For Physicians

A Close Look at Warfarin Mangement

Warfarin is prescribed to an estimated 2 million Americans, but a number of factors make this drug difficult to manage: delayed onset of action, narrow therapeutic margin, drug and dietary interactions, variations in patient sensitivity, and issues regarding how best to reverse its effect¹. Even with expert management, patients are within therapeutic range only 60% of the time² and 14% of patient time is spent with INRs higher than appropriate ⁸.

An increased INR is an independent predictor for major hemorrhage. A contemporary study of bleeding for patients on warfarin reported rates of 0.25, 1.1, and 6.2 per 100 patient years for fatal, major, and minor bleeds respectively ². Bleeding risk is highest during the first 90 days³. A major hemorrhage on warfarin potentially is a very serious event as confirmed by a meta-analysis that found a case fatality rate of 13.4%. The most serious is intracranial hemorrhage, which has an estimated 60% mortality ^1,4.

Patients commonly exhibit wide variations in sensitivity to warfarin. It is well-known that achieving and maintaining the proper therapeutic dose is influenced by clinical factors such as liver dysfunction, drug interactions of which there are many, an increase or decrease in dietary intake of vitamin K containing foodstuffs, and eradication of vitamin K producing intestinal bacteria by antibiotics.

Warfarin sensitivity also is influenced by genetically determined levels of certain enzymes. The two most important are CYP2C9, an enzyme responsible for the hepatic clearance of warfarin, and the enzyme VKORC1 that converts oxidized vitamin K to the active reduced form necessary for gamma-carboxylation of K-dependent clotting factors II, VII, IX, and X. VKORC1 is the crucial target inhibited by warfarin.

Individuals with certain specific polymorphisms of CYP2C9 are sensitive to warfarin because these alleles result in reduced enzyme activity, impairing drug clearance and thus producing higher drug levels. Such patients require lower doses to achieve a therapeutic effect ¹. On the other hand, warfarin resistance has been related to rare coding region genetic mutations of VKORC1. However, variations in warfarin dosage in the general population are more often related to certain polymorphisms of VKORC1, which have been used to stratify patients into sensitive (A/A), intermediate (A/B), and resistant groups B/B) based on changes in the VKORC1 enzyme ⁶. There is evidence that better INR control results when genetic information is incorporated into dosing decisions. Nevertheless, this type of testing has not been widely used because of the delay in obtaining results, expense, and absence of studies demonstrating improved clinical outcomes using pharmacogenetic-based dosing ⁶.

Oral vitamin K1 (plant-derived phytonadione) when used for non-bleeding or minor bleeding patients in doses of 1.0 to 2.5 mg does not produce warfarin resistance, skin reactions, or anaphylactic reactions. Subcutaneous vitamin K, although widely used, has unpredictable effects on the INR, and the only study comparing oral to subcutaneous favored the former ⁸. Intravenous vitamin K1 has been associated with anaphylactic reactions, although most cases occurred when large doses were given rapidly with little dilution, and also when castor oil used to be part of the formulation ².

No studies have directly compared the effects of differing doses of vitamin K. A recent analysis of randomized controlled trials of asymptomatic non-bleeding patients found that oral low dose vitamin K was as effective as IV doses when INR values were compared at 24 hours. Importantly, over-correction occurred infrequently when small oral doses ranging from 1.0 to 2.5 mg were administered ⁸. Furthermore, if a more rapid response is desirable with the goal of just returning to therapeutic range, IV doses of 0.5 mg seemed to be optimal ².

The dilemma in this situation is how to return the INR to the therapeutic range while not over-correcting, which in turn produces resistance to subsequent warfarin. The most common strategy is simply to withhold warfarin and allow the INR to drift into the desired range. Two studies of patients with INR > 6 found that the risk of hemorrhage was low (0.6%) with withdrawal of therapy only, and a third study confirmed that the 30 day incidence of major bleeding was low (1.3%), unless the initial INR was > 9, when major bleeding developed in 9.6% of patients. These authors concluded that active therapy was indicated in patients with very high INRs, defined in this series as > 9 ⁷. A randomized clinical trial of patients with INR values of 6.0 to 12.0 directly compared withholding warfarin, to withholding plus treatment with 1.0 mg of oral vitamin K. In the group receiving low-dose of vitamin K, the mean INR dropped from 7.22 to 2.99 within 24 hours, while in the withdrawal alone group the INR decreased from 7.72 to 5.23 ². A similarly designed trial found that despite a more rapid INR decline on low dose K, the incidence of major bleeding was the same when compared to withdrawal alone ⁷. Therefore, low dose oral vitamin K is effective and produces more rapid reversal of potential bleeding, therefore allowing earlier reinstitution of warfarin by avoiding over-reversal.

