Association of prepregnancy body mass index, rate of gestational weight gain with pregnancy outcomes in Chinese urban women

This was a multicenter, retrospective cohort study of postpartum women and their neonates conducted in 14 hospitals in urban areas of China, including 3 centers in Beijing, 2 centers in Guangdong province, 3 centers in Hunan province, 2 centers in Hubei province, 2 centers in Sichuan province and 2 centers in Shanxi province. Women aged ≥18 years with a gestational age of ≥28 weeks and live birth who delivered during 10th–19th of the last month of every quarter from June 2012 to March 2013 were recruited in order to control for seasonal variations. A structured questionnaire was designed to collect information on demographic characteristics, lifestyle behavior, obstetric and medical history of pregnancy, and pregnancy outcomes. During hospitalization for delivery, face-to-face interviews were conducted to collect information on demographic characteristics and lifestyle behavior, and clinical information was obtained retrospectively based on their medical records. Among 9152 participants with full medical history of regularly scheduled antenatal visits and delivery, 226 women with multiple gestation, pre-conception history of severe heart disease or chronic renal disease were excluded. Overall, the current analysis was limited to 8926 deliveries. All participants provided written informed consent, and the study was approved by the institutional review board of Peking University First Hospital.

The weight (kilograms) and height (meters) of all pregnant women were measured at each antenatal visits in light clothing without shoes. At the enrollment after delivery, weight and height at the first antenatal visit, weight at the last antenatal visit in the first trimester or the first antenatal visit in the 2nd trimester, and weight at the last antenatal visit or the time of delivery were recorded based on the medical records. Prepregnancy BMI was calculated as the weight in kilograms divided by the square of height measured in meters, and classified into four groups according to the Chinese standard [18]: underweight (BMI < 18.5 kg/m²), normal weight (18.5 kg/m² ≤ BMI < 24 kg/m²), overweight (24 kg/m² ≤ BMI < 28 kg/m²) and obese (BMI ≥ 28 kg/m²). The rate of GWG in the 2nd and 3rd trimesters was calculated as: [(final weight measured at the last antenatal visit or the time of delivery – weight measured at the last antenatal visit in the first trimester or the first antenatal visit in the 2nd trimester) / (gestational age at delivery – 13 weeks)] [16]. According to the 2009 IOM guidelines, the rate of GWG was classified into insufficient, adequate and excessive if it was below, within, or above the recommendations as follows: 0.44–0.58 kg/w (underweight), 0.35–0.50 kg/w (normal), 0.23–0.33 kg/w (overweight), and 0.17–0.27 kg/w (obese) [16].

The main outcomes of this study were cesarean delivery, preterm birth (defined as delivery before 37 weeks gestation [19]), large-for-gestational age (LGA) and SGA. LGA and SGA were indicated by birth weight less than and greater than the 10th and 90th percentile, respectively, for the same gestational age by sex, according to the Chinese neonatal birth weight curve [20].

Covariates included age, gestational age at delivery, education, drinking during pregnancy, passive smoking, annual household income, number of parity, GDM, and PIH. Age, education, drinking during pregnancy, passive smoking, annual household income and number of parity were assessed by the interviewer-administered questionnaire. Gestational age at delivery was determined from the date of last menstrual period to the date of delivery and expressed in the week after the last menstrual period. If the date was uncertain, ultrasonography was used to determine gestational age. GDM was diagnosed according to the diagnostic criteria amended by China’s Ministry of Health [21], which recommend that the diagnosis should be made when any one of the following values is met or exceeded in the 75 oral glucose tolerance test at 24–28 weeks: 0 h, 5.1 mmol/L; 1 h, 10.0 mmol/L; and 2 h 8.5 mmol/L. PIH was defined as systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg after 20 weeks of gestation [22].

Demographics characteristics and pregnancy outcomes were presented as numbers and frequency distributions for categorical variables, or median and interquartile range for continuous variables, and were compared using the chi-square test or Kruskal-Wallis test. Multivariable logistic regression models were conducted to estimate odds ratios (ORs) and their 95% confidence intervals (CIs) of pregnancy outcomes across categories of prepregnancy BMI or rate of GWG. Models were adjusted for age, gestational age at delivery (except for the outcome of preterm birth), education, drinking during pregnancy, passive smoking, annual household income, number of parity, and study centers. Further models included mutual adjustment of prepregnancy BMI or rate of GWG as appropriate. The normal weight group and adequate GWG rate group were used as the reference groups, respectively. Sensitivity analyses were conducted to additionally adjust for GDM and PIH. Restricted cubic spline (RCS) logistic regression models were used to assess nonlinear effects of GWG rate on adverse pregnancy outcomes by treating the median level of GWG rate as the reference, using 4 knots located at the 5th, 35th, 65th, and 95th percentiles of the distribution of GWG rate. Analyses were carried out using SAS software version 9.2 (SAS Institute, Cary, NC), and the RCS was performed using SAS RCS_Reg macro [23]. All P values are two-sided, and a 0.05 level was used to declare significant differences.

