Clarification of Optimal Anticoagulation Through Genetics - COAG


Although warfarin remains the most commonly used agent for anticoagulation, it has a narrow therapeutic index, with wide variation among patients in the daily doses required. Although a number of factors are responsible, pharmacogenomic studies have identified that two genes, CYP2C9 (cytochrome p450 isoform) and VKORC1 (vitamin K epoxide reductase complex subunit 1), account for a large portion of this variability. The current trial sought to study the efficacy of genotype-guided dosing of warfarin compared with standard management in patients requiring warfarin.


A strategy that incorporated both clinical variables and genotype data would be superior to a strategy incorporating clinical variables alone for achieving and maintaining therapeutic international normalized ratio (INR) levels in patients requiring warfarin for anticoagulation.

Study Design

  • Blinded
  • Parallel
  • Placebo Controlled
  • Randomized
  • Stratified

Patient Populations:

  • No prior treatment with warfarin
  • Clinical condition requiring anticoagulation with warfarin INR 2.0-3.0

    Number of enrollees: 1,015
    Duration of follow-up: 28 days
    Mean patient age: 58 years
    Percentage female: 49%

Primary Endpoints:

  • TTR (INR 2.0-3.0) during the 28 days after initiation of warfarin therapy

Secondary Endpoints:

  • Composite of any INR value ≥4.0, major bleeding, thromboembolism in the first 4 weeks
  • Time to first therapeutic INR
  • Time to determination of a maintenance dose
  • Time to an adverse event

Drug/Procedures Used:

Patients were randomized in a 1:1 fashion to a strategy that incorporated both clinical variables and genotype data or a clinically guided dosing strategy alone. Genotyping for CYP2C9*2, CYP2C9*3, and VKOR1C (–1639G →A) alleles was performed. Prespecified algorithms were utilized in both arms for the first 5 days of treatment initiation. After the 5-day initiation, the dose was adjusted during the first 4 weeks based on standardized adjustment techniques.

Principal Findings:

A total of 1,015 patients were randomized, 514 to the genotype-guided group and 501 to the clinically guided group. Baseline characteristics were fairly similar between the two arms. Approximately 22% of all patients were recruited for stroke prevention in patients with atrial fibrillation, 58% for venous thromboembolism, and the rest for other or multiple indications. As part of the study design, 27% of enrolled patients were black. Wild-type VKORC1 genotype was noted in 48%, heterozygous in 40%, and homozygous for G-A mutation in 12% of patients. Wild-type CY2C9*2 and CYP2C9*3 genotypes were noted in 83% and 91% of patients, respectively. Since genotype information was not point of care, it was available for 45% of patients before the first warfarin dose.

The primary endpoint of mean % of target therapeutic range (TTR) at 4 weeks was low, but similar in the genotype-guided arm compared with the clinically guided arm (45.2% vs. 45.4%, p = 0.91). Among black patients, mean % TTR was lower in the genotype-guided arm (35.2% vs. 43.5%, p = 0.01). The composite of any INR ≥4.0, major bleeding, or thromboembolism (20% vs. 21%, p = 0.93) was also similar.


The results of the COAG trial indicate that a strategy incorporating genotype and clinical variables is not superior to one incorporating clinical variables alone in increasing the amount of time spent in TTR over 4 weeks in patients requiring warfarin for anticoagulation. The average TTR in this trial at 4 weeks is consistent with that observed in other trials.

Genotype-guided management is not endorsed by current guidelines due to largely negative results from other similar randomized controlled trials. However, the EU-PACT Warfarin trial demonstrated better safety and efficacy with a point-of-care genotype-guided strategy. One reason for the difference in results between the COAG and EU-PACT Warfarin trials could be that the latter trial used fixed-dose warfarin in the comparison arm, as compared with a clinical variable-based algorithm in the COAG trial.


Kimmel SE, French B, Kasner SE, et al., on behalf of the COAG Investigators. A Pharmacogenetic Versus a Clinical Algorithm for Warfarin Dosing. N Engl J Med 2013;369:2283-93.

Presented by Dr. Stephen E. Kimmel at the American Heart Association Scientific Sessions, Dallas, TX, November 19, 2013.

Clinical Topics: Anticoagulation Management, Arrhythmias and Clinical EP, Pulmonary Hypertension and Venous Thromboembolism, Vascular Medicine, Anticoagulation Management and Venothromboembolism, EP Basic Science

Keywords: Mutation, Cytochrome P-450 Enzyme System, Stroke, Homozygote, Warfarin, Vitamin K Epoxide Reductases, Venous Thromboembolism, Genotype, Pharmacogenetics, Hemorrhage

< Back to Listings