Community and International Nutrition
Low Preconception Body Mass Index Is Associated with Birth Outcome in a Prospective Cohort of Chinese Women1
Alayne G. Ronnenberg,2 Xiaobin Wang,* Houxun Xing,† Chanzhong Chen, Dafang Chen,†**
Wenwei Guang,† Aiqun Guang,† Lihua Wang,** Louise Ryan‡ and Xiping Xu
Department of Environmental Health, Harvard School of Public Health, Boston, MA; *Department of
Pediatrics, Boston University School of Medicine and Boston Medical Center, Boston, MA; †Institute for
Biomedicine, Anhui Medical University, Anhui, China; **Center for Ecogenetics and Reproductive Health,
Beijing Medical University, Beijing, China; ‡Department of Biostatistics, Harvard School of Public Health,
Boston, MA

KEY WORDS: ● birthweight ● body mass index
● intrauterine growth restriction

●

China

Although vast improvement has been made in the survival of low birthweight (LBW3 Ͻ 2500 g) and preterm (Ͻ37 completed weeks of gestation) infants, these outcomes remain associated with virtually all causes of neonatal and postneonatal death (1–3). LBW and preterm birth are also associated with infant and childhood morbidity, including asthma (4) and neurodevelopmental delays (3,5). Recent evidence suggests that preterm birth and LBW are also linked to adverse health in adulthood (6), including insulin resistance (7), hypertension (8) and coronary heart disease (9). The potentially serious health consequences of these birth outcomes underscore the public health importance of preventing LBW and preterm birth by identifying and correcting modifiable risk factors. ●

head circumference

Maternal nutritional status is important to maternal and fetal well-being. BMI, weight (kg)/height squared (m2), is influenced by ethnicity and genetics but may also serve as a measure of adiposity and energy balance (10 –12). Although much recent research in developed countries has focused on the association between high maternal BMI and adverse pregnancy outcomes (13), in many developing countries, maternal underweight remains more common than overweight and therefore represents a more important risk factor for poor birth outcomes. Several studies have reported an association between low maternal prepregnancy BMI and increased risk of adverse pregnancy outcomes, including preterm birth (11),
LBW (12) and small-for-gestational age (SGA) (13). Other evidence indicates that the association between maternal weight gain during pregnancy and birth outcomes is influenced by maternal prepregnancy BMI (11,13–15). Available Chinese data also suggest that the risk of delivering a LBW (15,16) or
SGA infant (17) is higher among women with a low BMI before pregnancy than in their heavier counterparts.
Many previous studies have defined low prepregnancy BMI as Ͻ19.8 kg/m2 (11,18). This cut-off value, which includes women who are underweight to varying degrees, was published

1
Supported in part by grants 1R01 HD32505 and 1R01 HD41702 from the
National Institute of Child Health and Human Development; and 1R01 ES08337,
ES-00002, P01 ES06198, and 1R01 ES11682 from the National Institute of
Environmental Health Science.
2
To whom correspondence should be addressed.
E-mail: ronnenberg@comcast.net.
3
Abbrevations used: IUGR: intrauterine growth restriction; LBW: low birthweight; OR, odds ratio; PI, ponderal index; SGA: small for gestational age.

0022-3166/03 $3.00 © 2003 American Society for Nutritional Sciences.
Manuscript received 17 June 2003. Initial review completed 17 July 2003. Revision accepted 11 August 2003.
3449

Downloaded from jn.nutrition.org by guest on February 28, 2016

ABSTRACT Low maternal prepregnancy BMI is associated with adverse birth outcomes, but the BMI at which risk increases is not well defined. We assessed whether the relationship between prepregnancy BMI and birth outcomes is influenced by the extent to which mothers were underweight in a prospective study in Anhui, China.
The women (n ϭ 575) were 20 –34 y old, married, nulliparous and nonsmokers. All measures of infant growth increased with increasing maternal BMI until a plateau was reached at a BMI of 22–23 kg/m2. Infants born to the
27% of women who were severely underweight before pregnancy (BMI Յ 18.5 kg/m2) were at increased risk for fetal growth deficits associated with infant morbidity. Compared with a normal BMI, being severely underweight was associated with mean (Ϯ SEM) reductions of 219 Ϯ 40 g in infant birthweight and 6.7 Ϯ 1.3% in the birthweight ratio and an 80% increase in risk of intrauterine growth restriction [odds ratio (OR) 1.8; 95% CI: 1.0, 3.3; P ϭ 0.05].
Being severely underweight was also associated with smaller infant head circumference and lower ponderal index.
Being moderately underweight (18.5 Ͻ BMI Ͻ 19.8 kg/m2) was not significantly associated with adverse pregnancy outcomes. Gestational age and risk of preterm birth were not associated with maternal BMI. More than half of the women in this study were underweight before pregnancy. Although being moderately underweight was not associated with increased risk of adverse pregnancy outcomes, being severely underweight was an important risk factor for reduced fetal growth. J. Nutr. 133: 3449 –3455, 2003.

