Urinary antibiotic exposure across pregnancy from Chinese pregnant women and health risk assessment: Repeated measures analysis

Menglong Geng,Kaiyong Liu,Kun Huang,Yitian Zhu,Peng Ding,Jingjing Zhang,Baolin Wang,Wenwen Liu,Yan Han,Hui Gao,Sheng Wang,Guanjun Chen,Xiaoyan Wu,Fangbiao Tao

DOI: https://doi.org/10.1016/j.envint.2020.106164

Abstract:Multiple antibiotics are widely used in clinic practice and livestock husbandry, but exposure data based on repeated measurements are scarce among pregnant women. Here, we biomonitored 41 antibiotics and their two metabolites in urine samples from 3235 pregnant women over three trimesters. Spearman's correlation coefficient, intraclass correlation coefficient (ICC), percentile analysis and linear mixed models were employed to evaluate the correlations, variability, co-exposure patterns and predictors of antibiotics, respectively. Pregnant urinary creatinine-adjusted concentrations of antibiotics were used to estimate daily exposure dose and assessed health risks. The target antibiotics were detected in more than 90% of urine samples, primarily as preferred as veterinary antibiotics (PVAs), and the 95th percentile urinary concentrations of each individual antibiotics were range from below the limits of detection to 5.74 ng/mL. We observed considerable within-subject variation (ICC: 0.05-0.63) of urinary antibiotics concentrations during pregnancy. More than half pregnant women were co-exposed to two or more antibiotics of different usage classes, while both co-exposure to high percentiles of three usage antibiotics at one trimester or exposure to single usage antibiotics at high-dose through three trimesters were infrequent in the study population, and most pregnant women were continuously exposed to low-dose PVAs across pregnancy. A total of 4.5% samples were showed hazard index values exceeding 1 during entire pregnancy. Urinary levels of antibiotics associated with residence, maternal age and education, pre-pregnancy BMI, household income and gestational week, especially vary by sampling seasons. Taken together, most pregnant women were frequently exposure to low-dose PVAs across pregnancy and some were in a health risk associated with the disturbance of gut microbiota. Multiple measurements of urinary antibiotic concentrations are essential to more accurate charactering the exposure levels during pregnancy. Several predictors of urinary antibiotics should be taken into consideration in future researches.