Gestational Weight Gain and Pregnancy Outcomes in 481 Obese Glucose-Tolerant Women

The study was performed in four Danish university centers: the University Hospitals of Copenhagen (Glostrup and Rigshospitalet), Aarhus, and Odense; details have been previously described (8). Women referred to the centers from other hospitals because of a well-defined chronic disease and women with twin pregnancies were excluded beforehand. Only the first pregnancy during this period was included.

A 75-g 2-h oral glucose tolerance test (OGTT) was performed in 1,539 women based on a self-reported prepregnancy BMI ≥27 kg/m² as part of a risk factor–based screening for gestational diabetes mellitus (GDM). Of these, 967 women had a BMI ≥30 kg/m². Women with GDM according to World Health Organization criteria (17) (fasting blood glucose ≥6.1 mmol/l and/or 2-h blood glucose ≥7.8 mmol/l) (n = 323), known diet treatment (n = 10), or incomplete data on weight gain during pregnancy (n = 153) were excluded, leaving 481 obese women (BMI ≥30 kg/m²) eligible for analysis.

The participants in the study received standard obstetrical care as long as OGTT values were normal. Likewise, their infants were treated according to standard routines, and blood glucose was only measured when required by the clinical condition.

The subjects were divided into four groups according to their weight gain during pregnancy: weight gain <5.0 kg (n = 93), 5.0–9.9 kg (n = 134), 10.0–14.9 kg (n = 132), and ≥15 kg (n = 122).

We assessed a number of obstetric outcomes, including hypertension, cesarean section, induction of labor, shoulder dystocia, preterm delivery, birth weight, and 5-min Apgar score. Hypertensive complications were defined as persistently elevated blood pressure (office blood pressure ≥140/90 mmHg on more than two occasions) with or without proteinuria. Induction included vaginal prostaglandin or rupture of the membranes. Shoulder dystocia was defined when obstetrical maneuvers in addition to downward traction, episiotomy, and a mild suprapubic pressure were required to deliver the shoulders. The diagnosis was reevaluated by an obstetrician. Preterm delivery was defined as delivery before 37 completed weeks. Macrosomia was defined by birth weight ≥4,000 g or large for gestational age, which was defined as birth weight ≥90th percentile for a Danish standard population (18). SGA encompassed infants with birth weight <2 SD below the standard mean. Infant morbidity was assessed by admission to the neonatal intensive care unit and low Apgar score (<7 at 5 min).

All analyses were performed with the statistical program STATA 7.0 (STATA, College Station, TX). The effect of weight gain was analyzed by comparing the frequencies of various outcomes in the four gestational weight-gain groups by both uni- and multivariate logistic regression analysis. Adjustments were made for the following variables: 2-h OGTT result, maternal age, prepregnancy BMI, gestational age (continuous variables), parity, smoking, ethnic background, and clinical center (categorical variables). Trends across the weight-gain groups were evaluated by Cuzick’s nonparametric test for continuous variables (19). For categorical variables, weight-gain groups were scored 1–4, and logistic regression analysis was performed across the groups to estimate a possible trend. The results are expressed as ORs, corresponding 95% CIs, and P values. Backwards stepwise linear regression analysis was performed with birth weight as the dependent variable and gestational weight, maternal age, pregestational BMI, 2-h OGTT result, smoking, parity, ethnicity, and gestational age at delivery as the explanatory variables.

Information of weight gain was present for 481 and missing for 153 of the 634 obese women with normal OGTT. The two groups did not differ significantly regarding age (29.7 ± 5.1 vs. 29.2 ± 4.3 years, P = 0.46), BMI (34.5 ± 4.2 vs. 33.5 ± 2.9 kg/m², P = 0.07), and birth weight of the infants (3,668 ± 630 vs. 3,659 ± 608 g, P = 0.8). The median weight gain during gestation was 10.2 kg (interquartile range 6.0–15.0), ranging from −15.3 to 33.5 kg, and the median proportional weight gain (weight gain divided by pregestational weight) was 10.7% (interquartile range 6.2–16.5).

Maternal characteristics and infant birth weight according to the four gestational weight-gain groups are given in Table 1. Birth weight increased significantly across the groups (P < 0.001).

Fasting blood glucose (OGTT) increased with increasing weight gain (P for trend 0.05), and a significant inverse relationship between prepregnancy BMI and gestational weight gain was found (P < 0.001). Maternal age, 2-h OGTT result, ethnic background, smoking, and parity were not significantly associated with weight gain. Gestational age at delivery was the same in the four weight-gain groups.

The relationship between weight gain and adverse pregnancy outcomes is summarized in Tables 2 (univariate analysis) and 3 (multivariate analysis). The risk of hypertension, cesarean section, induction of labor, and macrosomia increased with gestational weight gain, whereas the frequencies of SGA infants, shoulder dystocia, preterm delivery, low Apgar score, and admission to neonatal intensive care unit were similar among the four groups.

In a backwards stepwise multivariate regression model, significant predictors for birth weight were gestational weight gain, prepregnancy BMI, 2-h OGTT result, smoking, maternal age, and gestational age (Table 4).

