Nutrition around Conception and the Prevention of Low Birthweight^*

Arthur Wynn

Executive Committee

McCarrison Society

About 80 per cent of perinatal deaths are associated with low birthweight. Mothers' prepregnancy weight for height is correlated with birthweight, but this is only a crude indication of the close connection between low birthweight and maternal consumption of energy, protein and a range of other nutrients before and around the time of conception. Diet influences follicular growth before ovulation and thus affects ovulatory maturation and the number and quality of ova produced. Immediately after fertilisation diet affects the rate at which ova proceed to first and subsequent cleavages and therefore the size of the subsequent fetus. Diet acts not only directly on follicular and embryonic growth but indirectly by affecting gonadotropin secretion. The endocrine system is sensitive to blood concentrations of amino acids and some vitamins and minerals. Gonadotropin secretion is also depressed by smoking, and by some drugs, poisons and diseases, which may as a result also affect birthweight.

In March of 1987, the Health Education Council published a report entitled, The Health Divide: Inequalities in Health in the 1980's. This report stressed the differing experience of health and disease in different parts of our society as did its predecessor, Inequalities in Health, often called the Black Report after Sir Douglas Black, chairman of the Working Group, published by DHSS in 1980. I quote from the more recent of these two reports:

"The risk of death for lower occupational classes in the 1980's was much higher.than that of the highest occupational classes at every stage of life."

Figure 1 shows the quite steep gradient in infant mortality ranging from 8 per 1,000 births for the professional self-employed to 31 for unskilled workers, a range of nearly 4 to 1 (Wynn & Wynn, 1979). There are similar gradients for perinatal death rates, stillbirths and death rates from the main causes including congenital malformations.

Mortality has biological causes and these pronounced occupational, social gradients prompt many questions about the factors that mediate between occupation of parents and raised infant or perinatal mortality. The first clue to these factors is to be found in the close connection between infant mortality and birthweight. In England today over 80 per cent of perinatal deaths and 77 per cent of neonatal deaths are of babies with birthweights below the optimum range of 3,500 to 4,500g. It is seen in Figure 2, which summarizes the data for England & Wales for 1985, that as birthweight declined from 3,500g to below 1,500g the infant mortality increased by a factor of nearly 100. It is not clear in Figure 2 that the optimum range goes up to 4,500g but data from other developed countries such as Sweden and the U.S.A. make it apparent that this is so.

This brief introduction prompts. three questions: The first question is : Is inadequate nutrition of the mother an important cause of low birthweight? Two more difficult questions follow: How does nutrition affect birthweight? When does it do so?

Prepregnancy weight is a crude indicator of prepregnancy nutrition. American and European studies show that the risk of having a baby of low birthweight increases as the mother's prepregnancy weight diminishes as illustrated in Figure 3 (Niswander & Gordon, 1972; US National Center for Health Statistics, 1979). The close correlation of birthweight with maternal prepregnancy weight is set out in Table 1. The correlation with maternal height is seen to be much weaker. This association of low prepregnancy weight and low birthweight certainly points to a connection between low birthweight and nutrition but without any indication of the nutritional factors responsible.

TABLE 1

CORRELATION OF BIRTHWEIGHT WITH MATERNAL PREPREGNANCY WEIGHT AND HEIGHT, 1974, 5,755 WHITE AND 6,012 BLACK MOTHERS

More than 40 years ago Dr. B. Burke and colleagues at the Harvard School of Public Health asked the first question: "Is inadequate nutrition of mothers a cause of low birthweight?" (Burke, 1943, 1948) Figure 4 is based on these studies and shows protein intake assessed by dietary recall during the second half of pregnancy. The association between birthweight and protein consumption for the 216 women studied was statistically very highly significant. Birth length was also recorded and the association with protein consumption is illustrated in Figure 5 showing that the association was not simply with body fat but with the size of the skeleton and indeed of all organs of the body.

This association between maternal protein consumption and birthweight has been found in subsequent epidemiological studies relating nutrient consumption by women to birth weight, length and head size of their babies. Professor Michael Crawford, Chairman of McCarrison Society, and his colleague Wendy Doyle have shown that the women who have low birthweight babies weighing under 2,500g consume not only less protein every day than the mothers who have large babies, but consume also a diet lower in energy, lower in fat and lower in a range of vitamins (Doyle etal., 1982). In a later study the same team showed the association of lower birthweight with a lower maternal intake of energy, protein, fat, a range of vitamins, and of calcium, iron and zinc (Crawford et al., 1986). This study also showed a significant association between low birthweight and a lower maternal consumption of essential fatty acids belonging to both the w3 and w6 groups. These later studies have shown that women who consume more protein and have larger babies also consume more of a range of other nutrients. The quite genuine and oft-repeated associations in Figures 4 and 5 were certainly accompanied by associations of birthweight with maternal intakes of a range of other nutrients too. It is now apparent that low birthweight is significantly correlated with a lower maternal cons umption of energy, protein and of many other nutrients, but most studies throw no light on when maternal diet has this effect.

