Statistical Analysis
Mean (standard deviation) and frequency (percentage), respectively, described the continuous and categorical variables. Baseline characteristics were compared by maternal age group using Student’s t-test for continuous variables and Pearson’s Chi-squared test for categorical variables. The potential confounding factors included BMI, nation, levels of education, history of passive smoking, work condition, history of adverse pregnancy outcomes (including stillbirth, spontaneous and induced abortion), parity, which were adjusted by following multivariate regressions. The associations between maternal age and hyperglycemia/macrosomia were assessed with multivariate logistic regression model by adjusted the potential confounding factors. Odds ratio (OR) and 95% confidence interval (95%CI) were estimated. We also conducted cubic restricted splines fitted in a logistic regression model with knots at the 5th, 35th, 65th, and 95th percentiles, to probe the association between macrosomia and maternal age / FPG (continuous variable). Furthermore, a causal mediation analysis using the mediation package in R software (27) was performed to explore the potential contribute of preconception FPG level to the effect of advanced maternal age on offspring birthweight , with 1000 Monte Carlo draws. Average causal mediation effects (ACMEs), average direct effects (ADEs), total effects (sum of a mediation effect and a direct effect), and the percentage of the variability in the total causal effect explained by the mediator were showed. Findings atp <0.05 were considered significant. Statistical analyses were performed using R-3.6.1 software.