Prediction of Drug-Drug Interactions for Highly Plasma Protein Bound Compounds

Area of Science:

• Pharmacokinetics and Drug Metabolism
• Drug Safety and Interactions

Background:

• Accurate prediction of drug-drug interactions (DDIs) is crucial for clinical study design and risk assessment.
• Regulatory agencies historically used a 1% lower limit for plasma unbound fraction (f_u,p) in DDI prediction for highly bound compounds, leading to frequent false positives.
• The International Council for Harmonisation (ICH) M12 guideline now permits experimental f_u,p values for DDI prediction in highly bound compounds.

Purpose of the Study:

• To evaluate the accuracy of DDI prediction for highly bound compounds using experimental f_u,p values.
• To assess the performance of basic and mechanistic static models in predicting clinical DDI risk for compounds with f_u,p < 1% and observed clinical DDIs > 20%.

Main Methods:

• Evaluation of a drug set with experimentally determined f_u,p < 1% and known clinical DDIs > 20%.
• Application of both basic and mechanistic static models utilizing experimental f_u,p values for DDI risk assessment.
• Comparison of model predictions against established clinical DDI data.

Main Results:

• The mechanistic static model accurately flagged all evaluated compounds for DDI risk using experimental f_u,p values, with no false negatives.
• The basic static model identified DDI risk for all compounds, with a single exception (CYP2D6 inhibition of almorexant).
• Experimental f_u,p measurements improved the accuracy of DDI prediction for highly bound compounds.

Conclusions:

• Experimental measurement of plasma protein binding (f_u,p) enables accurate prediction of DDI potential for highly bound compounds.
• The use of actual f_u,p values enhances the reliability of DDI risk assessment, aligning with recent regulatory guidelines.
• This approach supports improved clinical trial design and drug safety evaluations by reducing false positive DDI predictions.