Developing Chinese race-specific warfarin dose prediction algorithms.

Area of Science:

• Pharmacogenomics
• Clinical Pharmacology
• Medical Informatics

Background:

• Genotype-guided warfarin dosing algorithms explain only 47-52% of dose variability.
• Individualized warfarin dosing remains a challenge due to inter-individual differences.

Purpose of the Study:

• To develop novel warfarin dosing algorithms for the Chinese population.
• To compare the predictive performance of new algorithms against existing genotype-guided methods.

Main Methods:

• Developed a new warfarin algorithm (NEW-Warfarin) using multiple linear regression.
• Compared NEW-Warfarin with three major genotype-guided algorithms using Mean Absolute Error (MAE).
• Analyzed algorithm performance across different warfarin indications (AF, PE, CRD, DVT, OD).

Main Results:

• The NEW-Warfarin algorithm demonstrated superior predictive accuracy.
• Indication-specific analyses revealed significant variations in predictive performance (R² ranging from 0.424 to 0.902).
• Pulmonary embolism (PE) and deep vein thrombosis (DVT) indications showed the highest predictive accuracy.

Conclusions:

• Warfarin dosing algorithms are more accurate when based on patient indications.
• Developed indication-specific algorithms represent a novel strategy to enhance warfarin efficacy and safety.
• This approach offers improved warfarin dose prediction for diverse patient populations.