686 TRANSACTIONS OP I=
ROYAL SOCIETY OF TROPIWU. MEDICINE AND HYGIBNE, VOL. 73, No. 6, 1979
The quantitative
effects of maternal dietary energy intake on pregnancy and lactation in rural Gambian women
A. A. PAUL, E. M. MULLER* AND R. G. WHITEHEAD Medical Research Council Dunn Nutrition Unit, Milton Road, Cambridge, UK and Keneba, The Gambia
Summary The quantitative relationships between dietary energy intake and weight gain in pregnancy, birthweight and lactation performance during the first three months of infancy have been studied in such a way as to take account of major differences in the patterns of heavy manual labour at different times of the year in a subsistence farming community. Maternal weight gain and the accumulation of subcutaneous fat were significantly lower when the last trimester of pregnancy fell during the time of heaviest farm work and lowest energy intakes. The birth-weight of babies was also significantly correlated with differences in energy intake throughout the year. During early lactation breast milk yields were significantly related to concomitant alterations in the subcutaneous fat stores. Evidence has been produced which suggests that in undernourished nursing women there could be a competition for dietary energy between the repleting maternal subcutaneous fat organs and the mammary glands at the expense of milk production. Introduction Growth faltering in early childhood is widesoread in the develooina world. and The Gambia is’ no exception (WHI%EH~AD et al., 1977). Although the prevalence of severe clinical malnutrition is greatest in the second and third years of life, it is becoming increasingly obvious that the onset of the problem is much earlier, during infancy and perhaps even before the child is born (WATERLOW & RUTISHAUSER, 1974; WHITEHEAD et al., 1978). Whilst it is widely assumed that the food intake of a woman during pregnancy and lactation can have a marked effect on the baby’s birth-weight and on her capacity for lactation, there have been few quantitative studies to define the precise relation between maternal diet and these parameters. In The Gambia and in many other countries there is a major complicating factor. Women are an important part of the agricultural labour force, and they are not excused from heavy farming activities even during late pregnancy and lactation. A preliminary renort (WHITEHEAD et al., 1978) has already demonstrated-that maternal energy intakes in The Gambia are generally low, but there is a marked seasonal variation in which values fall to less than 50% of predicted energy requirements for moderately active pregnant and lactating women (FAO/WHO, 1973) during the rainy season. This period coincides with the heaviest farming activities so the women
are doubly subjected to energy deficit. During this traditional hungry period even pregnant women have been shown to lose weight (THOMSON et al., 1966). This study, which was carried out in 1977, concentrated on women during the last trimester of pregnancy and the first three months of lactation. The purpose was to define more exactly the relationship between food intake, maternal nutritional and anthropometric status, the baby’s birth-weight and the mother’s subsequent lactation performance, in relation to changes in the pattern of farming activities. The first three months post partum are of critical importance, as the relative-lactation capacity established at this time is maintained throughout the later stages of infancy (WHITEHEAD et al., 1978). Materials and Methods Keneba Village The village of Keneba is a relatively isolated subsistence farming community situated about 100 miles inland from the coast. It is surrounded by guinea Savannah and is about one mile from a tributary of the River Gambia. The Muslim community, with a population of almost 1,000, has been described in detail by MCGREGOR (1976), THOMPSON (1965), THOMSON et al. (1968) and TULLY (1978). There is a monomodal rainy season usually lasting from mid-June to mid-November, with most rain falling between July and September. Average annual rainfall is about 900 mm but it was exceptionally low, 580 mm, during 1977. Farming and foods The farming activities extend over the rainy season with field clearing starting in May and harvesting finally finishing about the end of December. The main crops are rice and millet. Rice, which is almost entirely grown by the women, is planted in the bush and then transferred to swamps by the river. Sorghum, maize and findo (Digitarum sp.) are also grown and groundnuts are the chief cash crop. In addition small amounts of fish are obtained over the year. Mangoes are in fairly plentiful supply from April to June, and leaves are collected from the bush during the rainy season and are made into sauces when groundnuts are scarce. Subjects and measurements From January to December 1977, 29 women were studied longitudinally from the seventh *Present address: Switzerland.
Wartstr.
