Multivariate fetal growth trajectory modeling and its association with maternal fatty acids

Maternal circulating fatty acids are closely linked to fetal development. However, few studies have investigated the relationship between the maternal fatty acid levels at 11-14 weeks of gestation and fetal growth trajectories phenotype. A total of 655 pregnant women were selected from the Complex Lipids in Mothers and Babies cohort. Twenty fatty acids in maternal serum at 11-14 weeks of gestation were quantified by GC-MS. The trajectories of fetal head circumference (HC), biparietal diameter (BPD), abdominal circumference (AC), and femur length (FL) z-score at 11-14, 22-28, and 32-34 weeks of gestation were determined by group-based multi-trajectory modeling (GBMTM). Multinomial logistic regression and quantile-based g-computation (Qgcomp) were used to investigate the relationship between individual and mixed fatty acid exposures and the fetal growth trajectories. GBMTM identified four trajectory groups of fetal ultrasound measurements: stable falling, stable increasing, high stable increasing, and dramatically falling. The incidence of premature rupture of membranes, placental implantation, large-for-gestational-age (LGA) and macrosomia differed significantly between the four trajectory groups (p-values of 0.044, 0.008, < 0.001, and < 0.001, respectively). Higher maternal serum linoleic acid (LA, C18:2 n-6) and α-linolenic acid (ALA, C18:3 n-3) were associated with increased odds in the stable falling (OR = 1.004; 95% CI: 1.001, 1.008) and dramatically falling trajectory groups (OR = 1.042; 95% CI: 1.001, 1.085). Meanwhile, eicosadienoic acid (EDA, C20:2 n-6) was associated with decreased odds in the stable falling trajectory group (OR = 0.845;95% CI: 0.716, 0.998). Fetal growth trajectory classification based on GBMTM reveals associations between different trajectory groups and adverse pregnancy outcomes. In addition, serum fatty acids LA, ALA and EDA at 11-14 weeks of gestation were associated with the fetal growth trajectories phenotype.