3450

RONNENBERG ET AL.

TABLE 1
Major outcomes examined in the study
Outcome

Unit

Gestational age
Infant birthweight
Preterm delivery
Low birthweight (LBW)
Birth weight ratio
Intrauterine growth restriction (IUGR)
Infant head circumference
Infant length
Ponderal index (PI)

Definition

wk kg %

Calculated from the number of days between d 1 of the last menstrual period and the day of delivery
Measured immediately after delivery to the nearest 0.01 kg
Spontaneous delivery of a live infant before 37 completed weeks of gestation
Live infant weighing Ͻ2500 g
Actual birthweight/mean birthweight of infants of the same gestational age within the cohort ϫ 1001

cm cm g ϫ 100/cm3

Birth weight ratio Ͻ85%1
Measured immediately after birth
Crown-heel length
Birth weight ϫ 100/crown-heel length cubed

1 Source: Reference 19.

SUBJECTS AND METHODS
Study site and subjects. The current study is part of a prospective reproductive health study among women textile workers in
Anqing, China, which is an urban area located ϳ200 km west of
Shanghai. All employees of the textile mills and their families receive

health care, including prenatal, delivery and postnatal care, in the nearby hospital. Most of the women in the study work rotating shifts, which are characterized by 2 d of day shifts, 2 d of afternoon shifts and
2 d of night shifts, followed by two days off. The eligibility criteria for the field enrollment in the reproductive health study were as follows:
1) full-time employed women workers; 2) newly married; 3) aged
20 –34 y; and 4) had obtained permission to have a child. All the women were nulliparous. Excluded criteria were as follows: 1) they were already pregnant before enrollment; 2) they had tried unsuccessfully to get pregnant for at least 1 y at any time in the past; and
3) they planned to quit/change jobs or to move out of the city over the 1-y course of follow-up. The study protocols were approved by the
Human Subject Committee of the Chinese Institutions involved in the study and by the Institutional Review Board of the Harvard
School of Public Health.

Data collection
Sociodemographic data. After obtaining informed consent, the interviewer administered a baseline questionnaire that collected his-

TABLE 2
Characteristics of the Chinese women in the study1
BMI category (kg/m2)
Characteristic

18.5 Ͻ BMI Ͻ 19.8

19.8 Յ BMI Ͻ 26

ANOVA

n ϭ 157
Age, y
Height, m
Weight, kg
BMI, kg/m2

BMI Ͻ 18.5

n ϭ 146

n ϭ 272

P-value2

24.9 Ϯ 1.5
1.58 Ϯ 0.04
44.1 Ϯ 2.7a
17.6 Ϯ 0.7a

24.9 Ϯ 1.5
1.58 Ϯ 0.05
47.6 Ϯ 2.8b
19.2 Ϯ 0.3b

24.9 Ϯ 1.6
1.57 Ϯ 0.05
53.0 Ϯ 4.6c
21.4 Ϯ 1.2c

0.88
0.16
Ͻ0.0001
Ͻ0.0001

␹2 P-value

%
Education completed
Elementary
Middle school
High school
College and above
Rotating work shift
Exposure to
Dust
Noise
Work stress
Passive smoking

0
68.2
31.2
0.6
93.6

2.1
67.1
28.8
2.1
94.2

1.5
68.0
29.4
1.1
94.5

95.5
90.4
33.8
69.4

93.8
88.4
30.1
65.1

96.7
90.4
32.7
63.2

1 Values are means Ϯ SD or %, n ϭ 575.
2 Means with different superscript letters differ, P Ͻ 0.05.

0.62

0.74
0.76
0.48
0.92
0.43

Downloaded from jn.nutrition.org by guest on February 28, 2016

by the Institute of Medicine in 1990 with the acknowledgment that there was “no scientific basis on which to accept or reject any of the existing reference standards for evaluating prepregnancy weight-for-height status” (18). The purpose of the current study was to assess the relation between maternal prepregnancy BMI and pregnancy outcomes in a prospective cohort of
Chinese women and to determine how the risk of adverse pregnancy outcomes is influenced by the extent to which women are underweight. Providing a scientific basis for maternal BMI recommendations during the periconceptional period will help to target women who may potentially benefit from dietary and lifestyle interventions.

