Background      Patients with advanced disease may be prescribed warfarin (Coumadin) even while receiving specialist palliative care or hospice care. For many patients, the primary goal of anticoagulation with warfarin is to prevent embolic stroke due to dilated cardiomyopathy, atrial fibrillation, or prosthetic valves, until close to the time of death (see Fast Fact #236 for information on venous thromboembolism). This Fast Fact reflects expert opinion on managing warfarin near the end-of-life.

Physiology     Vitamin K is an important cofactor in the hepatic synthesis of coagulation proteins. Warfarin inhibits vitamin K epoxide reductase, effectively reducing the amount of active vitamin K in the body. The principal source of vitamin K is ingested food, particularly leafy greens, which is acted upon by intestinal bacteria. Patients who are not eating well absorb less vitamin K and will have a higher INR so will need less warfarin to maintain the same anticoagulant effect. Antibiotics and liver synthetic failure will also raise the INR requiring warfarin dose reduction. Therefore, monitoring the INR is required for patients taking warfarin to prevent an increased risk of bleeding (GI bleed and hemorrhagic stroke).

Measurement of effect     An INR (International Normalized Ratio) of 2-3 is the therapeutic range for most patients. This is a standardized measure of clotting that reflects the sum total of all of the clotting factors that are dependent on Vitamin K: II, VII, IX and X as well as the anticoagulant proteins C and S. The long half-life of about 50 hours (2 days) of prothrombin (Factor II) synthesized by the liver means the antithrombotic effect of a changed warfarin dose is not approached until 3 half-lives (5-6 days).

Common error    When informed of an INR that is too high or too low, the less informed clinician will change the daily dose of warfarin and ask for a repeat value for the INR in 1 or 2 days. If it is too high or too low, the dose is changed again. Since steady state is not reached before the change, a see-saw of increasing and decreasing doses and repeated measurements is a burden to patients and health care professionals and puts the patient at risk from an INR that veers wildly from subtherapeutic to supratherapeutic levels. 

When to deprescribe warfarin    This is a challenging clinical decision. The management of the INR by pharmacists in many settings makes it even more difficult to keep the risks/benefits of warfarin therapy balanced with the clinical assessment of the patient at home.  

• Most experts would agree that the risk of intracranial hemorrhage for a patient taking warfarin who insists on being ambulatory but wobbly enough to fall is greater than the benefit of preventing embolic stroke. The one exception could be patients with mechanical heart valves.  Consult cardiologists in these cases since the risk of embolic stroke can be quite in patients who are not anticoagulated.
• Similarly, when dietary intake drops 50% or more from a usual balanced diet (especially when intake is only favorites like ice cream and not vitamin K rich foods) or when eating is highly erratic, the risk of bleeding escalates exponentially if warfarin dosing continues.
• No warfarin should be given to the patient who develops liver failure. If anticoagulation is still desired, seek an alternative source of anticoagulation (e.g., apixaban, lovenox).
• The patient who does not agree to checking an INR should stop taking warfarin.
• Stop warfarin therapy when there is frank bleeding.

References 

1. National Guideline Clearing House. Management of warfarin therapy. http://www.guideline.gov/search/search.aspx?term=management+of+warfarin.  Accessed Sept 25, 2015.
2. Ebell MH. A systematic approach to managing warfarin doses. Fam Pract Manag 2005;12:77-83.
3. Camm JA. Managing anticoagulation for atrial fibrillation: current issues and future strategies. J Intern Med. 2013 Jan;273(1):31-41.
4. Lip GYH, Wrigley BJ, Pisters R. Anticoagulation versus placebo for heart failure in sinus rhythm. Cochrane Database Syst Rev. 2012 Jun 13;6:CD003336. doi: 0.1002/14651858.CD003336.pub2.

Authors’ Affiliations:  Hospice & Palliative Care, Kobacker House, OhioHealth, Columbus, OH (CvG); Medical College of Wisconsin, Milwaukee, WI (DEW); Dana-Farber Cancer Institute and Harvard Medical School, Boston, MA (JA).
Conflict of Interest: The authors have disclosed no relevant conflicts of interest.
Version History:  First published May 2014. Re-copy-edited in September 2015. Reviewed and updated again in September 2025.