Originally, recommendations of gestational weight gain were made to secure adequate nutrition for the fetus. In recent years, the prevalence of obesity in the industrialized countries has increased dramatically, and attention has been drawn to the risk of adverse pregnancy outcomes in obese women, primarily delivery of macrosomic infants. GDM is also seen more frequently in obese women and is an important confounder when the risk of macrosomia is determined. Most studies have excluded women with overt GDM, but as even milder degrees of gestational glucose intolerance increase the risk of macrosomia (2,8), this might not be sufficiently accurate to evaluate the effect of obesity and weight gain per se.

In the present study, all women were systematically tested with a 75-g 2-h OGTT during the third trimester because of obesity and the use of World Health Organization criteria for GDM. Linear regression analysis showed that 2-h glucose values were independently associated with birth weight even at values <7.8 mmol/l, suggesting that this information is essential when evaluating the effect of other risk factors.

Our data confirmed that gestational weight gain was negatively associated with pregestational BMI. The reason for this is not obvious, but it has been shown that lean pregnant women accrue significantly more fat mass than obese women (20). During pregnancy, fat is stored to secure energy supply during fetal growth and lactation. In obese women, these fat stores are present already, and it could be speculated that this would tend to limit further weight gain.

Our study showed that gestational weight gain was an independent risk factor for the outcomes of macrosomia, hypertensive complications, induction of labor, and cesarean section. Rates of shoulder dystocia tended to increase across the weight-gain groups, but numbers were small. The risk of preterm delivery and neonatal morbidity was not related to weight gain in pregnancy. However, macrosomia might predispose to later morbidity of the child in terms of insulin resistance (21). Birth weight was strongly related to maternal weight gain in accordance with results of other studies of both normal-weight and obese women (3,22).

Relatively few large studies have specifically dealt with the clinical consequences of gestational weight gain in overweight and obese women (3,6,12), whereas pregestational obesity per se is a well-described risk factor for adverse pregnancy outcomes (4,5,7,9–11). In contrast to the findings of Edwards et al. (3) and Bianco et al. (6), we found that pregnancy complications in obese women were also affected by gestational weight gain. This was also reported by Kabiru and Raynor (12). Systematic information for OGTT was not collected in any of the studies.

As the study was observational and the women nondiabetic per definition, no formalized dietary advice was given, and information of energy intake is not available. It is not possible to conclude which level of weight gain should be recommended, but our data certainly suggest that gaining <5–10 kg is beneficial for the overall pregnancy outcome and that birth weight can be normalized by gestational weight gain restriction in obese women. This aim of a limited weight gain is stricter than the recommendations for overweight women mentioned in the Institute of Medicine report (6.8–11.5 kg) (13) and was obtained by 93 of the 481 women (20%) in our study. Only prospective intervention studies can show whether this goal is realistic.

Dietary treatment with energy restriction in pregnancy is widely used in obese women with GDM. Concern has been raised as to whether such regimens would compromise normal fetal growth and whether increased levels of plasma ketone bodies could be harmful to the fetus. In the present study, the rate of SGA infants was not significantly associated with gestational weight gain. The figures are too small to draw firm conclusions, but the frequencies are comparable with those of the background population. The effect of caloric restriction on ketone levels has previously been studied in women with GDM (23,24). It was concluded that a 50% caloric restriction was associated with a significant increase in ketonuria, whereas a 33% restriction was not. The latter is considered a safe treatment of GDM and is not accompanied by higher frequencies of SGA infants (23). As in GDM, moderate exercise in pregnancy would be of value to limit weight gain. Additionally, exercise before and during pregnancy has been shown to reduce the risk of a number of pregnancy complications, including cesarean section and macrosomia (25,26).

There is an ongoing debate regarding the impact of different maternal factors (glucose, weight gain, and prepregnancy BMI) on birth weight (27). The results from the present study show highly significant associations to birth weight for all three components in obese women with normal glucose tolerance. These factors, in addition to smoking, are all potentially modifiable, whereas maternal age and gestational age at delivery are not.

Apart from pregnancy complications and macrosomia, gestational weight gain is a strong predictor of long-term maternal obesity (28). Only few dietary intervention studies have been performed in pregnant nondiabetic women (29–31), and the results are ambiguous. Olson et al. (30) completed an intervention program in 179 normal- and overweight women consisting of guidance about and monitoring of gestational weight gain by health care providers and a by-mail patient education program. The intervention had no overall effect but prevented excessive weight gain in a low-income subgroup of obese women. Conversely, in the study of Polley et al. (29) intervention limited weight gain in normal-weight women only. Intervention among Cree Indians in Quebec (31) was also unsuccessful.

Overall, more efficient regimens for controlling weight gain in pregnancy are needed. This is of particular importance in obese women, who are already at risk for adverse pregnancy outcome, and dietary advice together with moderate exercise are important tools. Prospective studies should be performed to clarify the safety of such regimens.

This study was supported by the Faculty of Health Sciences, University of Southern Denmark; the NOVO Foundation; the Danish Diabetes Association; Handelsgartner O.V.B. Olesen og ægtefælle fru E. Buhl Olesen’s Mindelegat; Direktør Ib Henriksens fond; Poul og Erna Sehested Hansens fond; Fonden til Lægevidenskabens Fremme; the Danish Medical Research Council; and the Grant Committee of the Consultancy Council of Odense University Hospital.

The assistance with data collection from H. Graugaard, L. Thinggard, K. Bredmose, H. Kayam, and E. Bergholz is gratefully acknowledged.