It was widely assumed that because the kind of data about protein intake, shown in Figures 4 and 5 from the Harvard studies, was recorded during the second half of pregnancy, that the diet had its effect on birthweight during the second half of pregnancy. The 9th revised edition of the United States, Recommended Dietary Allowances 1980, recommended an increase of 30 g of protein for pregnant women. There is, however, no basis for the assumption that the major increases in birthweight, seen to be associated with increased protein intake in Burke's studies, were a result of an increase in protein consumption after diagnosis of pregnancy. It is now apparent that Burke was recording women's habitual diet. Not oniy are there important time-lags between diet and its physiological effects, but it had been shown, for example at the University of Colorado, that women's diet during the three trimesters of pregnancy is highly correlated with their preconception diet (Beal, 1971). Women's diet today is correlated with the same women's diet last week, last month and indeed a year or more previously. Figures 4 and 5 tell us that maternaldiet has an important effect on the outcome of pregnancy but provide no answer td the question: When does diet have its effect? It has been shown in quite expensive research, for example in New York City, that protein supplementation during the latter half of pregnancy has no significant effect on birthweight even on a population of women with an usually high incidence of low birthweight (Rush, 1980). The placenta protects the fetus from the effects of a low maternal protein consumption and indeed from deficiences of other nutrients notably vitamins.

It is not, of course, possible to do experiments on women to show when particular nutrient deficiencies cause low birthweight, but there is a long record of animal experiments. Figure 6 shows the average birthweight of rat pups following initiation of thiamine deficiency at various times before mating from the classical studies of Nelson & Evans at the University of California in the 1950s (Nelson & Evans, 1955). Thiamine is essential for DNA synthesis as part of the mechanism for providing the necessary energy. The Californian experiments examined the effects not only of thiamine but also of protein, riboflavine, pyridoxine and folic acid and showed that a deficiency of any of these could cause low birthweight already before mating. Crawford et al. (1986) reported that the Hampstead mothers in their studies who had the larger babies had higher daily intakes of these same nutrients than the Hackney mothers with the smaller babies. The mothers of the low birthweight babies weighing under 2,500g had lower intakes still of these nutrients and also of fat and calories and niacin, vitamin A and calcium.

Such nutrient deficiencies during the period immediately before mating in animals not only cause low birthweight but if begun early enough inhibit ovulation or cause resorption of the embryo as shown in Figure 7 also based on data from Nelson & Evans. It was shown by other teams and notably by Giroud and colleagues in Paris that maternal nutrient deficiencies could be adjusted to levels that produced congenital malformations in progeny. The papers of this French team published over 25 years ago were summarised in a book (Giroud, 1970). Deprivation before mating in animals or a deficient habitual diet could produce a whole spectrum of pathological pregnancy outcomes of which low birthweight was most commonly observed. But defective nutrition could produce infertility at one extreme and a range of birth defects at intermediate levels of nutrient deficiency.

These classical studies beginning in the 1930s have been greatly extended by recent research showing in greater detail how nutritional deficiencies affect embryonic growth from the very beginning. Immediately after fertilisation the new individual exists only as a single round cell called the zygote which has to divide many times even before implantation in the uterus of the mother. Figure 8 based on studies at the Venezuelan Institute of Research in Caracas, shows that the time taken for the zygote to proceed even to first and second cleavage is already affected by maternal diet before fertilisation (Munoz & Malav~, 1979). Eighty mice were divided into two groups. One group received a standard diet containing 27 per cent protein while the other group received a low protein diet containing only 8 per cent protein for at least 15 days before mating. The mice on the low protein diet produced 35 per cent fewer ova than the mice on the standard diet showing that the protein content of the diet affected ovulation already before mating. Figure 8 shows that 48 hours after mating ova from mice on the normal protein diet had divided more than 3 times and had therefore 16 or more cells. In contrast only 4 cells of the animals on the low protein diet had reached this stage while 64 cells had not achieved first cleavage. Figure 8 shows, indeed, a fast and a slow lane, the ova from mice on the normal protein diet travelling in the fast lane and ova from the mice on the 8 per cent protein diet travelling in the slow lane. A low protein diet is seen to slow down the rate of cell division. Twenty-four hours later, that is 72 hours after mating, the number of ova from the animals on the standard diet that had reached the 16 or more cell stage, had increased by 47 to 139 while only 14 ova from animals on the low protein diet had reached this stage as seen in Figure 9.

