Continuous data were tested for normality using the Kolmogorov-Smirnov test, and the continuous data of normal distribution were described as mean ± standard deviation (Mean ± SD), and the t-test was used for comparisons between groups. Non-normally distributed continuous variables were shown by median and quartile [M (Q1, Q3)], and the Wilcoxon rank sum test was used for comparisons between groups. Categorical data of groups were compared with the Pearson’s χ2 test, and expressed as cases and the constituent ratio [n (%)]. Statistical power was calculated (power = 1). Missing data were imputed using multiple imputation (Supplementary Table 1 ). Data before imputation were also used for multivariate analyses to conduct sensitivity analyses.
In order to study the association between GDM and preterm birth among vAMA women, we established three models, and odds ratios (ORs) with 95% confidence intervals (CIs) were estimated. Model 1 was a univariate model. Model 2 was a multivariate model adjusting for maternal age at delivery, race, education, and newborn sex. Then all variables were included in a multivariable model for stepwise regression, and the following variables were adjusted for in Model 3: maternal age at delivery, race, education, newborn sex, pre-pregnancy weight, pre-pregnancy BMI, delivery weight, weight gain, smoking before pregnancy, hypertension eclampsia, gestational hypertension, pre-pregnancy hypertension, number of prenatal visits, plurality, total birth order, prior birth now living, prior other terminations, birth weight, pregnancy method, and method of delivery. Subgroup analyses were then performed based on race and use of infertility treatment to demonstrate if and how the association between GDM and preterm birth in vAMA women varied by race and use of infertility treatment. Further, preterm birth was subdivided into extremely preterm, very preterm, and moderate or late preterm birth. Logistic regression was used to investigate the association between GDM and different stages of preterm birth. Model 1 was a univariate model. Model 2 was a multivariate model correcting for maternal age at delivery, race, education, and newborn sex. Model 3 was a multivariate model correcting for maternal age at delivery, race, education, newborn sex, delivery weight, smoking status 2nd trimester, hypertension eclampsia, gestational hypertension, pre-pregnancy hypertension, number of prenatal visits, WIC, plurality, prior other terminations, total birth order, birth weight, pregnancy method, and method of delivery. As for the associations of GDM with NICU admission, low birthweight and small for gestational age in vAMA women, analytical methods were the same as those for the association between GDM and preterm birth, and subgroup analyses by race and use of infertility treatment were also conducted for these outcomes.
All statistical analyses were two-sided, and P < 0.05 was considered to be statistically significant. All analyses were completed by SAS 9.4 software (SAS Institute Inc., Cary, NC, USA).
In order to study the association between GDM and preterm birth among vAMA women, we established three models, and odds ratios (ORs) with 95% confidence intervals (CIs) were estimated. Model 1 was a univariate model. Model 2 was a multivariate model adjusting for maternal age at delivery, race, education, and newborn sex. Then all variables were included in a multivariable model for stepwise regression, and the following variables were adjusted for in Model 3: maternal age at delivery, race, education, newborn sex, pre-pregnancy weight, pre-pregnancy BMI, delivery weight, weight gain, smoking before pregnancy, hypertension eclampsia, gestational hypertension, pre-pregnancy hypertension, number of prenatal visits, plurality, total birth order, prior birth now living, prior other terminations, birth weight, pregnancy method, and method of delivery. Subgroup analyses were then performed based on race and use of infertility treatment to demonstrate if and how the association between GDM and preterm birth in vAMA women varied by race and use of infertility treatment. Further, preterm birth was subdivided into extremely preterm, very preterm, and moderate or late preterm birth. Logistic regression was used to investigate the association between GDM and different stages of preterm birth. Model 1 was a univariate model. Model 2 was a multivariate model correcting for maternal age at delivery, race, education, and newborn sex. Model 3 was a multivariate model correcting for maternal age at delivery, race, education, newborn sex, delivery weight, smoking status 2nd trimester, hypertension eclampsia, gestational hypertension, pre-pregnancy hypertension, number of prenatal visits, WIC, plurality, prior other terminations, total birth order, birth weight, pregnancy method, and method of delivery. As for the associations of GDM with NICU admission, low birthweight and small for gestational age in vAMA women, analytical methods were the same as those for the association between GDM and preterm birth, and subgroup analyses by race and use of infertility treatment were also conducted for these outcomes.
All statistical analyses were two-sided, and P < 0.05 was considered to be statistically significant. All analyses were completed by SAS 9.4 software (SAS Institute Inc., Cary, NC, USA).
Free full text: Click here
