Contents - Previous - Next

Maternal body mass index: the functional significance during reproduction

The East Java Pregnancy Study (EJPS)

J. A. Kusini1, S. Kardjati2 and U. H. Renqvist1

1Royal Tropical Institute, Amsterdam, The Netherlands; and 2Nutrition Laboratories, School of Medicine, Airlangga University, Surabaya, Indonesia

Correspondence to: J. A. Kusin, Mauritskade 63, 1092 AD Amsterdam, The Netherlands.


In developing countries low energy intake of women is common, particularly during pregnancy and lactation (Prentice, 1980; McGuire & Popkin, 1989; NAS, 1990, 1991) and the high incidence of low birth weight and growth faltering at an early age has been attributed to maternal undernutrition (Kramer, 1987; Krasovec & Anderson, 1991). While there is no disagreement about the adverse effects of acute and severe energy deficits on the outcome of pregnancy and lactation (Stein et al., 1975; Prentice, 1980), no consensus has yet been reached on the relation between maternal nutrition and reproductive performance in communities having marginal energy intakes either seasonally or chronically (Naismith, 1981; Rush, 1983; Norgan, 1987; NAS, 1990). The trend in current opinion is towards an acceptance that energy requirements for reproduction are relatively low and can be compensated by protective mechanisms such as a lower basal metabolic rate (BMR) and reduced physical activity (Prentice, 1984; Adair & Pollitt, 1985; Prentice & Whitehead, 1987). Non-pregnant, non-lactating women may overcome seasonal hunger by using their energy reserves or by adjusting their activity pattern. The reported seasonal weight loss among adults seldom exceeds 2.5 kg and does not need to be detrimental to health if the lean season is followed by one of excess energy intake to replenish energy stores (Durnin, 1990). If these adjustments can accommodate fluctuations in food availability and the relatively low energy needs for reproduction, can one conclude that chronic energy deficiency (CED) among women of reproductive age is of minor importance? Such a conclusion is contradicted by the poor reproductive performance in developing countries or the inference is that fetal growth is not regulated by maternal energy status (Briend, 1985; Campbell-Brown & McFayden, 1985).

It is difficult to say whether the compensation mechanisms are a successful or desirable adaptation to dietary energy inadequacy. The issue relates to the concept of being 'apparently healthy'. Increased metabolic efficiency in the BMR component of energy expenditure in low body weight subjects may just be an illustration of a decreased mass of active tissue due to a change in body composition in response to chronic energy deficiency (Shetty, 1990). Gopalan (1990) rightly states that a reduction in physical activity only maintains the low productivity and quality of life among populations in energy balance on their present poor diets. The implications of moderate to severe chronic energy deficiency on reproductive performance are still largely unknown (IDECG, 1987). The functional meaning of metabolic adjustments to seasonal low energy intake or periodic extra needs for reproduction depend on whether or not they are superimposed on an already chronically energy deficient condition. Undernutrition in poor communities has been perpetuated for many generations. In Indonesia, for instance, no secular trend in adult height has been observed in the past 40 years or so. The average height of women in the 1970s was 149 cm and average body mass index (BMI) was 18.9 (Kusin et al., 1979). These values are not different from those of the 1940s-60s (Bailey, 1962; Eveleth & Tanner, 1976).

The limitation of research on maternal nutrition is the almost exclusive preoccupation with infant outcomes: birth weight, breast milk production and infant growth (Kramer, 1987; Krasovec & Anderson, 1991). Little consideration is given to the repercussions of reproduction on maternal nutritional status and vice versa. Maternal depletion over the course of numerous reproduction cycles is an often hypothesized but little-measured phenomenon (Merchant, Martorell & Haar 1990 a,b).

Recently BMI has been introduced to define CED (James, Ferro-Luzzi & Waterlow, 1988; Ferro-Luzzi et al., 1992). Three grades of BMI were suggested to categorize CED as mild, moderate and severe, i.e. 17.0-18.4 (grade D, 16.0-16.9 (grade II) and <16.0 (grade III). Limited data from India suggest that BMI among adult men is predictive for mortality (Satyanarayana et al., 1991) and 24h post-partum BMI categories were remarkably closely associated with birth weight (Naidu, Neela & Rao, 1991). More information is needed to justify a definition of CED based on BMI for pregnant and lactating women, and which is related to infant as well as maternal outcomes.