Among 8926 postpartum women, 22.6, 27.4 and 50.0% of women had insufficient, adequate and excessive rate of GWG, respectively. More than 70% of participants who were overweight or obese had excessive rate of GWG (Fig. 1). Distribution of pregnancy outcomes, prepregnancy BMI and rate of GWG by study centers is shown in Additional file 1: Table S1 and Additional file 2: Table S2, respectively. Table 1 shows demographic and clinical characteristics of study participants according to the prepregnancy BMI and rate of GWG categories. The prepregnancy BMI was inversely associated with rate of GWG (P < 0.001). Women with greater prepregnancy BMI were more likely to be multiparous, and have a higher risk of GDM and PIH, and were less likely to have a higher level of annual household income (all P < 0.05). Overweight women were slightly older and obese women were more likely to be exposed to passive smoking than women of normal weight (all P < 0.05). Women with excessive rate of GWG had a higher likelihood of PIH, and women with insufficient rate of GWG were more likely to be multiparous, and to have a lower education level than those with adequate rate of GWG (all P < 0.05).

Additional file 3: Table S3 illustrates the prevalence of adverse pregnancy outcomes by prepregnancy BMI and rate of GWG categories. Prepregnancy BMI and rate of GWG were positively associated with risk of cesarean delivery and LGA (all P < 0.001). Women with underweight and obese had higher prevalence of SGA compared with those of normal weight (P < 0.001). Both insufficient and excessive rates of GWG were associated with higher prevalence of preterm birth than those with adequate rate of GWG (all P < 0.01).

Table 2 shows associations of adverse pregnancy outcomes with prepregnancy BMI and rate of GWG categories. After adjustment for potential confounders, prepregnancy underweight was associated with higher risk for SGA (OR = 1.71, 95% CI: 1.40–2.09), and both prepregnancy overweight and obesity were associated with higher risk for cesarean delivery (overweight: OR = 1.80, 95% CI: 1.56–2.08; obese: OR = 2.34, 95% CI: 1.69–3.24), and LGA (overweight: OR = 1.75, 95% CI: 1.44–2.13; obese: OR = 2.48, 95% CI: 1.71–3.60). Insufficient rate of GWG was associated with increased odds of preterm birth (OR = 1.34, 95% CI: 1.08–1.65), and SGA (OR = 1.45, 95% CI: 1.16–1.82), while excessive rate of GWG was associated with higher risk for cesarean delivery (OR = 1.22, 95% CI: 1.10–1.35), preterm birth (OR = 1.44, 95% CI: 1.20–1.73), and LGA (OR = 1.58, 95% CI: 1.33–1.87). The results remained significant after further mutual adjustment for rate of GWG or prepregnancy BMI.

Sensitivity analyses were conducted to additionally adjust for GDM and PIH, and results were generally similar between models including and not including these two variables (Additional file 4: Table S4). We further examined the relationships between rate of GWG and pregnancy outcomes by prepregnancy BMI categories. Results were generally consistent pertaining to associations of GWG rate with the risk of pregnancy outcomes across different BMI groups, except for cesarean delivery (Additional file 6: Figure S1).

The RCS analysis presented in Additional file 5: Table S5 partly confirmed the results above. Higher levels of GWG rate were associated with higher risk of cesarean delivery and LGA, and associated with lower risk of SGA (all P _overall < 0.001). Both lower and higher levels of GWG rate were associated with increased risk of preterm birth compared with the median of GWG rate of 0.50 kg/w (P _overall < 0.001, Additional file 5: Table S5 and Fig. 2). Fig. 2 shows the shape of associations between continuous rate of GWG and the risk of preterm birth and LGA based on the RCS logistic regression models. Among all participants, rate of GWG was nonlinearly associated with the risk of preterm birth (U-shaped, P _nonlinear < 0.001) and LGA (P _nonlinear = 0.002) after adjustment for potential confounders. Subgroup analysis based on different BMI categories has been shown in Additional file 5: Table S5. Similar results for the shape and the magnitude of associations of GWG rate with the risk of pregnancy outcomes were found between women with normal weight and the whole population.