How then does a low protein diet have this effect of slowing down the early development of the embryo? Nelson & Evans made another contribution in their research in the 1950s. They showed that the effects of protein deficiency before mating were mediated by the endocrine system (Nelson & Evans, 1954). The effects of deficiencies of protein or thiamine around the time of mating could be largely offset by injection of two hormones oestrone and progesterone. The early development of the embryo depends absolutely on the secretion of the hormone progesterone, sometimes called the hormone of pregnancy, by the corpus luteum. This supporting role of the corpus luteum is needed for the first 6 or 7 weeks of human pregnancy. The corpus luteum is formed from the granulosa cells of the follicle before ovulation and there is no increase in number of these cells after ovulation. So the slow-down in early embryonic development, caused mainly by low hormone concentrations, is largely determined by follicular development before ovulation.

Is it then nutrition that affects follicular development before ovulation and hence determines the number of the granulosa cells? Research at the University of Ibaraki, Japan, has shown that a low protein diet, and also an abnormally high protein diet, affect the number of ova ovulated by female mice as shown in Figure 10 (Tagami & Sudo, 1982). Each number is the average of results from a group of 20 mice. It is noteworthy that a reduction or increase of only 40 per cent in protein intake has a major effect on ovulation. Protein intake has a significant effect if changed only for the 24 hours before ovulation, an effect only increased modestly if extended to 30 days. There is a period of high susceptibility in all mammals immediately preceding ovulation which is illustrated in Figure 10 for protein.

This Japanese research has shown that it is not only fertility and number of ova that are affected by protein intake but the quality of the ova is also affected. It is seen in Figure 11 based on examination of 10,300 ova from mice on 3 different diets that up to 72 per cent of ova on the low protein diet were abnormal. The abnormality is generally apparent on the first day after mating, that is before the first cleavage. It was shown in Figures 8 and 9 that low protein diets produced a slow lane of development in which embryos failed to progress, some being eliminated while others fell behind the normal rate of cell division. The Japanese studies in Figures 10 and 11 show a different aspect of the same effect of a low protein diet, which not only reduces the number of ova but produces many more abnormal ova. Furthermore an excessively high protein diet is also seen to be harmful. In most cases this abnormality of the ova on a low or excessively high protein diet has its origin in faulty follicular development before ovulation. The defective ova after or even before fertilisation are generally eliminated usually by resorption but a very small percentage survive to produce offspring with birth defects.

Animal research has shown that low birthweight, defective ova and birth defects can be readily produced by manipulation of diet around the time of conception. The animal research has also shown that these results are partly if not wholly mediated by the endocrine system. Reproduction requires a normal, healthy hormonal profile. Figure 12 shows how just one vitamin deficiency can upset the normal concentration of a hormone in sows, in this case progesterone, the hormone of pregnancy (Esch et al., 1981). Measured deficiencies of riboflavine and of other essential nutrients can produce congenital mal formations experimentally in animals with a 99 per cent certainty.

Congenital malformations are generally associated with reduced rates of DNA synthesis. The cells of the malformed fetus are reduced in number but unequally in different organs of the body. Figure 13 shows the differing amounts of DNA in organs of malformed fetuses reflecting differing cell numbers compared to cell numbers in the same organs of controls, in fetuses of rat dams deprived of pantothenic acid from 14 to 10 days before mating (Potier de Courcy, 1966).

This association of low birthweight and congenital malformation is apparent in human data from all countries. The association of perinatal death-rate attributed to congenital anomalies and birthweight is shown in Figure 14 from the official statistics for England & Wales for 1985 (OPCS, 1987). It is seen that as birthweight falls the risk of death attributed to congenital malformation increases nearly 100 times. The same association between low birthweight and malformation is found for the main specific types of malformation including congenital heart defects and congenital defects of the central nervous system, digestive system and skeleton. Reduced cell number is a common factor for both low birthweight and malformations in the great majority of cases. Thus the congenitally malformed heart normally has a reduced number of cells compared with the normal heart (Cheek et al., 1966).

