Benchmark dose profiles for bivariate exposures.

Area of Science:

• Environmental Health Sciences
• Toxicology
• Biostatistics

Background:

• Existing benchmark dose (BMD) methods are limited to single exposures and struggle with multidimensional exposures exhibiting nonlinear effects.
• Assessing the joint impact of multiple environmental factors, such as chemical mixtures or complex exposure patterns, requires advanced analytical approaches.
• Understanding the relationship between prenatal exposure to substances like alcohol and child cognitive deficits necessitates methods capable of handling dose-response complexities.

Purpose of the Study:

• To propose and validate a novel framework for benchmark dose (BMD) analysis applicable to two-dimensional exposures with nonlinear effects.
• To characterize the joint dose-response relationship between multiple exposure variables and a continuous health outcome.
• To provide a method for identifying combinations of exposures that result in a specific level of adverse effect.

Main Methods:

• Development of a generalized additive model to analyze the dose-response surface of two-dimensional exposures.
• Integration of propensity scores to adjust for potential confounding factors in the exposure assessment.
• Utilizing contour plots to visualize and interpret combinations of exposures yielding equivalent effects.

Main Results:

• The proposed BMD framework successfully characterizes the joint effects of multidimensional exposures on continuous outcomes.
• Application to prenatal alcohol exposure demonstrates an association between drinking frequency, amount consumed per occasion, and child cognitive deficits.
• The methodology generates a dose-response surface, visualized via contour plots, identifying critical exposure combinations.

Conclusions:

• The developed BMD framework offers a robust approach for evaluating the combined impact of multiple, potentially nonlinear, exposures.
• This methodology is broadly applicable to various environmental health scenarios, including chemical mixtures and dose-rate analyses.
• The study highlights the utility of advanced statistical modeling in understanding complex exposure-response relationships and informing public health strategies.