difference between Scotland and Holland, and that there are appreciably lower weight gains in Thailand and the Philippines, with the lowest in The Gambia. However, if these weight gains are examined as a proportion of the initial weight of the mother, that for the Gambian women is still low, and that for the Dutch women is lower than others, but the Scottish, Thai and Philippine women all have virtually the same proportional weight gain. It is important to remember this, since the absolute values look dissimilar enough to convey the impression that we are dealing with two diametrically different situations in the developed and in the developing countries.

Figure 1. Incremental changes in BMR during pregnancy.

Net additional cost of BMR during pregnancy



30,1 00



The Gambia






* from 10 weeks until term
**from 13 weeks until term

Table 2. Weight and fat gain during pregnancy (mean SE)

Total weight gain (kg) (from 10 weeks to term)

% of initial weight

Fat gain (kg) *

% of initial weight


11.7 0.4


2.3 0.3



10.5 0.5


1.3 0.3


The Gambia

7.3 0.4


0.6 0.3



8.9 0.4


1.4 1.1



8.5 0.4


1.3 0.3


*Difference between maternal fat stores at 4-6 weeks post partum and at 10 weeks gestation, calculated from the the sum of 4 skinfolds and body weight using the equations of DURNIN and WOMERSLEY (1974).

There are also differences in the actual amounts of maternal fat laid down, where the implication alters when they are expressed as a proportion of the initial body weight.

It is interesting that the Wageningen women, who were bigger and slightly fatter than the Glasgow women, deposited only about two-thirds the amount of maternal fat in spite of having a higher energy intake and little difference in energy expenditure.

One could even raise the question whether all the groups were unphysiological in the amounts of fat they laid down during pregnancy (Table 2). None of them come near the Hytten and Leitch value of 3.5 kg, although that value might well be an unnecessarily large quantity.

The quantity of fat laid down in the body of a healthy pregnant woman has an energy equivalent which Hytten and Leitch postulated as being almost half of the total extra energy cost of pregnancy, i.e., about 40,000 kcal. The fact that none of the present groups laid down anything like this amount has considerable bearing on the calculation of energy cost. It would be interesting to know whether the varying quantities of fat deposited during pregnancy were correlated with such things as successful breast-feeding, the maternal body weight during the first 6 months post partum, perhaps even with the weight of the baby. Certainly, the small amount of increased maternal fat of the Gambian women appears to have had little influence on their ability to breast-feed their infants for many months.

There is also a problem with regard to the estimation of maternal fat gain during pregnancy. None of the present techniques is devoid of some difficulty in interpretation because of the changes in maternal body composition in pregnancy. The accuracy of the assessment of body fat from either skinfold thickness or densitometry is affected by the altered density of the fat-free mass in pregnancy. Simultaneous measurement of body water will reduce these errors a little, but the intrinsic error, both in the experimental measurement of body water, and in the assumptions implicit in the derivation of fat from body water, make it unlikely that it will improve matters very much.

The other large component of the total energy cost of pregnancy is the increase in Basal Metabolic Rate (Figure 1). Hytten and Leitch estimate this to be almost half the total, with a value of about 36,000 kcal (150 MJ). The measured increases in the Scottish and Dutch women were not very different from this - 30,100 kcal (126 MJ) and 34,500 kcal (144 MJ) respectively. The values given for Thailand and the Philippines are rather different and significantly less - 24,000 kcal (100 MJ) and 19,000 kcal (79 MJ) respectively - but these were estimates made from measurements taken at 10 and 13 weeks' gestation. From the data obtained in the other centres, it seems clear that the values at 10-13 weeks are not baseline values for prepregnancy; the Glasgow and Wageningen data at 10-13 weeks have already shown an increase of approximately 40-80 kcal/d (170-340 kJ), whereas the Gambian values show a decrease of about 40 kcal/d (170 kJ).

Whatever situation applies to Thailand and the Philippines, the real increase in BMR may be significantly different from the estimates given here. From the body build, body mass, way of life, and general nutritional status, it is probable that the BMR of the Thai and Philippine women would have increased by 10-13 weeks' gestation. There does not seem to have been enough biological stimulation to expect them to behave like the Gambian women and show a reduction in BMR. Their BMR might therefore have increased by larger amounts than those calculated, perhaps to around 24,000-29,000 kcal (100-120 MJ), on the assumption that the real prepregnant baseline was, say, 20 kcal/d (83 kJ) lower than that at 10-13 weeks, which would make a difference of about 5500 kcal (23 MJ) over the whole of pregnancy.

The total increase in BMR over the whole of pregnancy of the Gambian women differs by at least a factor of 10 from the data from the other centres. Not only is this an enormous difference, but it obviously demonstrates a quite remarkable physiological adaptation to, presumably, not pregnancy only, but to pregnancy in the face of some severe nutritional stress. The stress might, of course, no longer exist at the present time in quite such a severe form as it might have done in the past to result in, presumably, a genetically-endowed alteration in hormonal function. The investigation of this might be very fruitful, in a general biological sense. There are obvious reasons why this is a fascinating physiological problem which needs intensive investigation because of its fundamental importance. For example, the energy expenditure of the Gambian women was about 2400 kcal/d (10 MJ) in early pregnancy and in prepregnancy. Presumably, as a long-term state, their intake must usually be of the same order. With intakes of this quantity, in a group of physically well-built women, with a substantial fat-free mass, it is superficially unlikely that they would develop such extraordinary adaptations in metabolic rate when faced with pregnancy. Also, when given food supplementation, the adaptation seems to disappear almost immediately (which would perhaps seem to argue against a genetic factor). This is clearly a problem of basic intrinsic interest and widespread physiological importance.

