CHAPTER 3: Reproductive Stress and Women’s Nutrition

Reynaldo Martorell and Kathleen Merchant, Division of Nutritional Sciences, Cornell University, and Stanford Center for Research in Disease Prevention, Palo Alto, USA.

Review of the literature on reproductive stress and nutrition indicate that researchers have focused on influences on fetal and infant outcomes and have paid little or no attention to the possible repercussions of reproductive stress on the nutrition of women. This is a serious gap in the literature which should be corrected.

The term “maternal depletion syndrome” has appeared frequently in the literature since Jelliffe and Maddocks used it in 1964. These authors proposed that a syndrome of impairment results from the increasing and cumulative nutritional stresses of successive pregnancies and lactation periods, apparently moved to use a dramatically descriptive name by the very poor physical condition of women living in a remote mountainous region of Papua-New Guinea.

The term “maternal depletion syndrome” has been used as a descriptor of the consequences of frequent childbearing under difficult living conditions and as an explanation for other relationships. For example, short birth intervals are associated with increased child mortality in several populations and maternal nutritional depletion has been postulated as a possible mechanism for this relationship. Not only infant mortality, but also low birthweight, child growth retardation, and increased morbidity are attributed sometimes to the effects of nutritional depletion of mothers.

While the term has been useful to call attention to an under-studied subject, the use of the word “syndrome”, which generally refers to clinical conditions identified by specific signs and symptoms, has led to debates as to whether the syndrome exists or not. The debate has not gone far since there is no agreement on the nature of the syndrome and therefore, no way of estimating its occurrence. In fact, most support for the existence of a syndrome has been indirect, based on inferences from studies of the repercussions of reproductive stress on maternal nutritional status. It is proposed that the word “syndrome” be avoided and that instead, the effects of reproductive stress on women’s nutrition be regarded as a continuum, to be measured in terms of the range of outcomes which are understood to relate to nutrition, such a diet, body composition and biochemical indicators. Also, a more interesting question is not whether maternal nutrition is affected by reproductive stress but under what circumstances effects are noted and to what degree and in what aspects.

The results to be shown below come from an unusually rich and widely-known study, the INCAP longitudinal study. The Institute of Nutrition of Central America and Panama, better known as INCAP, conducted research from 1969 to 1977 in four villages of eastern Guatemala. The villages were poor. There was convincing evidence of marked growth retardation and of high prevalence of infectious diseases, particularly diarrhoeal diseases in young children. The intent of the INCAP study was to document the effects of improved nutrition on the physical and mental development of children. As pan of the research design, mothers and children were provided with food supplements. Pregnancies were monitored through home visits every two weeks. Pregnant women were scheduled for a variety of measurements, including diet and anthropometric exams. Newborns were weighed within the first 24 hours and children were studied longitudinally using a rich battery of physical and developmental tests.

The research focused on the examination of women actively exposed to different degrees of reproductive stress. To achieve this aim, it was found helpful to focus on the energy demands imposed by reproductive events as depicted in Figure 1. At one extreme are women free of reproductive stress. If women are neither pregnant or lactating, all available energy will be devoted to physical activity and work or to meet the basic metabolic needs of maternal tissues. If dietary energy intake is insufficient to meet total energy demands, tissue stores will be used as fuel. Alternatively, excesses of intake will lead to tissue deposition. Potential repercussions of previous reproductive behavior such as increased work demands due to child care and the need to replenish body stores used during recent reproductive events are ignored in this discussion.

Figure 1. Energy Flows During Pregnancy and/or Lactation

According to current recommendations put forth by FAO, WHO and UNU, women need an additional 350 kcal during the last six months of pregnancy to meet the demands of fetoplacental tissue growth and maintenance (WHO, 1985). Assuming a baseline intake of 2000 kcal, this amounts to an increase of about 17% in energy intakes. In reality, many women in developing countries do not eat appreciably more during pregnancy. In fact, for reasons that are not entirely clear, intakes during the third trimester of pregnancy may be even lower than prior to pregnancy. Pregnancy, therefore, is an energetically stressful phase in developing countries.

Lactation is even more demanding than pregnancy in energy terms. Current recommendations assume that women will use energy derived from 4 kg of fat accumulated during pregnancy. This reserve is expected to provide about 200 kcal/day during the first six months of lactation. In addition, FAO, WHO and UNU recommend that dietary intakes during the first six months of lactation be increased by 500 kcal/day, representing an increase of approximately 25% over the non-pregnant, non-lactating state (WHO, 1985). In reality, many women begin lactation with notably less than the expected amount of fat reserves or even with a net loss of fat. Moreover, their energy intakes, while frequently greater than those of non-pregnant women, often fall substantially short of the recommended.