BMI AND BIRTH OUTCOME IN CHINESE WOMEN

RESULTS
A total of 575 women were included in the study (Table 2).
All women were newly married, nulliparous and were attempting to become pregnant over the course of the study. They tended to be both young and lean; the mean age of the group was 24.9 y and the mean BMI was 19.79 kg/m2 (mean BMI was

19.85 kg/m2 before the exclusion of four women with BMI
Ն 26 kg/m2). Among the women, 25% were moderately underweight (19.8 Ͼ BMI Ͼ18.5 kg/m2; n ϭ 146) and 27% were severely underweight (BMI Յ18.5 kg/m2; n ϭ 157); 5% of women had a BMI Յ 16.9 kg/m2. In addition to rotating shift work, the other major occupational exposures in this cohort were dust, noise and work stress.
The mean birthweight of infants was 3122 g and the mean gestational age was 39.6 wk. A total of 42 infants (7.3%) were born preterm, 37 (6.4%) were classified as LBW, and 69 (12%) had IUGR. Only five infants (0.9%) weighed Ͻ2000 g, and only one (0.2%) was born at Ͻ32 wk. Of the 42 preterm infants, 12 (28.6%) also had LBW; of the infants with IUGR,

Downloaded from jn.nutrition.org by guest on February 28, 2016

torical data on menstruation, contraceptive use, reproductive history, sociodemographic characteristics, active smoking and passive smoke exposure, alcohol use, and environmental and occupational exposures. If a woman reported a missed or late period or had early signs or symptoms of pregnancy, she was instructed to go to the affiliated hospital for a check-up and to give a urine sample to confirm a pregnancy. Once a woman was confirmed to be pregnant, she received regular prenatal care and delivery services at the designated hospitals according to standard clinical guidelines and was followed up for pregnancy outcomes by the research staff.
Anthropometry. At enrollment (which was before pregnancy), the women’s height and body weight in light clothing were measured to the nearest 0.1 cm and 0.1 kg, respectively. Maternal BMI was calculated. The definitions in Table 1 describe the major outcomes of interest. Statistical analysis. The central focus of our analysis was to examine the independent associations between birth outcomes and maternal prepregnancy BMI, modeled as both a continuous and categorical variable. In categorical models, mothers were divided into three subgroups on the basis of their prepregnancy BMI. The “normal
BMI” category included women whose BMI was Ն19.8 kg/m2 but
Ͻ 26 kg/m2 (18). Women with a BMI Ͻ 19.8 kg/m2 were considered underweight, and this category was subdivided into roughly equal groups, based on the median BMI (18.5 kg/m2) for the underweight group. The cut-off value of 18.5 kg/m2 was used by others to categorize “thin” women in Malawi (20). In the current study, women were considered moderately underweight if their BMI was Ͻ19.8 but
Ͼ18.5 kg/m2 and severely underweight if their BMI was Յ18.5 kg/m2.
For all of the subsequent analyses, women with a normal BMI served as the referent. Because only four women had a high prepregnancy
BMI (Ն 26 kg/m2), they were not included in subsequent analyses.
To evaluate the functional relation between maternal BMI as a continuous variable and covariate-adjusted birth outcomes, including birthweight, head circumference, ponderal index (PI), birthweight ratio and gestational age, we applied a generalized additive model with the LOESS smoothing function of BMI using SPLUS 2000
(Mathsoft, Seattle, WA). We plotted the predicted values from the
GAM model against observed maternal prepregnancy BMI. The models were adjusted for maternal age, height, education, work stress, maternal exposure to dust, noise, passive smoking and infant gender.
Birth weight, head circumference and PI models were also adjusted for gestational age (using the smoothing function). Although height is a component of BMI, it was included in adjusted models because it is thought to more adequately capture the joint relation of body composition and body size to health outcomes (10). Birth weight ratio models were not adjusted for gestational age because the definition of birthweight ratio is gestational age specific.
Significant differences in demographic characteristics and infant birth outcomes across BMI categories were assessed using ANOVA with Tukey’s adjustment for multiple comparisons or a ␹2 analysis of the equality of proportions. The associations between maternal BMI and birthweight, gestational age, birthweight ratio, head circumference and PI were assessed using multivariable linear regression with adjustment for the same potential confounding variables included in
GAM models. We evaluated possible effect modification by baby’s sex, maternal height and gestational age by adding product terms to the regression models.
The associations between maternal BMI and preterm birth, LBW and IUGR were each assessed using logistic regression and were expressed as the odds ratio (OR) and its 95% CI. Logistic models generally were adjusted for the same covariates as the LOESS and multivariable linear regression models. Differences were considered significant at P Յ 0.05. Statistical analyses were performed using SAS for Windows, release 8.2 (Cary, NC).