5,
8032
Zurich,
A. A. PAUL
month of pregnancy to the third month of lactation. Their basic anthropometric characteristics are given in Table I. The mothers ranged in age between 16 and 40 years and the older women had had up to 10 children. Food intake
The food intake of the women was measured on four to six days each month, by weighing all meals eaten. These meals usually consisted of boiled or steamed cereal eaten with a groundnut sauce. Snack foods such as raw groundnuts or mangoes eaten between meals were recorded by a quantitastandardized recall procedure. Trained tively Gambian nutrition assistants were present at each meal; normally the women and children would eat from a common bowl but for this study the women were asked to put their own portion of food into a separate bowl or calabash. In the farming season the midday meal was weighed after it was cooked at home and before being taken out to the fields. Nutrition assistants visited the mothers in the fields to determine the exact amount of the food eaten there. During Ramadan food is not eaten during the hours of daylight and it was therefore necessary for food intake to be measured as early as 4 a.m. and up to 10 p.m. There are no food composition tables appropriate for cooked Gambian foods. Thus 10 to 100 samples of each of the 40 most frequently eaten foods were collected and, after freeze drying, were sent to Cambridge for analysis. Gross energy (GE) was determined using a ballistic bomb calorimeter (MILLER & PAYNE, 1959), total nitrogen (N) by a semi-automated micro-Kjeldahl procedure (Technicon Ltd.) and unavailable carbohydrate (UC) by the Southgate procedure (SOUTHGATE, 1969). For each food, metabolizable energy was calculated using the relationshipME-GE ,,0.977-6.6 N-4UC (SOUTHGATE, 1975). On selected samples of food where a more complete analysis was carried out there were no appreciable differences between values obtained with this formula and metabolizable energy derived from separate protein, fat and available carbohydrate measurements using conventional energy conversion factors. Breast
milk
output
Breast milk output was measured over a 12-hr daylight period (7 a.m.-7 p.m.) at three months of age, by a test weighing procedure as previously described (WHITEHEAD et al.. 1978). The mothers followed their usual activities at home or in the fields as far as possible during the measurement periods. Full 24-hr measurements performed on other women in the village have shown a close correlation with 12-hr measurements (WHITEHEAD et al., 1978) and that the latter amount to 45 to 55 “A of the 24-hr total (Paul & Prentice, unpublished). Anthropometric
measurements
At two-weekly intervals the pregnant women were weighed to the nearest 0.05 kg, wearing light clothes; triceps and subscapular skinfolds were measured using Harpenden calipers. Skinfold
687
et al.
thickness measurements were carefully standardized and were made by one observer (E.M.M.) throughout the study. The same measurements were made on lactating women at six weeks and three months after delivery. Most of the infants studied were born in the village and birth-weights were obtained within 24 hr of birth. Activity
study
A minute by minute activity diary was kept for 24 of the women over one or more four-hour periods of the day, between 7 a.m. and 11 p.m. in the eighth month of pregnancy or second month post partum. Each woman was accompanied by the same observer (E.M.M.) throughout. Results Table I shows that the average Keneba mother is not particularly small by international standards. During the second month of lactation in 1977 the average woman studied was 97’:; and 96’;;, of the Boston 50th centiles for height and weight respectively of young 18-year-old-US women (NELSON, 1971). Her most important deficit was in triceps skinfold thickness which, on average, was only 62”; of the standard (JELLIFFE, 1966). Table II shows the variation in energy and protein intakes bv month during 1977. There was a striking variation-in both energyand protein intake, values being particularly low from July to September, the peak of the farming season. In August intakes were only 1325 kcal/d (5.54 MJ) for pregnant and 1184 (4.95 MJ) for lactating women. There was a tendency for lactating women to eat more than pregnant women, but only when food was plentiful. The seasonal differences in energy intake were a direct reflection of availabilitv of foods dictated bv the farming calendar. Throughout the year over 90 “(, of the dietary energy was supplied by cereals and groundnuts but there were variations in the relative contributions of the two components. In August groundnuts supplied only 13y; of the energy, but in November this rose to 35’:;. It was this increase in groundnut consumption that accounted for most of the seasonal variation in energy intake. One complicating factor was the Ramadan fast. In 1977 Ramadan fell between mid-August and mid-September, but the low energy intakes recorded during this time were not primarily due to the daytime fasting as the usual midday meal was replaced by a pre-dawn breakfast and the women when possible ate more in the evening as well. The energy intakes in the first two weeks of August were thus about the same as in the second half of August, during Ramadan. Table III shows the activitv nattern of the lactating women. The proportions of time recorded for farming refer to work itself, and time spent walking to work; rest periods in the fields have been included with home rest. In fact the women actually spent long hours away from home, from early in the morning till dusk, and often had to walk five miles or m&e to reach the fields, carrying heavy loads. The period of heaviest farm work occurred in June and July, when the fields were
688
EFFECTS OF DIETARY
Table I-Anthropometric those collected during tions plus ranges
INTAKE
ON PREGNANCY
characteristics the second month
AND LACTATION
of the 29 Gambian of lactation. Values
IN THE GAMBIA
women studied. Measurements are are means and their standard devia-
Mean Age, years Body weight, kg Height, cm Triceps skinfold, mm Subscapular skinfold, mm
Table II-Mean first trimester
Range
26h6.8 51.6f5.6 156.6f5.0 10.3 14.1 10*0&3*4
16-40 45*4- 64.9 147.6-165.8 6*4- 21.6 6*0- 14.8
daily energy and protein intakes by women in the last trimester of pregnancy of lactation during 1977. Values are means and their standard errors Pregnant women Energy intake (kcal)
Calendar month 1977
n
Feb Mar Apr May Dee Heavy farm work
Jun Jul
Aug Sep Light farm work Ott Nov
Lactating
Protein intake (9)
mean
se
mean
6 7 8 8 4
1471 1530 1580 1545 1495
202 115 100 137 171
Dry season 46.8 5.0 47.7 3.1 52.5 3.2 45.7 4.3 50.1 6.0
7 9 7 8
1652 1374 1325 1391
104 112 128 73
9 6
1611 1562
95 121
March-April
n
Protein intake (9)
mean
se
mean
1825 2042 2129
191 304 193
55.3 62.1 72.1
7.3 10.2 8.7
Rainy season 46.1 2.5 43.6 3.9 41.6 5.0 41.6 2.4
6 2107 8 1559 7 1184 7 1316
191 148 148 120
62.9 48.7 34.8 39.6
7.6 5.1 4.2 4.3
53.8 53.7
6 6
167 111
52.4 68.4
5.9 3.8
Table III-Time (% of 24 hr) spent performing women in the second month of lactation Non-farming
women
Energy intake (kcal)
se
and
various
4 5 4
3.5 3.9
activities
1603 1939
at different
season
Farming
times
se
-
of 1977, by
season
May
June- July
Aw
Sept
Nov
Resting and light activities
66.5
74.1
57.9
65.5
66.7
65.1
Home and village activities
33.5
24.7
18.5
23.0
15.2
28.0
1.2
23.6
11.5
18.1
6.9
Farming (excluding
resting)
being hoed in preparation
for planting.