Our scientific understanding of the causes of low birthweight and congenital malformation depends very much on 60 years of animal experimentation. There is, however, much human epidemiological information indicating that the causes are basically the same in man. There is quite extensive information about birth spacing following world-wide studies prompted by the World Health Organisation. The British Births Survey 1970 found that about 17 per cent of perinatal deaths and 18 per cent of low birthweight babies under 2,500g were attributable to birth spacing under 2 years. (Chamberlain et al., 1975). It is seen in Figure 15 that infant death-rate increases as the preconception interval falls, based on American data. Close birth spacing is associated with increased risk of the commonest reproductive disorders including low birthweight as seen in Figure 16 based on American official statistics. Close birth spacing is important in animal husbandry and it has been shown that its undesirable effects can be largely overcome by excellent nutrition before a new conception. The animal research suggests that close birth spacing is harmful because of the depletion of maternal nutrient reserves during the latter stages of the earlier pregnancy and by lactation. The hormonal profile is changed both by lactation and by the hypothalamus in response to blood concentrations of many nutrients.

Miscarriage is one kind of reproductive casualty found associated with low birthweight and perinatal death in times of food shortage and following too short an interval since the previous birth. Professor Laurence and colleagues (1980) have also shown the association of poor nutrition and miscarriage among women who had previously had a baby with a neural tube defect in a modern urban setting in South Wales. Figure 17 shows the association of miscarriage with close birth spacing from a study of 18,015 pregnancies in Turkey (Omran & Standley, 1976).

These and similar studies suggest two difficult questions: When does miscarriage have its origin? How does poor nutrition increase the risk of miscarriage?

The association of delayed ovulation with miscarriage and abnormal ova in women is illustrated in Figure 18 (Hertig, 1967; Iffy, 1981). Delay in ovulation is a consequence of slow-down in the growth of the follicle 0containing the ovum before ovulation. It is seen in Figure 18 that the normal ovum is ovulated around day 14 of the menstrual cycle. It seems that anything that slows down the growth of the follicle can cause miscarriage and abnormal ova. Figure 19 is a diagrammatic representation of the follicle on days 1 and 14 of the menstrual cycle. In only 14 days a normal follicle increases in diameter about 12 times and in mass about 1,000 times. This is one of the highest rates of growth achieved in any organ at any stage of the human life cycle. This high rate of growth suggests an answer to the question. How does diet influence the rate of miscarriage? The animal research shows that defective diets slow down and even stop the growth of the follicle and interfere with the maturation of the ovum it contains.

Figure 20 shows 3 sizes of the human follicle at the time of ovulation. Ova below 16mm do not generally survive. The normal range is from 22 to 30mm diameter. From 16 to 22mm there is an increased risk of defective ova, miscarriage and birth defects. Follicular development is sensitive to the actual concentrations of nutrients and hormones in the maternal blood supply particularly during this period before ovulation. As shown above the supply of hormones essential for follicular development is partly controlled by the hypothalamus which is sensitive to blood composition and to the mother's nutrient supplies and also influences appetite.

Nutrition is not the only human environmental factor that influences follicular and embryonic development. Many poisons and xenobiotic chemicals including psychotropic drugs can slow down and delay ovulation and otherwise interfere with follicular development (Wynn, 1987). Smoking, for example, can slow down follicular development and increase the risk of miscarriage and low birthweight by 100 per cent or more as shown in Figure 21 (Kline et al., 1981). Nutrition modulates the effect of xenobiotics. The liver of a well fed person is better able to catabolize poisons.

There are studies covering many years showing an association between infection and low birthweight and preterm birth. Nutrition modulates resistance to infection. The influence of nutrition on reproduction is in no sense limited to its direct effect on the developing follicle or spermatogenesis, or its indirect effect on hormone secretion mediated primarily by the hypothalamus. Nutrition also affects reproduction through its influence on the immune system. Good nutrition is moreover the only effective defence against many viral infections.

We conclude by answering the questions at the beginning of our paper:

1. Inadequate nutrition is a major cause not only of low birthweight, but of miscarriage and congenital anomalies, but deficiencies producing these results are multiple and complex.

2. Nutrient deficiencies have their most serious effects on the outcome of pregnancy during the days before ovulation through conception and very early pregnancy.

3. While nutrient deficiencies can affect the developing ovum and embryo directly, a major effect of nutrient deficiencies is to depress the level of hormones essential to development of both ovum and embryo.

4. Nutrition modulates the effects of poisons including smoke and psychotropic drugs. Many infections are damaging to reproduction and their effects are aggravated by poor nutrition.

REFERENCES

Beal, V.A. (1971). Nutritional studies during pregnancy. Journal of the American Dietetic Association, 58, 312-320.

Burke, B.S. (1948). Nutritional needs in pregnancy in relation to nutritional intakes, as shown by dietary histories. Obstetric and Gynecological Survey, 3, 716-727.

Burke, B.S., Harding, V.V. and Stuart, H.C. (1943). Nutrition studies during pregnancy. Journal of Pediatrics, 23, 506-515.

