How Many Urine Samples are Needed to Accurately Assess Exposure to Non-Persistent Chemicals? The Biomarker Reliability Assessment Tool (BRAT) for Scientists, Research Sponsors, and Risk Managers

• Main Authors: Marc-André Verner, Hassan Salame, Conrad Housand, Linda S. Birnbaum, Maryse F. Bouchard, Jonathan Chevrier, Lesa L. Aylward, Daniel Q. Naiman, Judy S. LaKind
• Format: Article
• Language: English
• Published: MDPI AG 2020-12-01
• Series: International Journal of Environmental Research and Public Health
• Subjects: biomonitoring; non-persistent chemicals; exposure assessment; within- and between-person variability; environmental epidemiology; urine sampling
• Online Access: https://www.mdpi.com/1660-4601/17/23/9102

In epidemiologic and exposure research, biomonitoring is often used as the basis for assessing human exposure to environmental chemicals. Studies frequently rely on a single urinary measurement per participant to assess exposure to non-persistent chemicals. However, there is a growing consensus that single urine samples may be insufficient for adequately estimating exposure. The question then arises: how many samples would be needed for optimal characterization of exposure? To help researchers answer this question, we developed a tool called the Biomarker Reliability Assessment Tool (BRAT). The BRAT is based on pharmacokinetic modeling simulations, is freely available, and is designed to help researchers determine the approximate number of urine samples needed to optimize exposure assessment. The BRAT performs Monte Carlo simulations of exposure to estimate internal levels and resulting urinary concentrations in individuals from a population based on user-specified inputs (e.g., biological half-life, within- and between-person variability in exposure). The BRAT evaluates—through linear regression and quantile classification—the precision/accuracy of the estimation of internal levels depending on the number of urine samples. This tool should guide researchers towards more robust biomonitoring and improved exposure classification in epidemiologic and exposure research, which should in turn improve the translation of that research into decision-making.