3451

FIGURE 1 Independent relationships between birthweight (panel
A), birthweight ratio (panel B), head circumference (panel C), and ponderal index (panel D) and maternal prepregnancy BMI (kg/m2) in Chinese subjects. Results were derived from generalized additive models using SPLUS 2000. Birth weight ratio was defined as the infant’s actual birthweight divided by the mean birthweight for that particular gestational age (values were multiplied by 100 for convenience). The models were adjusted for maternal age, height, education, work stress, maternal exposure to dust, noise, and passive smoking, infant gender and gestational age (using the smoothing function). The birthweight ratio model was not adjusted for gestational age because the definition of birthweight ratio is gestational age specific.

3452

RONNENBERG ET AL.

TABLE 3
Relationship between maternal prepregnancy BMI and infant birthweight
Birth weight
Multiple linear regression

Birth weight,2 g
BMI,1 kg/m2
Category
Normal BMI
Moderately underweight
Severely underweight

Low birthweight (Ͻ2500 g)
Logistic regression3

n

Mean

SD

␤

SEM

P

Odds ratio

95% CI

P

272
146
157

3192
3159
2966

477
418
401

Referent
Ϫ32
Ϫ219

41
40

0.43
Ͻ0.0001

Referent
0.5
2.0

0.2, 1.6
0.9, 4.5

0.21
0.08

1 BMI categories: normal (19.8 Յ BMI Ͻ 26 kg/m2), moderately underweight (18.5 Ͻ BMI Ͻ 19.8 kg/m2), severely underweight (BMI Յ 18.5 kg/m2).
2 The mean birthweight of infants born to severely underweight women was significantly less than that of infants born to women who were

moderately underweight or who had a normal BMI, P Ͻ 0.0001.
3 n ϭ 39 LBW infants.

TABLE 4
Relationship between maternal prepregnancy BMI and infant birthweight ratio and intrauterine growth restriction
Birth weight ratio2 BMI,1 kg/m2
Category
Normal BMI
Moderately underweight
Severely underweight

Birth weight ratio
Multiple linear regression

Intrauterine growth restriction
Logistic regression3

n

Mean

SD

␤

SEM

P

Odds ratio

95% CI

P

272
146
157

102
101
95

14.2
12.9
12.3

Referent
Ϫ1.0
Ϫ6.7

1.3
1.3

0.45
Ͻ0.0001

Referent
1.1
1.8

0.5, 2.2
1.0, 3.3

0.86
0.05

1 BMI categories: normal (19.8 Յ BMI Ͻ 26 kg/m2), moderately underweight (18.5 Ͻ BMI Ͻ 19.8 kg/m2), severely underweight (BMI Յ 18.5 kg/m2).
2 Birth weight ratio defined as (infant’s actual birthweight/mean birthweight for gestational age) ϫ 100. The mean birthweight ratio of infants born

to severely underweight women is significantly less than that of infants born to moderately underweight women or those with a normal BMI, P
Ͻ 0.0001.
3 Intrauterine growth restriction (IUGR) defined as infant birthweight ratio Ͻ85%; n ϭ 69 IUGR infants.

Downloaded from jn.nutrition.org by guest on February 28, 2016

maternal height and the infants’ sex and gestational age were significantly associated with birthweight and length, inclusion of these variables in adjusted models did not substantially alter the relationship between maternal BMI and any of the pregnancy outcomes. Being moderately underweight before pregnancy was not significantly associated with infant birthweight.
However, being severely underweight was associated with a predicted 7% reduction (219 g) in birthweight (P Ͻ 0.0001).
Moreover, the risk of LBW tended to be doubled among severely underweight women compared with those with a normal BMI (P ϭ 0.08). No evidence of effect modification by baby’s sex, maternal height or gestational age was observed.
The birthweight ratio of infants born to severely underweight women was 7% less than that of infants whose mothers had a normal BMI (Table 4). Compared with women with a normal BMI, the risk of IUGR (birthweight ratio Ͻ85%) was
80% greater among those who were severely underweight
(95% CI: 1.0, 3.3; P ϭ 0.05), but being moderately underweight was not significantly associated with increased risk of
IUGR.
Infant head circumference and birth length were available for a subset of infants (n ϭ 504 and 518, respectively; Table
5). The predicted head circumference of infants born to severely underweight women was 0.83 cm less than that of infants whose mothers had a normal BMI (P Ͻ 0.0001). Being moderately underweight was not significantly associated with decreased head circumference. The mean PI of infants born to severely underweight women was less than that of infants born to women with a BMI Ͼ 18.5 kg/m2 (P Ͻ 0.005). Maternal