0
There was a of rice in September. It must be emphasized, however, that during 1977 the women’s harvesting load was less than normal because of the low rainfall and the almost total failure of the rice crop. The pregnant women could only be studied from February to July, but their activity pattern during this period
resurge in farm work during the transplanting
was essentially similar to that of the lactating women. As there were such differences in both food intake and heavy farming activities according to the time of the year, the women for whom the majority of their last trimester of pregnancy or first trimester of lactation fell between July and September have been grouped separately (in Table IV) from those
689
A. A. PAUL et al.
0
0 0
:
4I
' 1000
1250
.
I 1500
Energy intake
1750
1 zoo0
kcal/day
Fig. 1. Relationbetweenbirth-weightand energyintakeduring the lasttrimester of pregnancy.0 Heavyfarmingperiod; 0 Remainder of year (r = 0.50, P< 0.05). who reached these critical periods at a less arduous time. In the heavy farming period the pregnant women gained very little weight and even lost an average of 1 .O kg during August. There was also a quite abnormal reduction in triceps skinfold thickness; the small loss at other times was not surprising as this is a normal finding for the last trimester of pregnancy even in well-nourished women (TAGGART et al., 1967). Fig. 1 shows the relationship between the level of maternal energy intake during the last trimester of pregnancy and the weight of the baby at birth. This significant correlation (r = O-50, P < 0.05) demonstrates that full-term children born below 2.5 kg were associated with average maternal energy intakes of less than 1600 kcal/d (6 * 69 MJ). MALETNLEMA & BAW (1974) in Tanzania have also published a similar correlation. Although the children in this particular sub-sample born during the heavy farming period showed no difference in birth-weight from those born in the remainder of the year, an analysis of birth data covering all village infants in 1976 and 1977 has shown that those born during the rains were significantly smaller (WHITEHEAD et al., 1978). The women whose first trimester of lactation fell mainly between July and September also lost weight (Table IV), whilst surprisingly those studied during the remainder of the year actually gained
690
EFFECTS
OF DIETARY
INTAKE
ON
PREGNANCY
+3
AND
LACTATION
+2
f f .P D B 2 .5
+1
E 2”
THE
GAMBIA
600 I
I % D
IN
’
O
-1 o
-2
0
. a
-3
. 0
' 1000
I
I
1500
2000
I
2500
t
3000
Energ” intake. kcawday Fig. 2. Relation between change in body-weight and energy intake over the second six weeks of lactation. 0 Heavy farming period; 0 Remainder of year (I =0.57, P< 0.01).
0
weight, in spite of the fact that they were actively secreting considerable amounts of milk. Their skinfold thicknesses followed a similar pattern. In babies of this age there was no significant difference in milk intake between the two groups, but this was expected as we have already shown that seasonality tends to have its major effect on lactation capacity at the later stages of infancy (WHITEHEAD et uZ., 1978). Lactation performance was, however, still influenced by the mother’s over-all nutritional status and energy balance as can be seen in Figs. 2 and 3. Fig. 2 shows that over the year as a whole there was a significant positive relationship between changes in body-weight during lactation and dietary energy intake (r = 0.57, P < 0.01). Even at levels of energy intake as low as 1700 kcal/d some of the dietary energy was beginning to be utilized for the replenishment of maternal body stores; a similar relationship was observed with triceps skinfold thickness (r = 0.49, P < 0.05). From these surprising findings it can be deduced that if during pregnancy women have been unable to store fat for subsequent use in milk production, as discussed by FAOWHO (1973), there can be an actual competition betweenmaternal body replenishment and the synthesis of milk. This concept is supported by Fig. 3, which illustrates a significant negative correlation (r = -0.69, P < 0 .Ol) between breast milk output at three months and changes in skinfold thickness during the second six weeks of lactation. In women who were actively replenishing their subcutaneous fat stores, milk output was low. During the farming season virtually all the women were producing considerable amounts of milk at the expense of their subcutaneous fat stores.
Z l-
1
4
1
1
-2
0
Skinfold
change
+2
mm/6
+4
1
+6
weeks
Fig. 3. Relation between 12-hr breast milk output at 3 months and changes in the sum of triceps plus subscapular skinfold thickness l Heavy farming period; over the second six weeks of lactation. 0 Remainder of vear (r = -0.69, P < 0.01).