What if women are faced with pregnancy and lactation simultaneously? This would result, surely in a pronounced degree of reproductive stress. The flow of energy would proceed along all pathways identified in Figure 1. That is, in addition to the ordinary needs of women, this situation would demand energy for milk synthesis as well as for fetal and placental tissues. But, can this situation occur and with sufficient frequency to permit study? The answer is yes to both questions.

The notion that pregnancy and lactation can occur simultaneously is so startling that it deserves some comment. Lactation and pregnancy are generally studied as separate phenomena. To an extent, these physiological states are incompatible because of the antagonistic effect produced by their corresponding hormonal controls. For example, the steroids of pregnancy inhibit the onset of lactation; breastfeeding in turn delays cyclic ovarian activity through poorly understood hormonal mechanisms. However, pregnancy and lactation can and do overlap. Lactation is not prevented by pregnancy if it has been established before conception.

In traditional societies with long durations of lactation, conceptions will occur despite breastfeeding. This is not to negate the fact that breastfeeding generally prolongs the period of postpartum amenorrhea. Ovulation may be delayed for many months but eventually it will reappear in women who breastfeed, generally after the introduction of other foods to the infant’s diet. Therefore, in societies where the use of artificial contraceptives is rare, many women will become pregnant while breastfeeding unless there are strong taboos against sexual intercourse during the period of lactation.

Behavioral studies in the Third World indicate that pregnancy is cited frequently as a reason for weaning (Vis and Hennert, 1978). The belief is widespread that pregnancy and lactation are incompatible states. For example, breast milk sometimes is viewed as harmful to the fetus; conversely, pregnancy sometimes is thought to spoil or damage the milk. The fetus, the toddler, or even the mother may be viewed as being at risk if pregnancy and lactation overlap. But, overlap is probably common because many women may not realize they are pregnant for several months after conception. Also, not all women may choose to wean their children upon discovering they are pregnant.

Surprisingly, little is known (Merchant and Martorell, 1988) about the frequency and extent to which pregnancy and lactation overlap. Where information is available, however, indications are that overlap is a common occurrence. A study of breastfeeding practices in Central Java found that 40% of the mothers who weaned their children were known to be pregnant (Bracher and Santow, 1982). In Senegal, 30% of the study sample who were breastfeeding became pregnant. A substantial proportion of these women continued to breastfeed a previous child: 62% were breastfeeding at 3 mo, 19% at 6 mo, and over 4% breastfed into the ninth month of pregnancy and beyond (Cantrelle and Leridon, 1971). It was estimated by use of cross-sectional data from Bangladesh that 12% of the women who were pregnant at the time of one survey also were breastfeeding a previous child. The cumulative probability of lactation during pregnancy was calculated using data from several successive surveys in the same study. These calculations indicate that among women who were pregnant and breastfeeding a previous child, 45% continued breastfeeding through the sixth month of pregnancy and nearly 20% were breastfeeding at the beginning of the ninth month (Huffman et al., 1980).

The INCAP prospective data set permits careful estimation of the extent of overlap for a rural Guatemalan sample. These estimates are based on 504 pregnancies for which complete information was available. The results are that overlap occurred in 50.2% of the sample. Overlap was defined as two or more weeks of breastfeeding during pregnancy.

There was substantial variability in Guatemala in the duration of overlap (Figure 2). In 56% of cases of overlap, breastfeeding was confined to the first trimester of pregnancy. In 41% of cases, breast-feeding continued into the second trimester. In 3% of cases, breastfeeding continued into the third trimester. Obviously, some women must have known they were pregnant but made a conscious decision to continue breastfeeding, at least for some time. Others may have weaned their infants immediately after discovering their state. Unfortunately, ethnographic information necessary for understanding how women in Guatemala view the phenomenon of overlap were not collected.

Figure 2. Duration of Overlap of Lactation with Pregnancy (n=253)

To fully appreciate the nature of the stresses brought on by overlap, the components of the birth interval in pregnancies with and without breastfeeding need to be examined. A representation of a common pattern seen when there is no overlap is shown at the top of Figure 3. The extremes of the interval are defined by the first and second pregnancies (P₁ and P₂). Two components separate one pregnancy from the other. The first is the period of lactation (L) which can vary widely in duration and the second is the recuperative interval (R). The latter is an important component for it represents a time when active reproductive stress is absent and therefore, an opportunity for maternal stores to be replenished. Ideally, both long durations of lactation as well as long recuperative intervals are desired.