Data of a longitudinal study on the interrelationship between maternal and child nutrition in Madura, East Java will be used to address this issue. The preliminary results have been reported earlier (Kusin et al., 1992b, 1993).

The East Java Pregnancy Study (EJPS)

This longitudinal study was conducted in a rural community on the island of Madura, East Java. Full data collection started in September 1982. The emphasis of the study in the first phase (until December 1985) was on maternal nutrition in pregnancy in relation to fetal growth and breast milk intake. The second phase focused more on nutrition of the mother-child pair in the preschool years (January 1987-December 1989). Except for birth weight recording, there were no field activities in 1986.

Details of the study population and methodology of data collection have been described in earlier publications (Steenbergen et al., 1989; Kardjati et al., 1990; Kusin et al., 1992 a, b, 1993).

Procedures and methods

In summary, the framework of EJPS was a surveillance of all mothers with children, age 0-60 months, and newly married couples through 4-weekly home visits to detect a pregnancy and to monitor the growth of children. On these occasions the weight of the mothers was also measured, providing prepregnant, and post-partum weights, and birth weight was measured within 24 h of delivery. In the first phase of the study pregnant women were studied intensively. Habitual dietary intake of all pregnant women was recorded by the weighing technique with a field worker staying over 3 consecutive days, once in early pregnancy and monthly in the last trimester (Launer et al., 1991). They were asked to come to the project's antenatal clinic at 4-weekly intervals for an obstetric examination and anthropometric measurements (height, weight, skinfold thickness at 4 sites, upper arm circumference). All anthropometric measurements were taken in duplicate according to standard techniques (Weiner & Lourie, 1969). Standardization exercises were done at 3-6-monthly intervals (Habicht, 1973). In a 10% random sample (n = 72) women's activities were recorded chronologically over a 12h period using a digital watch. Only a subsample of lactating women was covered for recording of food consumption (n = 78) and physical activities (n = 108).

General background

The two study villages were located in an area with two rain periods and moderate seasonality in food availability (Ferro-Luzzi, Pastore & Sette, 1987). The annual range in mean energy intake was 1500-1700 kcal per day (6.37.1 MJ/day) for pregnant women (Kardjati et al., 1988; Launer et al., 1991). The energy intake of lactating women was about 200 kcal (0.84 MJ) higher (unpublished data).

As in any subsistence economy, women are housewives and productive labourers. The activity records indicate that women were engaged in light or at most moderate activities. Irrespective of physiological state only 5-15% of the day (1-3 h) was spent on strenuous work, such as working in the fields, selling fish at the market and fetching water (unpublished data). The level of physical activity would thus correspond to 1.5-1.7 x BMR (FAD/WHO/UNU, 1985).

Study population

For the assessment of the functional significance of BMI during reproduction, data from women living in the two villages are used. They had given birth to a liveborn singleton baby in the period September 1982-December 1989, and involved 982 women and 1783 newborns. The coverage of birth weights was 88%; of maternal height 90%, pre-pregnant weight 65% and postpartum weight at 4-7 weeks 61%. Missing data were mainly due to a lapse in data collection between the two phases and during the phasingin of data collection procedures.

Table 1. Pre-pregnant body mass index (BMI) by pre-pregnant weight and height categories: mean (SD)

Height (cm)

Pre-pregnan weight (kg)






BMI for each weight category


17.0 (0.7)

15.6 (0.8)

14.8 (1.0)



-16.2 (1.1)

(n = 39)


18.3 (0.9)

16.9 (0.4)

16.0 (0.3)

15.4 (0.1)


17.2 (1.1)

(n = 95)


19.4 (0.7)

18.3 (0.6)

17.3 (0.6)

16.5 (0 5)

16.2 -

18.1 (1.0)

(n = 298)


21.0 (0.8)

19.8 (0.5)

18.6 (0.5)

17.6 (0.5)


19.1 (1.1)

(n = 279)


22.2 (0.5)

21.3 (0.7)

20.0 (0.7)

18.8 (0.5)

18.3 (0.6)

20.2 (1.2)

(n = 283)

> 50


24.6 (2.1)

22.9 (1.6)

21.6 (1.6)

20.8 (2. ())

22.5 (2.1)

(n = 93)

BMI total

19.4 (1.6)

19.3 (2.0)

19.0 (1.9)

18.7 (1.9)

19.3 (2.l)

19.1 (1.9)

for height categories

(n = 146)

(n = 400)

(n = 372)

(n = 146)

(n = 23)

(n = 1087)

Contents - Previous - Next