Discussion The results have shown that in spite of the multiplicity of other environmental constraints on the mother, including infections such as malaria, maternal energy balance has a quantifiable effect both on weight gain during pregnancy and on the weight of the baby at birth. Importantly, however, the precise relationship between food intake and the progress of pregnancy depends not only on the energy content of the food but also on the level of activity of the mother, and this is always something that must be taken into account when the adequacy of the maternal diet to the successful outcome of pregnancy is being considered (WHO, 1965). Our findings also provide strong circumstantial evidence that dietary intake during pregnancy is also of crucial importance for lactation capacity because of the building up of subcutaneous fat stores. In tropical paediatric circles much emphasis is now being given to the philosophy that it is better to feed the mother and hence the child, via breast milk, rather than to concentrate just on the provision of milk products for the supplementary feeding of the young infant. Unfortunately most people’s
A. A. PAUL et al.
attention has concentrated only on the lactation period itself and, by and large, the results of interventions involving the supplementary feeding of nursing mothers have not led to dramatic improvements in milk production (HOLEMANS et al., 1954; GOPALAN, 1958; CHAVEZ et al., 1975). In many developing countries, such as The Gambia, where lactation is universal until the infant is 18 to 24 months of age, women under 45 are pregnant or lactating virtually all the time. If mothers are to improve their milk output then we must become responsible for the nutritional well-being of women during the whole of their reproductive lives. Levels of energy intake as low as 1500 kcal during pregnancy and little more during lactation are so much below the recommended dietary intakes that incomplete measurements might be suspected. Indeed, THOMSON et al. (1966) estimated on the basis of assumed levels of energy expenditure that energy intakes in the Keneba women must have been about 3,000 kcal/day. In the present study, however, there was scrupulous care taken to avoid all obvious sources of error, such as snack foods. Subsequent work in the same village with much larger numbers of women has confirmed the essential features of the present findings and we are confident that the low food intake values presented here are a true representation of what women really are eating. It is relevant to point out that similar intakes have been found in other carefully carried out quantitative studies of food intake, for example NORGAN et al. (1974) reported intakes of 1400 kcal (5.85 MJ) in pregnant and lactating women in the coastal zone of New Guinea; MALETNLEMA & BAVU (1974) found intakes of 1850 kcal (7.72 MJ) in pregnant women in Tanzania; and GEBRE-MEDHIN & GOBEZIE (1975) reported intakes of 1540 kcal (6.44 MJ) in non-privileged Ethiopians during pregnancy. While such levels of energy intake are clearly inadequate, doubts must also be expressed about the need for the very much higher energy intakes currently recommended. We hope that the more extensive studies at present being carried out by us will produce data which will help us predict more realistic dietary requirements for conditions prevailing in countries such as The Gambia. Acknowledgements We would like to thank the project leader, Dr. M. G. M. Rowland, Sister M. Tully in whose care the women were, M. Hutton for the breast milk measurements, the Gambian nutrition assistants and the villagers of Keneba for their constant and cheerful co-operation. We are grateful to G. J. Hudson, P. M. V. John and B. Bailey for analysis of the foods, K. C. Day for the computer data analysis and T. J. Cole for statistical advice. E. M. M. was supuorted by ‘Co-ooeration Techniaue Suisse’ and‘ ‘Stiftung - zur Forderung der Ernahrungsforschung in der Schweiz’. References Chavez, A., Martinez, C. & Bourges, H. (1975). Role of lactation in the nutrition of low socioeconomic groups. Ecology of Food and Nutrition, 4, 159-169.
691
FAOjWHO (1973). Energy and Protein Requirements. World Health Organization Technical Report Series No. 52%Jood and Agriculture Organization Nutrition Meetings Report Series No. 52. Geneva: World Health Organization. Gebre-Medhin, M. & Gobezie, A. (1975). Dietary intake in the third trimester of pregnancy and birthweight of offspring among non-privileged and privileged women. American Journal of Clinical Nutrition, 28, 1322-1329. Gopalan, C. (1958). Studies on lactation in poor Indian communities. Journal of Tropical Paediatrics and Environmental Child Health, 4, 87-97. Holemans, K., Lambrechts, A. & Martin, H. (1954). Etude qualitative et quantitative du lait de femmes indigenes du Kwango (Congo belge). Revue Medicale de Liege, 9, 714-719. Jelliffe, D. B. (1966). The assessment of nutritional status in the‘community. World Health Organization Monotrraah Series No. 53. Geneva: World Health Organization. Maletnlema, T. N. & Bavu, J. L. (1974). Nutrition studies in pregnancy. Part I. Energy, protein and iron intakes of pregnant women in Kasarawe, Tanzania. East African Medical Journal, 51, 515-528. McGregor, I. A. (1976). Health and communicable disease in a rural African environment. Oikos, 27, 180-192. Miller, D. S. & Payne, l?. R. (1959). A ballistic bomb calorimeter. British Journal of Nutrition, 13, 501-508. Nelson, W. E. (1971). Textbook of Paediatrics. (9th edit.). Philadelphia: Saunders. Norgan, N. G., Ferro-Luzzi, A. & Durnin, J. V. G. A. (1974). The energy and nutrient intake and energy expenditure of 204 New Guinean adults. PhilisophLal Transactions of the Royal Society of London. Series B. 268. 309-348. Southnate,
