The PFAS exposure burden calculator: A novel tool for cross-study harmonization, biomonitoring and report-back - PROJECT SUMMARY Almost every person in the US has detectable levels of per- and polyfluoroalkyl substances (PFAS) in their blood. PFAS are a large class of >4,000 chemicals, known as toxic “forever chemicals”, that contribute to worse cardio-metabolic health, decreased immune functioning, and cancer risk. Focusing on total exposure to PFAS as a chemical class, rather than on individual PFAS chemicals, has been emphasized for regulatory purposes and clinical biomonitoring. However, the field lacks a standardized way to quantify cumulative PFAS exposure burden, which is a key obstacle for advancing PFAS risk assessment and epidemiological findings. Different studies and laboratories measure different sets of PFAS biomarkers. Over time, there is regrettable substitution, in which some PFAS chemicals are phased out of use by industry and substituted with other PFAS chemicals that exert similar health impacts. Our research team is at the forefront of developing a common, standardized scale to quantify PFAS exposure burden, building off of ESI PI Shelley Liu’s R03 and K25. We introduced novel applications of item response theory (IRT), traditionally used to create standardized scales for high-stakes educational testing, and demonstrated how it could be applied to PFAS biomarker data. We developed the 2017-2018 US PFAS exposure burden calculator (R/Shiny app) based on nationally representative data for 2017-2018. We showed that IRT enables us to set a standardized scale that makes full use of all available PFAS biomarker data, even if different studies measure different sets of PFAS biomarkers, so that exposure burden scores are on the same scale across studies, enabling cross-study harmonization. In Aim 1, we will extend our preliminary calculator to additional years (1996-2022) and extend to the peri-natal and early life stages, using nationally representative NHANES data and data from 6 US birth cohort studies for life stages not available in NHANES (HOME Study, Healthy Start, BBC, GenC, Project Viva, CIOB), total N=20,911. We will create personalized exposure burden metrics to account for exposure source heterogeneity due to diet/lifestyle habits, using differential item functioning analysis and mixture IRT; and account for time trends/period effects due to regrettable substitution using IRT vertical scaling. We will create a point-and-click PFAS burden calculator with software maintenance, active user tracking and dissemination. In Aim 2, we will develop new methods to estimate PFAS burden scores for emerging high-dimensional PFAS-omics data, generated from advances in non-targeted high-resolution mass spectrometry exposure science. We will create novel multi-dimensional IRT methods with mixed item response functions for continuous, zero-inflated data, yielding subscores and overall score to PFAS-omics data. Simulations and applications will be based on n=4,188 participants with PFAS-omics data from the Sister Study. Over the course of this R01, we expect that >1,000 studies on PFAS epidemiology with be published; our common, standardized scale of PFAS exposure burden will enable harmonization, pooled analyses, meta-analyses, biomonitoring and report-back.