Chamberlain, C., Philipp, F., Howlett, B. & Claireaux, A. (1975). British Births 7970, Vol.2. Obstetric Care. London, Heinemann p. 32.

Cheek, D.B., Craystone, J. and Mehrizi, A. (1966). The importance of muscle cell number in children with congenital heart disease. Bulletin Johns Hopkins Hospital, 118, 140-150.

Crawford, MA., Doyle, W., Craft, I.L. & Laurance, B.M. (1986). A comparison of food intake during pregnancy and birthweight in high and low socioeconomic groups. Progress in Lipid Research, 25, 249-254.

.. Doyle, W., Crawford, M.A., Laurance, B.M. & Drury, P. (1982). Dietary survey during pregnancy in a low socio-economic group. Human Nutrition: Applied Nutrition, 36A, 95-106.

Esch, M.W., Easter, RA. & Bahr, J.M. (1981). Effect of riboflavine deficiency on estrous cyclicity in pigs. Biology of Reproduction, 25, 659-665.

Ciroud, A. (1970). Nutrition of the Embryo. Thomas, Springfield. Hertig, A.T. (1967). Human trophoblast: normal and abnormal. American Journal of Clinical Pathology, 47, 249-270.

Iffy, K. (1981). The concept of the aging ovum. In Iffy, L. and Kaminetzky, H.A. (Eds.) Principles and Practice of Obstetrics and Perinatology, John Wiley, Toronto.

Kline, J., Levin, B., Stein, Z., Susser, M & Warburton, D. (1981). Epidemiologic detection of low dose effects on the developing fetus. Environmental Health Perspectives, 42, 119-126.

Laurence, KM., James, N., Miller, M. & Campbell, H. (1980). Increased risk of recurrence of pregnancies complicated by neural tube defects in mothers receiving poor diets and possible benefit of dietary counselling. British Medical Journal, 281, 1592-1594.

Munoz, C. & Malavd, J.B. (1979). Influence of dietary protein restriction on ovulation, fertilization rates and pre-implantation embryonic development in mice. Journal of Experimental Zoology, 210, 253-258.

Nelson, M.M. & Evans, H.M. (1954). Maintenance of pregnancy in the absence of dietary protein with estrone and progesterone. Endocrinology, 55, 543-549.

Nelson, M.M. & Evans, H.M. (1955). Relation of thiamine to reproduction in the rat. Journal of Nutrition, 55, 151-163.

Niswander, KR. & Cordon, M. (1972). The women and their pregnancies. U.S. Department of Health, Education and Welfare.

Niswander, K.R. & Jackson, E.C. (1974). Physical characteristics of the gravida and their association with birthweight and perinatal death. American Journal of Obstetrics and Gynecology, 119, 306-313.

Office of Population Censuses & Surveys (1987). Infant and perinatal mortality 1985: birthweight. Series DI-13 87/1.

Omran, A.R. & Standley, CC. (1976). Family Formation Patterns and Health. Geneva, World Health Organisation.

Potier de Courcy, C. (1966). Caracteres globaux des m6tabolismes nucl6ique et protdique chez le foetus de rat carence en acide pantothenique. Archives des Sciences Physiologiques, 20, 43-63.

Rush, E., Stein, Z. & Susser, M. (1980). Diet in Pregnan0cy: a Randomised Controlled Trial o Nutritional Supplements. Alan R. Liss N.Y. for March of Dimes.

Tagami, S. & Sudo, K. (1982). Influence on ovulation and ovulated ova in mice of different dietary protein levels and feeding periods. Japanese Journal of Zootechnical Science, 53, 266-271.

US National Center for Health Statistics (1978). Characteristics of Births, United States 1973-1975. U.S. Department of Health, Education, and Welfare.

US National Center for Health Statistics (1979). Weight by Height and Age for Adults 18-74 Years: United States, 197 19774. US Department of Health, Education, and Welfare.

Weiss, W. & Jackson, E.C. (1969). Maternal factors affecting birthweight. In Perinatal Factors Affecting Human Development. Regional Office, World Health Organization, Washington.

Wynn, A. (1987). Preconception Care, Proceedings of a one-day symposium at the City University. Wyeth, London.

Wynn, M. & Wynn, A. (1979). Prevention of Handicap and the Health of Women. Routledge and Kegan Paul, London.

Yerushalmy, J., Bierman, J.M., Kemp, D.H., Connor, A. & French, F.E. (1956). Longitudinal studies of pregnancy on the island of Kauai, Territory of Hawaii. American Journal of Obstetrics and Geynecology, 71, 80-96.

*Based on a paper presented to a joint meeting of UNICEF and the McCarrison Society, October 1987.