ObjectivesTo compare different formula calculated dosages with the actual doses of warfarin from patients in Beijing Hospital so as to investigate suitable warfarin dosing models for Chinese patients.MethodsOne hundred and three Chinese patients with long-term prescription of warfarin were randomly selected from Beijing Hospital from July 2012 to May 2013. The CYP2C9 and VKROC1 genotypes and basic statistical information were collected. SPSS 18.0 software was used to compare the differences between different formula calculated dosages and the actual dosages of warfarin.ResultsFive genotypes were found in 103 patients, including: CYP2C9 AA genotype + VKORC1 AA genotype (n=72, 69.9%), CYP2C9 AA genotype + VKORC1 AG genotype (n=17, 16.5%), CYP2C9 AC genotype + VKORC1 AA genotype (n=10, 9.7%), CYP2C9 AC genotype + VKORC1 AG genotype (n=3, 2.9%) and CYP2C9 AA genotype + VKORC1 GG genotype (n=1, 1%). Compared with the actual dosages of warfarin, the degree of coincidence was highest for dosages calculated by Jeffrey’s formula.Conclusions Using Jeffrey’s formula to calculate warfarin dosages may be more suitable for Chinese patients with using long-term warfarin. Due to limited sample size, prospective and large sample size studies are required to verify the above conclusion.
Warfarin, a classic oral anticoagulant, is characterized by a narrow therapeutic window and considerable interindividual variability in dosing requirements. This makes precise dose adjustment challenging in clinical practice and increases the risk of bleeding or thrombosis. To improve dose prediction, this study developed a streamlined multilayer perceptron (MLP) model using real-world data from the International Warfarin Pharmacogenomics Consortium (IWPC) database. The LASSO-proj algorithm was applied for high-precision feature selection prior to model construction. The resulting model demonstrated strong predictive performance on the test set, achieving a coefficient of determination (R2) of 0.456, a mean absolute error (MAE) of 8.92 mg/week, and 48.522% of its predictions falling within ±20% of the actual stable therapeutic dose. Through SHAP-based interpretation using DeepExplainer, key features influencing warfarin dosing were identified, including the VKORC1 genotype, body weight, age, and ethnicity. The interpretable MLP framework incorporating LASSO-proj not only maintains high predictive accuracy, but also significantly enhances model transparency, providing a valuable tool for guiding warfarin therapy.