ROYAL SOCIETY OF TROPIWU. MEDICINE AND HYGIBNE, VOL. 73, No. 6, 1979
The quantitative
effects of maternal dietary energy intake on pregnancy and lactation in rural Gambian women
A. A. PAUL, E. M. MULLER* AND R. G. WHITEHEAD Medical Research Council Dunn Nutrition Unit, Milton Road, Cambridge, UK and Keneba, The Gambia
Summary The quantitative relationships between dietary energy intake and weight gain in pregnancy, birthweight and lactation performance during the first three months of infancy have been studied in such a way as to take account of major differences in the patterns of heavy manual labour at different times of the year in a subsistence farming community. Maternal weight gain and the accumulation of subcutaneous fat were significantly lower when the last trimester of pregnancy fell during the time of heaviest farm work and lowest energy intakes. The birth-weight of babies was also significantly correlated with differences in energy intake throughout the year. During early lactation breast milk yields were significantly related to concomitant alterations in the subcutaneous fat stores. Evidence has been produced which suggests that in undernourished nursing women there could be a competition for dietary energy between the repleting maternal subcutaneous fat organs and the mammary glands at the expense of milk production. Introduction Growth faltering in early childhood is widesoread in the develooina world. and The Gambia is’ no exception (WHI%EH~AD et al., 1977). Although the prevalence of severe clinical malnutrition is greatest in the second and third years of life, it is becoming increasingly obvious that the onset of the problem is much earlier, during infancy and perhaps even before the child is born (WATERLOW & RUTISHAUSER, 1974; WHITEHEAD et al., 1978). Whilst it is widely assumed that the food intake of a woman during pregnancy and lactation can have a marked effect on the baby’s birth-weight and on her capacity for lactation, there have been few quantitative studies to define the precise relation between maternal diet and these parameters. In The Gambia and in many other countries there is a major complicating factor. Women are an important part of the agricultural labour force, and they are not excused from heavy farming activities even during late pregnancy and lactation. A preliminary renort (WHITEHEAD et al., 1978) has already demonstrated-that maternal energy intakes in The Gambia are generally low, but there is a marked seasonal variation in which values fall to less than 50% of predicted energy requirements for moderately active pregnant and lactating women (FAO/WHO, 1973) during the rainy season. This period coincides with the heaviest farming activities so the women
are doubly subjected to energy deficit. During this traditional hungry period even pregnant women have been shown to lose weight (THOMSON et al., 1966). This study, which was carried out in 1977, concentrated on women during the last trimester of pregnancy and the first three months of lactation. The purpose was to define more exactly the relationship between food intake, maternal nutritional and anthropometric status, the baby’s birth-weight and the mother’s subsequent lactation performance, in relation to changes in the pattern of farming activities. The first three months post partum are of critical importance, as the relative-lactation capacity established at this time is maintained throughout the later stages of infancy (WHITEHEAD et al., 1978). Materials and Methods Keneba Village The village of Keneba is a relatively isolated subsistence farming community situated about 100 miles inland from the coast. It is surrounded by guinea Savannah and is about one mile from a tributary of the River Gambia. The Muslim community, with a population of almost 1,000, has been described in detail by MCGREGOR (1976), THOMPSON (1965), THOMSON et al. (1968) and TULLY (1978). There is a monomodal rainy season usually lasting from mid-June to mid-November, with most rain falling between July and September. Average annual rainfall is about 900 mm but it was exceptionally low, 580 mm, during 1977. Farming and foods The farming activities extend over the rainy season with field clearing starting in May and harvesting finally finishing about the end of December. The main crops are rice and millet. Rice, which is almost entirely grown by the women, is planted in the bush and then transferred to swamps by the river. Sorghum, maize and findo (Digitarum sp.) are also grown and groundnuts are the chief cash crop. In addition small amounts of fish are obtained over the year. Mangoes are in fairly plentiful supply from April to June, and leaves are collected from the bush during the rainy season and are made into sauces when groundnuts are scarce. Subjects and measurements From January to December 1977, 29 women were studied longitudinally from the seventh *Present address: Switzerland.
Wartstr.
