Reconstructing Breastmilk and Infant Formula Intake using Novel Biomarkers in Baby Teeth

Abstract

TPS 911: Air pollution, epigenetics, biomarkers, Exhibition Hall, Ground floor, August 26, 2019, 3:00 PM - 4:30 PM Background/Aim: Early life diet is a major determinant of lifelong health trajectories. However, studies of the long term health outcomes of breastfeeding are difficult to study due to cost and time restraints in collecting data prospectively and the inherent limitations of questionnaires. A biomarker that can retrospectively measure the dose and duration of breastmilk and formula feeding would be a major advancement in the study of health trajectories associated with infant diet, particularly for health outcomes measured much later in life, years after the cessation of breastfeeding. We show that chemical signals in baby teeth can reconstruct breastfeeding histories. Methods: We measure chemical signatures in aged dentine layers to generate weekly estimates of biomarkers of infant diet. Using reversed distributed lag models (DLM) we identified chemical signatures in teeth associated with breastmilk or formula intake and developed an index of breastmilk intake using a lagged weighted quantile sum (WQS) regression approach. Biomarkers of milk intake were developed in a non-human primate model and refined in a human cohort. We tested the index against prospectively collected breastfeeding histories in a human cohort based in Mexico (n=225) and a replication cohort from the US (n=100). Results: Preliminary studies identified barium as a sensitive marker of infant diet transition from breastmilk to infant formula, identifying the timing of infant formula introduction in 70% of teeth. Reversed DLMs indicated strontium as a sensitive marker to infant formula intake and lithium as a marker for breastfeeding after 6 months. Using WQS, lithium and strontium were found to be the strongest contributors within a mixture to differentiate breastfed and formula fed infants. Conclusions: Breastmilk intake during childhood can be reconstructed using chemical signatures in teeth. This method can be applied to more accurately investigate associations of breastfeeding dose and duration with long term health outcomes.