When overlap occurs, the nature of the birth interval changes dramatically as shown in the bottom of Figure 3. The lactation interval may be shortened considerably. Overlap (O) may extend through part or all of the second pregnancy. Also, when overlap occurs, the recuperative interval is absent. Thus, there are two potential consequences when overlap occurs: energy demands are pronounced due to the simultaneous challenges of pregnancy and lactation and there is a loss of the recuperative interval, leading to reduced energy stores available during pregnancy.

Figure 3. Reproductive Interval Components

After considering the characteristics of the birth interval, four groups which vary in the degree of reproductive stress experienced were constructed (Table 1). Groups 1 and 2 refer to pregnancies where no overlap occurred. Group 1 refers to cases where the recuperative interval exceeds 6 months in duration whereas Group 2 includes cases with shorter birth intervals. Groups 3 and 4 involve overlap and hence in these, the recuperative interval is absent. Group 3 involves overlap only during the first trimester of pregnancy whereas Group 4 involves a longer duration of overlap. Clearly the degree of stress increases from left to right and is least for Group 1 and greatest for Group 4.

Table 1: Nutritional Stress During Pregnancy

	Group 1	Group 2	Group 3	Group 4
Recuperative Interval	Long (> 6 months)	Short (<6 months)	Absent	Absent
Overlap	Absent	Absent	Short (1st Trimester only)	Long (2nd or 3rd Trimester)
Stress	+	++	+++	++++

In a series of publications, the implications of reproductive stress on mothers and infants have been examined in analyses which use all groups (Merchant, Martorell and Haas, 1990a, 1990b; Martorell and Merchant, 1990). In this brief report, only the extreme comparison between Groups 1 and 4 is presented.

It will be recalled that the women in the study were provided with liquid food supplements. These were available twice a day and in any desired amount. Consumption of these supplements was carefully measured each time and very precise measures of daily energy intake from the supplement are available.

Striking differences were found between supplement intake of least and most stressed pregnancies (Figure 4). These differences were evident throughout pregnancy but were most pronounced during the first trimester. Differences during the first trimester amount to about 80 kcal per day. Interestingly, intake levels were similar at three months postpartum. This lends support to the view that the increased consumption was driven by a greater energy demand during pregnancy. It should be added that there were no differences between least and most stressed mothers in home dietary intakes, measured through 24-hour recall surveys every three months. These data may have lacked the necessary precision to allow us to detect small differences in intake.

Figure 4. Supplement Intake (+ SEM) in Least and Most Stressed Mothers⁺

⁺ Adjusted for supplement type and study month.

The thickness of the thigh fatfold was used as an indicator of maternal fat stores (Figure 5). It was found that thigh fatfold thicknesses were lower in most as opposed to least stressed mothers. This occurred in spite of the increased supplement intake of the most stressed mothers. Again, differences were most pronounced earlier in pregnancy; at three months postpartum, fatfold thicknesses were similar in both groups. The interpretation of the thigh fatfold data becomes more complex when the results of all four comparison groups are considered. See Merchant et al., (1990a) for a detailed description and possible interpretation of these data.

Figure 5. Maternal Thigh Fatfold Thickness ( + SEM) for Least and Most Stressed Pregnancies⁺

⁺ Longitudinal cases. Adjusted for maternal age, parity, relative measurement date, and study month.

Comparison of the mean birthweight corresponding to least and most stressed pregnancies shows a difference of 115 g in the expected direction (Table 2). This difference is statistically significant if a one-tail t-test is carried out. Since clear directionality was hypothesized, a one-tail test may be appropriate. When the response to reproductive stress was examined in consecutive pregnancies of the same woman the difference in birthweight between the two groups of stress could not be confirmed, perhaps because this sub-sample was too small. See Merchant et al., (1990b) for more information.

Table 2: Birthweight (g) in Least (Group 1) and Most (Group 4) Stressed Pregnancies⁺

	Least (n=85)	Most (n=103)
Mean	3,204	3,089
SEM	51	51
Difference	115 g t = 1.68

+ Adjusted for sex and gestational age of newborn and for maternal height, age and parity.

Differences in growth between infants arising from least and most stressed pregnancies were also examined (Table 3). At one year of age, infants from least stressed pregnancies were larger than those from most stressed pregnancies. This was the case for length, weight, and head circumference. In all cases, the differences were statistically significant.