5,
8032
Zurich,
A. A. PAUL
month of pregnancy to the third month of lactation. Their basic anthropometric characteristics are given in Table I. The mothers ranged in age between 16 and 40 years and the older women had had up to 10 children. Food intake
The food intake of the women was measured on four to six days each month, by weighing all meals eaten. These meals usually consisted of boiled or steamed cereal eaten with a groundnut sauce. Snack foods such as raw groundnuts or mangoes eaten between meals were recorded by a quantitastandardized recall procedure. Trained tively Gambian nutrition assistants were present at each meal; normally the women and children would eat from a common bowl but for this study the women were asked to put their own portion of food into a separate bowl or calabash. In the farming season the midday meal was weighed after it was cooked at home and before being taken out to the fields. Nutrition assistants visited the mothers in the fields to determine the exact amount of the food eaten there. During Ramadan food is not eaten during the hours of daylight and it was therefore necessary for food intake to be measured as early as 4 a.m. and up to 10 p.m. There are no food composition tables appropriate for cooked Gambian foods. Thus 10 to 100 samples of each of the 40 most frequently eaten foods were collected and, after freeze drying, were sent to Cambridge for analysis. Gross energy (GE) was determined using a ballistic bomb calorimeter (MILLER & PAYNE, 1959), total nitrogen (N) by a semi-automated micro-Kjeldahl procedure (Technicon Ltd.) and unavailable carbohydrate (UC) by the Southgate procedure (SOUTHGATE, 1969). For each food, metabolizable energy was calculated using the relationshipME-GE ,,0.977-6.6 N-4UC (SOUTHGATE, 1975). On selected samples of food where a more complete analysis was carried out there were no appreciable differences between values obtained with this formula and metabolizable energy derived from separate protein, fat and available carbohydrate measurements using conventional energy conversion factors. Breast
milk
output
Breast milk output was measured over a 12-hr daylight period (7 a.m.-7 p.m.) at three months of age, by a test weighing procedure as previously described (WHITEHEAD et al.. 1978). The mothers followed their usual activities at home or in the fields as far as possible during the measurement periods. Full 24-hr measurements performed on other women in the village have shown a close correlation with 12-hr measurements (WHITEHEAD et al., 1978) and that the latter amount to 45 to 55 “A of the 24-hr total (Paul & Prentice, unpublished). Anthropometric
measurements
At two-weekly intervals the pregnant women were weighed to the nearest 0.05 kg, wearing light clothes; triceps and subscapular skinfolds were measured using Harpenden calipers. Skinfold
687
et al.
thickness measurements were carefully standardized and were made by one observer (E.M.M.) throughout the study. The same measurements were made on lactating women at six weeks and three months after delivery. Most of the infants studied were born in the village and birth-weights were obtained within 24 hr of birth. Activity
study
A minute by minute activity diary was kept for 24 of the women over one or more four-hour periods of the day, between 7 a.m. and 11 p.m. in the eighth month of pregnancy or second month post partum. Each woman was accompanied by the same observer (E.M.M.) throughout. Results Table I shows that the average Keneba mother is not particularly small by international standards. During the second month of lactation in 1977 the average woman studied was 97’:; and 96’;;, of the Boston 50th centiles for height and weight respectively of young 18-year-old-US women (NELSON, 1971). Her most important deficit was in triceps skinfold thickness which, on average, was only 62”; of the standard (JELLIFFE, 1966). Table II shows the variation in energy and protein intakes bv month during 1977. There was a striking variation-in both energyand protein intake, values being particularly low from July to September, the peak of the farming season. In August intakes were only 1325 kcal/d (5.54 MJ) for pregnant and 1184 (4.95 MJ) for lactating women. There was a tendency for lactating women to eat more than pregnant women, but only when food was plentiful. The seasonal differences in energy intake were a direct reflection of availabilitv of foods dictated bv the farming calendar. Throughout the year over 90 “(, of the dietary energy was supplied by cereals and groundnuts but there were variations in the relative contributions of the two components. In August groundnuts supplied only 13y; of the energy, but in November this rose to 35’:;. It was this increase in groundnut consumption that accounted for most of the seasonal variation in energy intake. One complicating factor was the Ramadan fast. In 1977 Ramadan fell between mid-August and mid-September, but the low energy intakes recorded during this time were not primarily due to the daytime fasting as the usual midday meal was replaced by a pre-dawn breakfast and the women when possible ate more in the evening as well. The energy intakes in the first two weeks of August were thus about the same as in the second half of August, during Ramadan. Table III shows the activitv nattern of the lactating women. The proportions of time recorded for farming refer to work itself, and time spent walking to work; rest periods in the fields have been included with home rest. In fact the women actually spent long hours away from home, from early in the morning till dusk, and often had to walk five miles or m&e to reach the fields, carrying heavy loads. The period of heaviest farm work occurred in June and July, when the fields were
688
EFFECTS OF DIETARY
Table I-Anthropometric those collected during tions plus ranges
INTAKE
ON PREGNANCY
characteristics the second month
AND LACTATION
of the 29 Gambian of lactation. Values
IN THE GAMBIA
women studied. Measurements are are means and their standard devia-
Mean Age, years Body weight, kg Height, cm Triceps skinfold, mm Subscapular skinfold, mm
Table II-Mean first trimester
Range
26h6.8 51.6f5.6 156.6f5.0 10.3 14.1 10*0&3*4
16-40 45*4- 64.9 147.6-165.8 6*4- 21.6 6*0- 14.8
daily energy and protein intakes by women in the last trimester of pregnancy of lactation during 1977. Values are means and their standard errors Pregnant women Energy intake (kcal)
Calendar month 1977
n
Feb Mar Apr May Dee Heavy farm work
Jun Jul
Aug Sep Light farm work Ott Nov
Lactating
Protein intake (9)
mean
se
mean
6 7 8 8 4
1471 1530 1580 1545 1495
202 115 100 137 171
Dry season 46.8 5.0 47.7 3.1 52.5 3.2 45.7 4.3 50.1 6.0
7 9 7 8
1652 1374 1325 1391
104 112 128 73
9 6
1611 1562
95 121
March-April
n
Protein intake (9)
mean
se
mean
1825 2042 2129
191 304 193
55.3 62.1 72.1
7.3 10.2 8.7
Rainy season 46.1 2.5 43.6 3.9 41.6 5.0 41.6 2.4
6 2107 8 1559 7 1184 7 1316
191 148 148 120
62.9 48.7 34.8 39.6
7.6 5.1 4.2 4.3
53.8 53.7
6 6
167 111
52.4 68.4
5.9 3.8
Table III-Time (% of 24 hr) spent performing women in the second month of lactation Non-farming
women
Energy intake (kcal)
se
and
various
4 5 4
3.5 3.9
activities
1603 1939
at different
season
Farming
times
se
-
of 1977, by
season
May
June- July
Aw
Sept
Nov
Resting and light activities
66.5
74.1
57.9
65.5
66.7
65.1
Home and village activities
33.5
24.7
18.5
23.0
15.2
28.0
1.2
23.6
11.5
18.1
6.9
Farming (excluding
resting)
being hoed in preparation
for planting.