11 (15.9%) were born preterm and 35 (50.7%) were also classified as LBW. Only live births were included in this analysis. The proportion of LBW infants, stratified by BMI category, was 5.9% for normal-weight women, 3.4% for moderately underweight women, and 10.2% for women who were severely underweight (␹2: P ϭ 0.05). However, the proportion of preterm births (7.4, 6.9, and 7.6%, respectively) did not differ significantly across BMI categories (␹2: P ϭ 0.96).
The covariate-adjusted relations between maternal prepregnancy BMI and infant birthweight, birthweight ratio, head circumference and PI, all as continuous variables, were generally linear (Fig. 1). The association between BMI and birthweight (panel A) had a relatively steep positive slope up to a
BMI of ϳ22 kg/m2, after which the slope became much less steep but continued to increase up to the maximum BMI of
25.9 kg/m2. The relations between BMI and birthweight ratio
(panel B) and head circumference (panel C) were also linear up to a BMI of ϳ21.5–22.0 kg/m2, where they reached a plateau, indicating that further increases in maternal BMI would likely have little influence on these variables. Although somewhat less consistent than the other three measures, the relation between maternal BMI and PI (panel D) was also linear, with a fairly steep slope up to a BMI of ϳ23 kg/m2, after which a plateau was reached. A linear relation was not evident between gestational age and maternal BMI (data not shown).
Crude (data not shown) and adjusted linear and logistic regression analyses indicated that severely underweight women gave birth to infants weighing significantly less than those born to women with a higher BMI (Table 3). Although

3453

BMI AND BIRTH OUTCOME IN CHINESE WOMEN

TABLE 5
Relationship between maternal prepregnancy BMI and infant head circumference and ponderal index
Head
circumference,2 cm BMI,1 kg/m2

Head circumference
Multiple linear regression

Ponderal index
Multiple linear regression

Ponderal index3

n

Category
Normal
Moderately underweight
Severely underweight

Mean

SD

␤

SEM

P

n

Mean

SD

␤

SEM

P

245
121
140

33.8
33.6
32.9

1.9
1.5
1.9

Referent
Ϫ0.13
Ϫ0.83

0.18
0.18

0.47
Ͻ0.0001

252
124
144

2.5
2.5
2.4

0.2
0.3
0.2

Referent
0.01
Ϫ0.07

0.03
0.03

0.76
0.006

1 BMI categories: normal (19.8 Յ BMI Ͻ 26 kg/m2), moderately underweight (18.5 Ͻ BMI Ͻ 19.8 kg/m2), severely underweight (BMI Յ 18.5 kg/m2).
2 n ϭ 504. The mean head circumference of infants born to severely underweight women was smaller than that of infants born to moderately

underweight women or those with a normal BMI, P Ͻ 0.0001.
3 n ϭ 518. Ponderal index (PI) is defined as 100 ϫ infant weight (g)/(length in cm)3. Mean PI of infants born to severely underweight women was less than that of infants born to women with higher BMI, P ϭ 0.005.

DISCUSSION
The relationship between low maternal prepregnancy weight and adverse pregnancy outcomes has been recognized for several decades (21,22). In this prospective cohort of young, healthy and generally thin Chinese women, we found that infant birthweight and other measures of infant growth increased with increasing maternal BMI until an apparent plateau was reached at a BMI of 22–23 kg/m2. Infants born to the 27% of women who were in the lowest end of the BMI distribution before pregnancy (BMI Յ 18.5 kg/m2), which we classified as “severely underweight,” were at increased risk for fetal growth deficits associated with increased infant morbidity. Being moderately underweight, however, had only a minimal effect on pregnancy outcome. This information could be used to focus potentially costly nutrition and lifestyle interventions on the young women of reproductive age who are more likely to deliver growth-restricted infants if they become pregnant. Such targeted interventions might be expected to have greater effect on infant growth at a population level than those aimed generally at women attempting to become pregnant.
Understanding the effect of maternal adiposity on infant growth parameters has public health importance because these measures are associated with infant health and survival and

TABLE 6
Relationship between maternal prepregnancy BMI and infant gestational age and preterm birth
Gestational
age,2 wk
BMI,1 kg/m2
Category
Normal
Moderately underweight
Severely underweight

Gestational age
Multiple linear regression

Preterm birth
Logistic regression3

n

Mean

SD

␤

SEM

P

Odds ratio

95% CI

P

272
146
157

39.5
39.7
39.5

1.7
1.8
1.7

Referent
0.17
Ϫ0.11

0.17
0.17

0.33
0.52

Referent
0.9
1.1

0.4, 2.0
0.5, 2.3

0.83
0.85

1 BMI categories: normal (19.8 Յ BMI Ͻ 26 kg/m2), moderately underweight (18.5 Ͻ BMI Ͻ 19.8 kg/m2), severely underweight (BMI Յ 18.5 kg/m2).
2 Mean gestational age did not differ across BMI categories.
3 Preterm birth defined as the spontaneous delivery of a live infant before 37 completed weeks of gestation; n ϭ 42.