Hemorrhage can occur as a complication even with the INR in the therapeutic range, but is much more likely when the INR is elevated. In the presence of major bleeding the goal of management is to completely correct the INR. Decisions should be individualized as they depend on severity of bleeding, need for a rapid response, and site of hemorrhage (brain being the most serious), as well as the degree of INR elevation. All patients should receive vitamin K and coagulation factor replacement. The effect of IV vitamin K is more rapid than other routes; reduction of the INR begins within 2 hours and it generally normalizes within 24 if hepatic function is normal and the dose is large enough (small doses in this setting are inappropriate). Vitamin K should be infused slowly over 30 minutes at doses of 2.5-10 mg.

The most common method of coagulation factor replacement in the U.S. is transfusion of fresh frozen plasma (FFP), which supplies factors II, VII, IX, and X. At the usual dose of 15 mL/kg reversal is fairly rapid but incomplete. When clotting factors are very low and the INR very high, increasing the titer of clotting factors to hemostatic levels may require clinically intolerable large volumes of FFP. Other disadvantages are that factor IX may remain lower than required for hemostasis for an extended period ⁹, there is a delay for thawing, and a risk (very low) of transfusion-transmitted infection.

An alternative to FFP is prothrombin complex concentrates (PCC), which are intermediate purity pooled plasma products containing factors II, IX, and X, with variable amounts of VII and the natural anticoagulant proteins C and S. There are no randomized, controlled clinical trials proving that PCCs improve the clinical outcome for seriously bleeding patients, but observational and retrospective studies clearly indicate that PCCs are much more rapidly effective in reducing the INR than either IV vitamin K or FFP ¹⁰. In one study, complete correction of the INR occurred in 28 of 29 patients within 15 minutes 9, and this has been confirmed by others. The optimal initial dose for correction has not been standardized, although doses as low as 8.8 U/kg have been successful for INR < 5. This dose probably is inadequate for higher INRs where doses commonly are 20-25 U/kg, and may increase to 50 U/kg. The INR must be checked within 30-60 minutes, to determine if more treatment or higher doses are needed. Also, vitamin K usually is administered simultaneously to prevent a late rise in INR within the next 12-24 hours. Thromboembolic complications can occur with PCC and this fact probably accounts for limited usage. However, a recent review of 14 studies with a total of 460 patients found only 7 such events, and retrospectively, several may not have been related.These authors and other experts have concluded that fear of thromboembolic events with PCC has been overestimated ^2,10.

Recombinant factor VIIa (rFVIIa) has been shown to rapidly correct the INR in over-anticoagulated patients, including those with CNS bleeding. However, this product does cause a significant number of thrombotic events. Since there is little evidence that it improves clinical outcome, most experts recommend that it be avoided until there is better quality evidence ².

References:

1. Francis C., ASH Educational Book, 2008, p.259
2. Dentali F., ASH Educational Book, 2008, p.266
3. Fitzmaurice D., BMJ, 2002, 325:828
4. Linkins L., Ann Int Med, 2003, 139:893
5. Ansell J., Chest, 2004, 126:204S
6. Rieder M. NEJM, 2005, 352:2285
7. Crowther M., Ann Int Med, 2009, 150:125
8. Wilson S., CMAJ, 2004, 170:821
9. Greaves M., Thromb Haemost, 1997, 77:477
10. Leissinger C., Am J Hemat, 2008, 83:137
11. Liumbruno G., Blood Transf, 2009, 7:325
This newsletter is a review and not meant as a guideline for medical treatment.