0
There was a of rice in September. It must be emphasized, however, that during 1977 the women’s harvesting load was less than normal because of the low rainfall and the almost total failure of the rice crop. The pregnant women could only be studied from February to July, but their activity pattern during this period
resurge in farm work during the transplanting
was essentially similar to that of the lactating women. As there were such differences in both food intake and heavy farming activities according to the time of the year, the women for whom the majority of their last trimester of pregnancy or first trimester of lactation fell between July and September have been grouped separately (in Table IV) from those
689
A. A. PAUL et al.
0
0 0
:
4I
' 1000
1250
.
I 1500
Energy intake
1750
1 zoo0
kcal/day
Fig. 1. Relationbetweenbirth-weightand energyintakeduring the lasttrimester of pregnancy.0 Heavyfarmingperiod; 0 Remainder of year (r = 0.50, P< 0.05). who reached these critical periods at a less arduous time. In the heavy farming period the pregnant women gained very little weight and even lost an average of 1 .O kg during August. There was also a quite abnormal reduction in triceps skinfold thickness; the small loss at other times was not surprising as this is a normal finding for the last trimester of pregnancy even in well-nourished women (TAGGART et al., 1967). Fig. 1 shows the relationship between the level of maternal energy intake during the last trimester of pregnancy and the weight of the baby at birth. This significant correlation (r = O-50, P < 0.05) demonstrates that full-term children born below 2.5 kg were associated with average maternal energy intakes of less than 1600 kcal/d (6 * 69 MJ). MALETNLEMA & BAW (1974) in Tanzania have also published a similar correlation. Although the children in this particular sub-sample born during the heavy farming period showed no difference in birth-weight from those born in the remainder of the year, an analysis of birth data covering all village infants in 1976 and 1977 has shown that those born during the rains were significantly smaller (WHITEHEAD et al., 1978). The women whose first trimester of lactation fell mainly between July and September also lost weight (Table IV), whilst surprisingly those studied during the remainder of the year actually gained
690
EFFECTS
OF DIETARY
INTAKE
ON
PREGNANCY
+3
AND
LACTATION
+2
f f .P D B 2 .5
+1
E 2”
THE
GAMBIA
600 I
I % D
IN
’
O
-1 o
-2
0
. a
-3
. 0
' 1000
I
I
1500
2000
I
2500
t
3000
Energ” intake. kcawday Fig. 2. Relation between change in body-weight and energy intake over the second six weeks of lactation. 0 Heavy farming period; 0 Remainder of year (I =0.57, P< 0.01).
0
weight, in spite of the fact that they were actively secreting considerable amounts of milk. Their skinfold thicknesses followed a similar pattern. In babies of this age there was no significant difference in milk intake between the two groups, but this was expected as we have already shown that seasonality tends to have its major effect on lactation capacity at the later stages of infancy (WHITEHEAD et uZ., 1978). Lactation performance was, however, still influenced by the mother’s over-all nutritional status and energy balance as can be seen in Figs. 2 and 3. Fig. 2 shows that over the year as a whole there was a significant positive relationship between changes in body-weight during lactation and dietary energy intake (r = 0.57, P < 0.01). Even at levels of energy intake as low as 1700 kcal/d some of the dietary energy was beginning to be utilized for the replenishment of maternal body stores; a similar relationship was observed with triceps skinfold thickness (r = 0.49, P < 0.05). From these surprising findings it can be deduced that if during pregnancy women have been unable to store fat for subsequent use in milk production, as discussed by FAOWHO (1973), there can be an actual competition betweenmaternal body replenishment and the synthesis of milk. This concept is supported by Fig. 3, which illustrates a significant negative correlation (r = -0.69, P < 0 .Ol) between breast milk output at three months and changes in skinfold thickness during the second six weeks of lactation. In women who were actively replenishing their subcutaneous fat stores, milk output was low. During the farming season virtually all the women were producing considerable amounts of milk at the expense of their subcutaneous fat stores.
Z l-
1
4
1
1
-2
0
Skinfold
change
+2
mm/6
+4
1
+6
weeks
Fig. 3. Relation between 12-hr breast milk output at 3 months and changes in the sum of triceps plus subscapular skinfold thickness l Heavy farming period; over the second six weeks of lactation. 0 Remainder of vear (r = -0.69, P < 0.01).