Table 3: Infant Size ( ± SEM) at 1 Year of Age from Least (Group 1) and Most (Group 4) Stressed Pregnancies⁺

	Least (n=78 to 90)	Most (n=84 to 95)	Difference	t
Length (cm)	69.08 + 0.31	68.34 + 0.30	0.74	1.72
Weight (kg)	7.96 + 0.12	7.68+0.11	0.28	1.70
Head Circumference (cm)	44.28 + 0.15	43.85 + 0.15	0.43	2.02

+ Adjusted by sex of the child, type of supplement and maternal height.

Discussion

It was found that half the pregnancies in the Guatemalan sample studied involved 2 or more weeks of breastfeeding and that many women breastfed well into the second trimester of pregnancy. Two types of pregnancies were compared which represented the extremes of active reproductive stress in the study population. One was subjected to the lowest level of stress and involved women who benefitted from a recuperative interval of 6 or more months prior to pregnancy. The other group was subjected to the highest level of stress and included women who breastfed into the second trimester of pregnancy or beyond. To emphasize, one group was not exposed to overlap and had a long recuperative interval and the other had no recuperative interval and had a long duration of overlap. We find that both maternal and fetal outcomes were affected. Maternal intakes rose but fat stores declined in women exposed to the greatest amount of reproductive stress during pregnancy. Birthweights were lighter and infant sizes were smaller for the group of mothers exposed to the greatest amount of stress during pregnancy. Thus, a consistent pattern of effects in the expected direction was detected.

What are the policy implications of the study? First, there is the phenomenon of overlap. These results and those few gleaned from the literature suggest that the overlap of pregnancy and lactation is quite common in traditional societies in the Third World. It is surprising that no one had called attention to this condition. Should overlap be avoided? What are mothers confronted with overlap to do? The answers depend partly on what the effects of overlap are on mother and child. The results presented suggest small but consistent adverse effects on mothers and children. This should not be taken to mean that overlap is harmless. The Guatemalan population was receiving supplements and this probably offset some of the reproductive stress on maternal fat stores and fetal growth. Also, as poor as the Guatemalan population was, there are populations that are markedly worse off elsewhere. Women and children in many regions of the world may show more dramatic repercussions arising from overlap than observed in Guatemala. Clearly, this is an issue that deserves to be explored in future research. Obviously, the condition of overlap should be prevented. Breastfeeding is not an effective means of birth control once any of the following three events occur: the infant becomes older than 6 months, the mother resumes menstruation, or the infant ceases to be fully breastfed (Kennedy et al., 1989). Therefore, alternative methods should be used once one of these events occurs. This way, not only will overlap be prevented, but a long recuperative interval will be made possible. But what if overlap occurs anyway? What can one recommend to mothers? This is a difficult question. To answer it properly one should consider not only the health of the mother and her fetus but that of the older infant being breastfed, an issue not yet explored in this study. Whether to wean the child or not might depend partly on the age and health of the breastfed child and on the nutritional status of the mother. If the child is more than say 18 months of age and in good health, then the mother might be encouraged to wean the child to encourage adequate fetal growth and maternal fat reserves. Alternatively, the child might be younger and malnourished. If the mother is able, perhaps she should continue to breastfeed in these cases. If the mother decides to continue breastfeeding, for whatever reason, then she should be encouraged to eat substantially more, particularly energy rich foods. Clearly, no satisfactory answer to the problems posed by overlap can be given presently. The safest course is to avoid it through effective family planning programmes.

Earlier, it was stated that it would not be productive to try to define the signs and symptoms of a syndrome. Rather, it is proposed that researchers focus on the effects of reproductive stress on a continuum of women’s nutritional status. This study shows strong evidence that reproductive stress does affect women’s nutritional status, and subsequent infant growth under certain conditions (Merchant et al., 1990a; Merchant et al., 1990b; Martorell and Merchant, 1990). Therefore, the question is not whether or not maternal nutrition is affected by reproductive stress, but rather under what circumstances are effects noted and to what degree and in what aspects. Within this framework, the research reported here merely begins this process.

Acknowledgements

Supported by grant IC-58/03 from the International Center for Research on Women through Cooperative Agreement DAN-1010-A-00-7061-00 with the Offices of Nutrition Health of the US Agency for International Development.

Literature cited

Bracher, M.D. and G. Santow (1982). Breastfeeding in Central Java. Population Studies, 36:413-429.

Cantrelle, P. and H. Leridon (1971). Breast Feeding, Mortality in Childhood and Fertility in a Rural Zone of Senegal. Population Studies, 25:505-533.

