A Review of A Priori Regression Models for Warfarin Maintenance Dose Prediction

摘要: A number of a priori warfarin dosing algorithms, derived using linear regression methods, have been proposed. Although these algorithms may validated patients from the same centre, rarely they patient cohort recruited another centre. In order to undertake external validation, two cohorts were utilised. One formed by prospective trial and second in control arm EU-PACT trial. Of these, 641 identified as having attained stable dataset used for validation. Predicted maintenance doses six criterion fulfilling models then compared individual dose. Predictive ability was assessed with reference several statistics including R-square mean absolute error. The explained different amounts variability dose requirements validation cohorts; adjusted R-squared values ranged 24.2% 68.6%. An overview summary demonstrated that no one algorithm could be considered optimal. larger produced more consistent across algorithms. study found all performed worse when derivation cohort. Further, there little difference between contained pharmacogenetic coefficients containing just non-pharmacogenetic coefficients. inconsistency results suggests unaccounted population specific factors cause performance. Better methods take into account inter- intra-individual variability, at initiation phases treatment, are needed.