Downloaded from jn.nutrition.org by guest on February 28, 2016

also may influence development and health in later life. Infant birthweight, for instance, is not only associated with measures of infant health and neonatal morbidity (1,3), but it has also been linked to increased risk of numerous adult disorders
(5,7,9). In addition, reduced PI, which indicates asymmetric growth restriction, is associated with increased systolic blood pressure in children (8) and hypertension in adults (23).
Smaller infant head circumference has been correlated with reduced brain size (24) and smaller head size in childhood
(25), and it may be related to reduced cognitive development and lower intelligence quotient in children (24 –26).
The strength of the associations, the steady increase in the risk of adverse fetal growth outcomes with the increasing degree of underweight, the demonstration of a temporal sequence (with low BMI preceding conception) and the consistency of the results for multiple measures of fetal growth suggest a possible causal relation between severe low prepregnancy BMI and fetal growth deficits. There are several plausible mechanisms for such an association. For instance, low maternal BMI could result from chronically poor energy intake, which would reduce fat stores and compromise visceral and somatic protein status. The recommended increment for energy intake during pregnancy is estimated at 836-1254 kJ/d
(200 –300 kcal/d) (18). If maternal prepregnancy energy deficiency persisted during the first 10 –30 wk of pregnancy, when energy requirements are greatest (18), the substrates necessary to support fetal tissue growth would not be available. Poor maternal nutritional status during pregnancy has also been

prepregnancy BMI was not significantly associated with infant gestational age or the risk of preterm birth in any analysis
(Table 6).

3454

RONNENBERG ET AL.

These observations suggest that BMI and micronutrient status influence pregnancy outcomes through different mechanisms, and they imply that the associations between BMI and pregnancy outcomes are not confounded by maternal micronutrient status.
Understanding the effect of maternal prepregnancy BMI on pregnancy outcomes and the mechanism through which BMI influences fetal growth and development is important from both clinical and public health perspectives. Unlike many determinants of infant growth, low prepregnancy BMI is a risk factor that is easily identified. Although it is not known whether efforts to increase low maternal BMI before conception can reduce the risk of adverse pregnancy outcomes, there is evidence that optimized weight gain during pregnancy may help prevent the adverse effects associated with low maternal
BMI before pregnancy or in the early months of gestation
(13,14,32). Although our study was conducted in a Chinese population, it has implications for women in developed countries as well because the prevalence of overt eating disorders among women of reproductive age is relatively high (33), and these conditions as well as less severe energy restriction may contribute to reduced maternal BMI (34). Studies aimed at identifying severely underweight women before they attempt to become pregnant and intervening in ways that increase prepregnancy weight, optimize pregnancy weight gain and enhance nutritional status should be implemented to determine whether such interventions improve infant growth parameters and reduce infant morbidity.
LITERATURE CITED
1. Vermeulen, G. M. (2000) Spontaneous preterm birth: prevention, management and outcome. Eur. J. Obstet. Gynecol. 93: 1–3.
2. Buehler, J. W., Strauss, L. T., Hogue, C.J.R. & Smith, J. C. (1987) Birth weight-specific causes in infant mortality, United States, 1980. Public Health Rep.
102: 162–171.
3. McCormick, M. C. (1985) The contribution of low birth weight to infant mortality and childhood morbidity. N. Engl. J. Med. 312: 82–90.
4. Wjst, M., Popescu, M., Trepka, M. J., Heinrich, J. & Wichmann, H. E.
(1998) Pulmonary function in children with initial low birth weight. Pediatr.
Allergy Immunol. 9: 80 –90.
5. Liu, X., Sun, Z. & Neiderhiser, J. M. (2001) Low birth weight, developmental milestones, and behavioral problems in Chinese children and adolescents.
Psychiatry Res. 101: 115–129.
6. Rich-Edwards, J. W., Stampfer, M. J., Manson, J. E., Rosner, B., Hankinson, S. E., Colditz, G. A., Willett, W. C. & Hennekens, C. H. (1997) Birth weight and risk of cardiovascular disease in a cohort of women followed up since 1976.
Br. Med. J. 315: 396 – 400.
7. Mi, J., Law, C. & Zhang, K. L. (2000) Effects of infant birthweight and maternal body mass index in pregnancy on components of the insulin resistance syndrome in China. Ann. Intern. Med. 132: 253–260.
8. Law, C. M., Egger, P., Dada, O., Delgado, H., Kylberg, E., Lavin, P., Tang,
G. H., von Hertzen, H., Shiell, A. W. & Barker, D. J. (2001) Body size at birth and blood pressure among children in developing countries. Int. J. Epidemiol. 30:
52–57.
9. Eriksson, J. G., Forsen, T., Tuomilehto, J., Winter, P. D., Osmond, C. &
Barker, D. J. (1999) Catch-up growth in childhood and death from coronary heart disease: longitudinal study. Br. Med. J. 318: 427– 431.
10. Michels, K. B., Greenland, S. & Rosner, B. A. (1998) Does body mass index adequately capture the relation of body composition and body size to health outcomes? Am. J. Epidemiol. 147: 167–172.
11. Schieve, L. A., Cogswell, M. E., Scanlon, K. S., Perry, G., Ferre, C.,
Blackmore-Prince, C., Yu, S. M. & Rosenberg, D. (2000) Prepregnancy body mass index and pregnancy weight gain: associations with preterm delivery.
Obstet. Gynecol. 96: 194 –200.
12. Allen, L. H., Lung’aho, M. S., Shaheen, M., Harrison, G. G., Neumann, C.
& Kirksey, A. (1994) Maternal body mass index and pregnancy outcome in the
Nutrition Collaborative Research Support Program. Eur. J. Clin. Nutr. 48. (suppl.
3): S68 –S77.
13. Ekblad, U. & Grenman, S. (1992) Maternal weight, weight gain during pregnancy and pregnancy outcome. Int. J. Gynecol. Obstet. 39: 277–283.
14. Mitchell, M. C. & Lerner, E. (1989) Weight gain and pregnancy outcome in underweight and normal weight women. J. Am. Diet. Assoc. 89: 634 –
638.
15. Wong, W., Tang, N.L.S., Lau, T. K. & Wong, T. W. (2000) A new recommendation for maternal weight gain in Chinese women. J. Am. Diet. Assoc.
100: 791–796.

