Pre- and early post-natal exposure to phthalates and DINCH in a new type of mother-child cohort relying on within-subject pools of repeated urine samples

Abstract

For non-persistent chemicals such as phthalates, a single spot urine sample only reflects exposure in the past few hours. Collecting repeated urine samples for each participant over windows of sensitivity is expected to improve exposure characterization but has rarely been done. We aimed to rely on within-subject pools of repeated urine samples to assess phthalate exposure during pregnancy and infancy.Women of the French SEPAGES mother-child cohort were asked to collect three urine samples per day over seven consecutive days, twice during their pregnancy (approximatively second (T2) and third (T3) trimesters). For their infants they also collected one sample per day during a week at two (M2) and twelve months (M12). Samples were pooled (within-subject, within-period) prior to phthalate and DINCH metabolite concentrations assessment. Number of pooled samples assayed was 477, 456, 152 and 100 for T2, T3, M2 and M12, respectively. All metabolites were detected in more than 95% of the pooled samples except for the two DINCH metabolites (oh- and oxo-MINCH), MMCHP and oh-MPHP at M2 for which detection frequencies ranged between 64% and 88%. Maternal concentrations of MiBP, MBzP, DEHP metabolites and oxo-MiNP decreased between 2014 and 2017, whereas concentrations of oh-MiNP and the two DINCH metabolites increased (Mann-Kendall p-values < 0.05). While improved compared to studies that relied on spot samples, Intraclass Correlation Coefficients for the pregnancy were below 0.40 for most metabolites. Spearman correlation coefficients between pooled samples collected in infancy were lower than those observed during pregnancy, and were all below 0.30. Exposure to emerging phthalate substitutes such as DINCH and DPHP seems widespread among pregnant women and infants. Collecting repeated urine samples in pregnant women and infants is feasible. The relatively low correlation across trimesters and between maternal and infant samples highlights the need to collect biospecimens in the assumed sensitive time window.