Huffman S.L., A.K.M.A. Chowdury, J. Charkraborty, and N. Simpson (1980). Breastfeeding Patterns in Rural Bangladesh. Am J Clin Nutr, 33:144-153.

Jelliffe, D.B. and I. Maddocks (1964). Notes on Ecologic Malnutrition in the New Guinea Highlands. Clin Ped, 3(7):432-438.

Kennedy, K.I., R. Rivera, and A.S. McNeilly (1989). Consensus statement of the Use of Breastfeeding as a Family Planning Method. Contracep., 39:477-496.

Martorell, R. and K. Merchant (April 1990). The Effect of Lactation during Pregnancy on Subsequent Child Growth. Abstract. The 74th Annual FASEB Meeting, Washington, D.C. Abstracts of the Federation of the American Societies for Experimental Biology, 4(3):A362.

Merchant, K. and R. Martorell (1988). Frequent Reproductive Cycling: Does It Lead to Nutritional Depletion of Mothers? Progress in Food and Nutrition Science, 12:339-369.

Merchant, K. R. Martorell and J. Haas (August 1990a). Maternal and Fetal Responses to the Stresses of Lactation Concurrent with Pregnancy and of Short Recuperative Intervals. American Journal of Clinical Nutrition, 52:280-288.

Vis, H.L. and P.H. Hennert (1978). Decline in Breastfeeding: About Some of Its Causes. Acta Paediatr Belg, 31:195-206.

WHO. Energy and Protein Requirements (1985). Report of a Joint FAO/WHO/UNU Expert Consultation. Technical Report Series 724. World Health Organization: Geneva.

Comments

Barry Popkin, Population Center, University of North Carolina, USA.

The concept of maternal depletion, that is the idea of the confluence of poverty and recurrent reproductive cycling on the nutritional status of women and children have been a focus of concern over the last three decades. A recent report of the NAS National Research Council on the effects of short birth intervals focused our attention on the potential importance of this topic by stating that maternal depletion is one of the important potential pathways by which repeated reproductive cycles affect maternal and infant health. Only recently have researchers begun to try to quantify carefully the various dimensions of reproductive stress on the mother and the infant.

Women in less-developed countries experience repeated pregnancies followed by long periods of lactation. Up to 60% of their reproductive life is spent being pregnant or lactating. The length of lactation is key determinant of the percentage of time under reproductive stress. In the past much of our attention has been on the effects of these patterns on pregnancy outcome and infant health. Only recently have researchers began to examine the effects on the health of the mother. In addition most of the early research has focused on the effects of short birth intervals, or the length of the recuperative interval.

The presentation of Drs Martorell and Merchant on concurrent states of pregnancy and lactation fits into this larger topic. They provide evidence of a pattern of reproductive behaviour which has heretofore been largely ignored. They show that a large proportion of pregnant women in their sample fit into the role of pregnancy concurrent with lactation, by presenting data revealing this as a pattern found in Central Java, Senegal and Bangladesh. My colleague Linda Adair and I looked at the data from a prospective study of over 3300 pregnant women representative of the Cebu region of the Philippines. We find a similar incidence of the proportion of pregnant women who lactate concurrent with pregnancy. Our incidence level is 46.2% with a higher overlap in rural than urban areas. Most of this occurs before the women were aware they were pregnant; we found a median duration of overlap of 10 weeks. This is a high proportion for such a potentially nutritionally compromising state. Most important is that a small percentage of our women continued to breast-feed their infant throughout their entire pregnancy. The nutritional status of these women must have been severely compromised.

Other researchers have tried to disagreggate the reproductive pattern and examine its effects on maternal health. My colleague, Linda Adair and I are doing this. Among our preliminary findings are that very long duration breast-feeding is more likely to be associated with reduced maternal weight and body fat stores. Our work on this process is still in progress but points in the same direction as that reported by Martorell and Merchant and the NRC NAS report. This same research shows that dietary intake increases can easily offset the harmful effects of reproductive stress.

There is no doubt that there is a large potential for women of child-bearing age to be under considerable stress from pregnancy. The effects on the women themselves have only recently begun to be studied. It is clear that there is considerable potential for the woman to be nutritionally compromised by the reproductive process and we need to study this phenomena more carefully as noted by Drs Martorell and Merchant, the National Research Council, and others. At the same time, as we have found in our research, the harmful effects of pregnancy and lactation can be offset by improved maternal energy intake.

The lesson of this symposium is obviously that lactation has important health and family planning benefits and we must consider programmes which should accompany breast-feeding promotion programmes which pay attention to the needs of women.