Downloaded from jn.nutrition.org by guest on February 28, 2016

associated with reduced placental weight and surface area
(27), which could limit nutrient transfer from the maternal circulation to the fetus, even if dietary intake increased later in pregnancy. Whether smaller placentas result directly from maternal malnutrition or whether the effect is mediated through disruption of normal endocrine function is not known. However, severe food deprivation is associated with reduced serum concentrations of hormones such as leptin and estrogen (28), and it is possible that hormonal factors contribute to fetal growth restriction.
An important limitation of the current study is the lack of data on maternal weight gain during pregnancy, which could modify the relation between prepregnancy BMI and pregnancy outcomes. It is possible, for instance, that some of the women who were very thin before conception gained adequate weight during pregnancy to produce a normal-weight or near-normalweight infant. Had data on pregnancy weight gain been available, we might have been able to stratify low BMI women into those with “adequate” and those with “inadequate” pregnancy weight gain to assess the influence of pregnancy weight gain on birth outcomes in women with low BMI at conception. Several studies have suggested that greater weight gain during pregnancy compensates for the adverse effects on fetal growth associated with low maternal BMI during early gestation
(13,14). However, other evidence indicates that women with low prepregnancy BMI are still more likely to have smaller infants than heavier women, even when their gestational weight gain is the same (18). It is also possible that parity modifies the relation between maternal BMI and pregnancy outcomes. Because all of the women in our study were pregnant for the first time, we are unable to assess this potential effect and can only generalize our results to primiparae.
Another potential limitation of the study is the lack of dietary and physical activity data, which prevents our determining whether low prepregnancy maternal BMI was associated with inadequate energy intake or increased physical activity level. Low BMI is often used as a marker of reduced tissue nutrient reserves (18), but the relationship between
BMI and energy balance is not fully understood. For instance,
Allen et al. (12) reported a negative correlation between energy intake and periconceptional BMI in both Egyptian (r ϭ Ϫ0.71, P Ͻ 0.01) and Mexican (r ϭ Ϫ0.63, P Ͻ 0.001) women and no correlation between the two in women from
Kenya. Possible differences in physical activity level or genetic predisposition may be other important determinants of maternal BMI. Dietary and activity data would help to clarify the relation between energy intake and BMI in this relatively homogeneous Chinese cohort.
Unlike previous reports (11,18), we did not find an association between maternal BMI and gestational age or preterm birth in this group of women. We cannot explain this finding, although it should be noted that many of the genetic and environmental factors associated with gestational age are distinct from those related to infant growth. It is also possible that variables we did not measure, such as gestational weight gain, might have had a more powerful effect on gestational age than on infant growth.
We previously reported a high prevalence of prepregnancy
B-vitamin deficiencies and anemia in a subset of women from this cohort (29); we subsequently found that B-vitamin status was associated with increased risk of clinical spontaneous abortion (30) and preterm birth (31) but was not associated with LBW or SGA (31). Although we observed a small correlation between BMI and plasma folate concentration (r ϭ 0.11; P Ͻ 0.05), BMI was not significantly related to plasma concentrations of homocysteine or vitamins B-6 or B-12.

