On testing for drug/chemical interactions: definitions and inference

Area of Science:

• Toxicology
• Statistics
• Pharmacology

Background:

• The concept of "zero interaction" varies between statistical and toxicology fields.
• Analyzing complex chemical mixtures (more than three components) presents design challenges, as traditional factorial designs become impractical.

Purpose of the Study:

• To clarify the equivalence of "zero interaction" definitions in statistical and toxicology literature.
• To propose efficient experimental designs for dose-response modeling of chemical combinations.
• To enable comparison of mixture effects against an additivity model.

Main Methods:

• Discussed the conceptual differences and potential equivalencies in "zero interaction" definitions.
• Introduced an economical experimental design using single chemical dose-response data.
• Utilized the estimated additivity surface for comparative analysis of combination data.

Main Results:

• Identified conditions under which statistical and toxicology definitions of zero interaction align.
• Demonstrated that single chemical dose-response data is sufficient for estimating an additivity model.
• Showcased the utility of the additivity surface for evaluating mixture effects.

Conclusions:

• Clarified the definition of zero interaction in mixture toxicology.
• Proposed a practical and efficient design strategy for studying chemical mixtures.
• Enabled robust assessment of dose-response relationships in chemical combinations.