Mechanical efficiency, particularly in locomotion but perhaps in a more general sense as well, has also been studied in some of its aspects in this multicentre project. Small differences have arisen in the results from the different centres, but they are not of much significance. The net energy cost of walking at a fixed rate on the treadmill, calculated as energy per unit of body weight, becomes slightly less in most of the various population groups, but the decrease is small in degree and never results in a diminished energy expenditure in absolute terms: even though a pregnant woman in her third trimester expends less energy in walking per kg of her body mass than she did in her first trimester, she still expends more energy as a total (Figure 2).

There are serious methodological problems in interpreting the data on patterns of physical activity and of energy expenditure. Those of us with wide experience in the technique used to measure energy expenditure in this project know of the variability which occurs in these measurements. The result of this variability can be illustrated by simple examples. "Sitting" is an "activity" which takes up a very large part of anyone's day - 7 to 8 hours or more as an average for these pregnant women. It is improbable that more than 3 or 4 measurements by indirect calorimetry will usually have been made on any individual, and this has been the case here, even in the Gambian study where most measurements were done. If the mean value of sitting in any individual was found to be say, 1.2 kcal/min (5 kJ), the SD will be about 0.1-0.2 kcal/min (0.4 to 0.8 kJ). Applying these values to 7-8 hours of sitting would mean that there is a 1 in 3 chance that, over the whole day, the total calculated energy of sitting could vary by plus or minus at least 90 kcal (380 kJ). For the total 40 weeks of pregnancy this comes to more than 25,000 kcal (105 MJ).

The time and energy spent in bed and in standing are subject to similar variations as for sitting, and there is no guarantee that these variations will cancel each other out; a systematic bias could easily be introduced.

Figure 2. The energy cost of walking on a treadmill at 3.8 kph (Scotland, Holland, The Gambia) or at 3 kph (Thailand) at various stages of pregnancy.

It seems at least possible therefore that, by its nature, the method of estimating energy expenditure by indirect calorimetry can not be of sufficient precision for present needs. It is informative in that it indicates the approximate energy expended in light or moderately heavy activities, it may give some information about possible trends in the energy expended in various tasks at different stages of pregnancy, it allows some comparison between different individuals, and it can signify trends in energy expenditure related to bodyweight changes. But it is not possible to make enough measurements of the energy expended in any one activity to know that the final average value makes proper allowance for variations due to time of day, the effect of meals - which themselves will be of different quantities and composition - previous physical activity, and naturally occurring variation. The final energy expenditure assessment, if done meticulously and extensively enough, will give information which can put an individual or a group into a general category, but these techniques are incapable of discriminating to levels of 100-150 kcal/d (420-630 kJ) which for current purposes leaves scope for important errors to arise. Perhaps the newer technique of measuring energy expenditure by the use of stable isotopes will remove some of these difficulties.

The energy cost of pregnancy, as calculated for the women studied in each of the five centres, is shown in Table 3. There are apparent differences in the data from the different localities, but in fact, if they are standardized for body weight, total energy costs come to about 60,000 kcal (250 MJ) for all the groups except the Gambian women, and the small differences between the other groups are the result mostly of the variable amounts of maternal fat stored during pregnancy. Based on very comprehensive measurements, there is thus a high degree of uniformity in the energy cost of pregnancy, except for the rather unusual Gambian group.

Table 3. Energy cost of pregnancy in different centres (in kcal)


















Uterus, breasts, blood vol., amniotic fluid, ECF, etc. )






Maternal fat


* 14300


* 15400














*Estimated from 10 weeks.
**Estimated from 13 weeks.

These costs, which may of course be even higher because they do not take into account the possibility that energy expenditure may be increased due to the greater body weight of the mother, must, then, be met by appropriate increases in energy intake.

However, energy intakes, with the exception of the Thai data, do not conform, even remotely, to the theoretically expected quantities. In the Glasgow, Wageningen, Philippine and almost certainly also the Gambian populations, energy intakes increased overall by very small quantities - by no more than 100 kcal/d (420 kJ), but perhaps by very much less than this even (Figure 3).

Figure 3. Incremental changes in energy intake during pregnancy.

On the other hand, the exception, the Thai women, are very different, and if an adjustment is made for their lower body weight compared to the "reference" woman with a body mass of 55-58 kg, the mean daily increase in energy intake of the Thai women was 250 kcal/d (1.0 MJ) above the baseline value.

There is no doubt that this result raises difficulties in relation to general recommendations of the energy requirements of pregnancy. Are the Thai women measured in this study more representative of women in the developing countries than the Gambian women or the Philippine women ? They appear reasonably nourished women with moderate amounts of body fat and not subject to any particular nutritional stress throughout pregnancy. From the data, their workload does not seem importantly different from that of the Philippine women. Nothing about them appears fundamentally different from all other groups except their food intake. If the data are representative of Thai women in general, the biological implications are interesting in that there seems to be a complete absence of any form of adaptation to the energy needs of pregnancy, clearly demonstrated in all the other groups. Further investigation is necessary.

Contents - Previous - Next