BMI AND BIRTH OUTCOME IN CHINESE WOMEN
16. Sun, B., Li, J. & Song, Q. (1998) Influence of prepregnancy weight and maternal weight gain on pregnancy outcome [in Chinese]. Zhonghua Fu Chan Ke
Za Zhi 33: 71–73.
17. Xu, B., Jarvelin, M.-R., Xu, X., Wang, Z., Qin, H. & Rimpela, A. (1997)
Maternal menstrual history and small-for-gestational-age births in a populationbased Chinese birth cohort. Early Hum. Dev. 49: 183–192.
18. Institute of Medicine (1990) Committee on Nutritional Status during
Pregnancy and Lactation. Nutrition during Pregnancy, pp. 81– 82. National Academy Press, Washington, DC.
19. Kramer, M. S., Platt, R., Yang, H., McNamara, H. & Usher, R. H. (1999)
Are all growth-restricted newborns created equally? Pediatrics 103: 599 – 602.
20. Neufeld, L., Pelletier, D. L. & Haas, J. D. (1999) The timing of maternal weight gain during pregnancy and fetal growth. Am. J. Human Biol. 11: 627– 637.
21. Niswander, K. & Jackson, E. C. (1974) Physical characteristics of the gravida and their association with birth weight and perinatal death Am. J. Obstet.
Gynecol. 119: 306 –313.
22. Kramer, M. S. (1987) Intrauterine growth and gestational duration determinants. Pediatrics 80: 502–511.
23. Barker, D. J., Godfrey, K. M., Osmond, C. & Bull, A. (1992) The relation of fetal length, ponderal index and head circumference to blood pressure and the risk of hypertension in adult life. Paediatr. Perinat. Epidemiol. 6: 35– 44.
24. Tramo, M. J., Loftus, W. C., Stukel, T. A., Green, R. L., Weaver, J. B. &
Gazzaniga, M. S. (1998) Brain size, head size, and intelligence quotient in monozygotic twins. Neurology 50: 1246 –1252.
25. Chasnoff, I. J., Griffith, D. R., Freier, C. & Murray, J. (1992) Cocaine/ polydrug use in pregnancy: two-year follow-up. Pediatrics 89: 284 –289.

3455

26. Hack, M. & Breslau, N. (1986) Very low birth weight infants: effects of brain growth during infancy on intelligence quotient at 3 years of age. Pediatrics
77: 196 –202.
27. Lechtig, A., Yarbrough, C., Delgado, H., Martorell, R., Klein, R. E. & Behar,
M. (1975) Effect of moderate maternal malnutrition on the placenta. Am. J.
Obstet. Gynecol. 123: 191–201.
28. Mantzoros, C. S. (2000) Role of leptin in reproduction. Ann. N.Y. Acad.
Sci. 900: 174 –183.
29. Ronnenberg, A. G., Goldman, M. B., Aitken, I. W. & Xu, X. (2000)
Anemia and deficiencies of folate and vitamin B-6 are common and vary with season in Chinese women of childbearing age. J. Nutr. 130: 2703–2710.
30. Ronnenberg, A. G., Goldman, M. B., Chen, D., Aitken, I. W., Willett,
W. C., Selhub, J. & Xu, X. (2002) Preconception folate and vitamin B6 status and clinical spontaneous abortion in Chinese women. Obstet. Gynecol. 100:
107–113.
31. Ronnenberg, A. G., Goldman, M. B., Chen, D., Aitken, I. W., Willett, W. C,
Selhub, J. & Xu, X. (2002) Preconception homocysteine and B-vitamin status and birth outcomes in Chinese women. Am. J. Clin. Nutr. 76: 1385–1391.
32. Bracero, L. A. & Byrne, D. W. (1998) Optimal maternal weight gain during singleton pregnancy. Gynecol. Obstet. Investig. 46: 9 –16.
33. Grinspoon, S., Thomas, E., Pitts, S., Gross, E., Mickley, D., Miller, K.,
Herzog, D. & Klibanski, A. (2000) Prevalence and predictive factors for regional osteopenia in women with anorexia nervosa. Ann. Intern. Med. 133: 790 –794.
34. Franko, D. L. & Spurrell, E. B. (2000) Detection and management of eating disorders during pregnancy. Obstet. Gynecol. 95: 942–946.

Downloaded from jn.nutrition.org by guest on February 28, 2016