Warfarin dose calculator (IWPC 2009 algorithm)

Warfarin maintenance dose from the IWPC 2009 equation, the anchor for choosing a starting dose: age, height, weight, VKORC1 and CYP2C9, amiodarone. Then INR.

Updated 31 Aug 2026 · Methodology

How to use this calculator

Who this is for

This page addresses healthcare professionals: cardiologists, haematologists, internists, clinical pharmacologists and anticoagulation clinic staff. The calculation is not intended for patients to adjust their own dose.

What the calculation shows

The form reproduces the published IWPC 2009 algorithms (Klein TE et al., N Engl J Med 2009;360:753-764), derived in 4043 patients and validated in 1009. The output is the value the model returns for the parameters entered, not a prescription. Spread is wide: mean absolute error was 8.5 mg per week for the pharmacogenetic variant and 9.9 mg for the clinical one, and the models explain 43 % and 26 % of dose variability. Age enters the algorithm in decades, so 60 and 69 years give the same result.

What this calculator does not do

The calculation does not select a dose for an individual patient and does not replace the treating physician. It ignores liver and kidney function, the indication for anticoagulation, comorbidities, diet, and most drug interactions apart from amiodarone and four enzyme inducers. The model predicts a maintenance dose rather than a loading dose, and it never replaces INR monitoring. It is not validated in children or adolescents. Coumarin anticoagulants are contraindicated in pregnancy. For mechanical heart valves and INR targets above 3.0 the algorithm falls outside its training data, and a separate management protocol applies. Evigrade provides educational reference information: the calculator carries no CE marking and is not intended for diagnostic or therapeutic decision-making.

Spanish context

In Spain the standard vitamin K antagonist is acenocoumarol (Sintrom); warfarin covers roughly 5 % of patients on a VKA for non-valvular atrial fibrillation (PAULA subanalysis, Rev Esp Cardiol. 2015;68(12):1181-1182, PMID 26507538). IWPC algorithms were derived on warfarin and do not transfer to acenocoumarol, which has a different half-life and dose range. The ALDOCUMAR ficha tecnica in CIMA AEMPS makes no mention of genotyping before initiation, and ACCP recommends against routine pharmacogenetic testing (Grade 1B). Where the calculation and the AEMPS-approved ficha tecnica disagree, the ficha tecnica prevails.

INR monitoring is mandatory

Per the ALDOCUMAR ficha tecnica, INR is measured before treatment starts and then every 24 hours until the maintenance dose is established, weekly through the first month, and monthly thereafter. Dosing from then on follows the INR, not the calculation. A large starting dose raises bleeding rates without speeding up protection from thrombosis, which is why the label advises against it.

About this calculator

The IWPC 2009 algorithm was derived from 5052 patients across 21 sites and predicts the square root of the weekly maintenance dose, so the output of the equation is squared. The pharmacogenetic variant needs VKORC1 rs9923231 and CYP2C9; the clinical variant runs without genotype. Accuracy is modest for both: mean absolute error 8.5 mg per week for the pharmacogenetic model against 9.9 mg for the clinical one, with 43 % and 26 % of variance explained. The model was built on adults stable at INR 2.0–3.0 and does not extend to children, pregnancy, mechanical valves or target INR above 3.0. In patients of African ancestry CPIC 2017 advises against the genetic variant, because the panel covers only *2 and *3 alleles and the genotype-guided arm lost to the clinical arm in the COAG trial, so the calculator falls back to the clinical model. Spanish context: the standard vitamin K antagonist in Spain is acenocoumarol (Sintrom), and the IWPC equation does not transfer to it; a separate algorithm for the Spanish population was published as PGX-ACE 2016 (Tong HY et al., PLoS ONE 2016). The estimate never replaces INR monitoring: the first measurement comes before the first dose, and the local protocol then runs until the dose is stable.

Source

International Warfarin Pharmacogenetics Consortium (Klein TE et al.). Estimation of the warfarin dose with clinical and pharmacogenetic data. N Engl J Med. 2009;360(8):753-764. Coefficients from Supplementary Appendix 1, sections S1e (pharmacogenetic) and S1f (clinical).

Formula version: iwpc-2009-v1

Calculators

eGFR calculator online: CKD-EPI 2021 formula without race coefficientEstimated glomerular filtration rate by the updated CKD-EPI 2021 formula without race coefficient. Used in nephrology, cardiology, and endocrinology for CKD staging and drug dose adjustment in adults.
Child-Pugh Score (Liver Cirrhosis)Liver cirrhosis severity and survival prognosis assessment.
CHA₂DS₂-VASc calculator online: stroke risk in atrial fibrillationCHA₂DS₂-VASc score for annual ischemic stroke and systemic embolism risk in non-valvular atrial fibrillation. Includes 7 components: heart failure, hypertension, age, diabetes, prior stroke/TIA, vascular disease, female sex. Guides oral anticoagulation decisions.
Wells score for DVT online: deep vein thrombosis probability calculatorPretest probability of lower-extremity deep vein thrombosis using 9 clinical Wells criteria. Output: low, moderate, or high DVT probability – guides D-dimer and venous ultrasound decisions.
HAS-BLED (Bleeding Risk on Anticoagulation)Bleeding risk in patients on anticoagulation. Used alongside CHA₂DS₂-VASc.
Wells Score for Pulmonary EmbolismPre-test probability of pulmonary embolism using Wells criteria.

Drugs

Carbamazepine, Rivaroxaban, Phenytoin, Apixaban, Clopidogrel, Dabigatran, Rifampicin, Warfarin, Acetylsalicylic acid, Amiodarone, Phytomenadione, Fluconazole

About this calculator

Source

Related