Discussion The results have shown that in spite of the multiplicity of other environmental constraints on the mother, including infections such as malaria, maternal energy balance has a quantifiable effect both on weight gain during pregnancy and on the weight of the baby at birth. Importantly, however, the precise relationship between food intake and the progress of pregnancy depends not only on the energy content of the food but also on the level of activity of the mother, and this is always something that must be taken into account when the adequacy of the maternal diet to the successful outcome of pregnancy is being considered (WHO, 1965). Our findings also provide strong circumstantial evidence that dietary intake during pregnancy is also of crucial importance for lactation capacity because of the building up of subcutaneous fat stores. In tropical paediatric circles much emphasis is now being given to the philosophy that it is better to feed the mother and hence the child, via breast milk, rather than to concentrate just on the provision of milk products for the supplementary feeding of the young infant. Unfortunately most people’s
A. A. PAUL et al.
attention has concentrated only on the lactation period itself and, by and large, the results of interventions involving the supplementary feeding of nursing mothers have not led to dramatic improvements in milk production (HOLEMANS et al., 1954; GOPALAN, 1958; CHAVEZ et al., 1975). In many developing countries, such as The Gambia, where lactation is universal until the infant is 18 to 24 months of age, women under 45 are pregnant or lactating virtually all the time. If mothers are to improve their milk output then we must become responsible for the nutritional well-being of women during the whole of their reproductive lives. Levels of energy intake as low as 1500 kcal during pregnancy and little more during lactation are so much below the recommended dietary intakes that incomplete measurements might be suspected. Indeed, THOMSON et al. (1966) estimated on the basis of assumed levels of energy expenditure that energy intakes in the Keneba women must have been about 3,000 kcal/day. In the present study, however, there was scrupulous care taken to avoid all obvious sources of error, such as snack foods. Subsequent work in the same village with much larger numbers of women has confirmed the essential features of the present findings and we are confident that the low food intake values presented here are a true representation of what women really are eating. It is relevant to point out that similar intakes have been found in other carefully carried out quantitative studies of food intake, for example NORGAN et al. (1974) reported intakes of 1400 kcal (5.85 MJ) in pregnant and lactating women in the coastal zone of New Guinea; MALETNLEMA & BAVU (1974) found intakes of 1850 kcal (7.72 MJ) in pregnant women in Tanzania; and GEBRE-MEDHIN & GOBEZIE (1975) reported intakes of 1540 kcal (6.44 MJ) in non-privileged Ethiopians during pregnancy. While such levels of energy intake are clearly inadequate, doubts must also be expressed about the need for the very much higher energy intakes currently recommended. We hope that the more extensive studies at present being carried out by us will produce data which will help us predict more realistic dietary requirements for conditions prevailing in countries such as The Gambia. Acknowledgements We would like to thank the project leader, Dr. M. G. M. Rowland, Sister M. Tully in whose care the women were, M. Hutton for the breast milk measurements, the Gambian nutrition assistants and the villagers of Keneba for their constant and cheerful co-operation. We are grateful to G. J. Hudson, P. M. V. John and B. Bailey for analysis of the foods, K. C. Day for the computer data analysis and T. J. Cole for statistical advice. E. M. M. was supuorted by ‘Co-ooeration Techniaue Suisse’ and‘ ‘Stiftung - zur Forderung der Ernahrungsforschung in der Schweiz’. References Chavez, A., Martinez, C. & Bourges, H. (1975). Role of lactation in the nutrition of low socioeconomic groups. Ecology of Food and Nutrition, 4, 159-169.
691
FAOjWHO (1973). Energy and Protein Requirements. World Health Organization Technical Report Series No. 52%Jood and Agriculture Organization Nutrition Meetings Report Series No. 52. Geneva: World Health Organization. Gebre-Medhin, M. & Gobezie, A. (1975). Dietary intake in the third trimester of pregnancy and birthweight of offspring among non-privileged and privileged women. American Journal of Clinical Nutrition, 28, 1322-1329. Gopalan, C. (1958). Studies on lactation in poor Indian communities. Journal of Tropical Paediatrics and Environmental Child Health, 4, 87-97. Holemans, K., Lambrechts, A. & Martin, H. (1954). Etude qualitative et quantitative du lait de femmes indigenes du Kwango (Congo belge). Revue Medicale de Liege, 9, 714-719. Jelliffe, D. B. (1966). The assessment of nutritional status in the‘community. World Health Organization Monotrraah Series No. 53. Geneva: World Health Organization. Maletnlema, T. N. & Bavu, J. L. (1974). Nutrition studies in pregnancy. Part I. Energy, protein and iron intakes of pregnant women in Kasarawe, Tanzania. East African Medical Journal, 51, 515-528. McGregor, I. A. (1976). Health and communicable disease in a rural African environment. Oikos, 27, 180-192. Miller, D. S. & Payne, l?. R. (1959). A ballistic bomb calorimeter. British Journal of Nutrition, 13, 501-508. Nelson, W. E. (1971). Textbook of Paediatrics. (9th edit.). Philadelphia: Saunders. Norgan, N. G., Ferro-Luzzi, A. & Durnin, J. V. G. A. (1974). The energy and nutrient intake and energy expenditure of 204 New Guinean adults. PhilisophLal Transactions of the Royal Society of London. Series B. 268. 309-348. Southnate,
