Reversal of warfarin: case-based practice recommendations.

Case Presentation: A 74-year-old woman presents to the emergency department with bruising. She takes warfarin for atrial fibrillation. She has recently begun taking trimethoprim/ sulfamethoxazole. Her international normalized ratio (INR) is reported as 8.6. Supratherapeutic INR values are common in warfarin-treated patients. In this case, the antibiotic is the likely cause, but it is not unusual for an INR measurement to exceed 3.0 without explanation. Irrespective of whether a cause for the INR increase can be identified, the patient should be interviewed and examined to ensure she is not bleeding. For an asymptomatic patient whose INR is 5, warfarin should be withheld for at least 1 dose, and close follow-up monitoring should be arranged. This patient’s INR will return to the therapeutic range more quickly if she receives low-dose oral vitamin K (as opposed to simple warfarin withdrawal).1 Low-dose oral vitamin K is often considered in such situations because INR elevations like the one described here can be quite alarming to both the patient and the clinician. However, there is uncertainty about the short-term risk of major bleeding in such a patient. In one observational cohort of 1104 warfarintreated asymptomatic patients with a single INR value between 5.0 and 9.0 (90% of whom were managed with simple warfarin withdrawal), only 0.96% experienced major hemorrhage within 30 days.2 However, an earlier observational study of 114 asymptomatic patients taking warfarin with an INR 6.0 managed without vitamin K reported major bleeding in 5 patients (4.4%; 95% confidence interval, 1.4%–9.9%) during 14 days of followup.3 To address this uncertainty, we randomized 355 nonbleeding warfarintreated patients whose INR was 5.0 and 9.0 to receive either 1.25 mg of oral vitamin K or placebo. Although INR correction was more robust for the vitamin K–treated patients, the rate of major bleeding was low in both groups at 7 days (no major bleeds in either group) and at 90 days (2.5% with vitamin K versus 1.1% in the placebo group, P 0.22).1 On the basis of these results, we suggest that for asymptomatic warfarin-treated patients whose INR is 5.0 and 9.0, low-dose oral vitamin K will more quickly lower the INR (and possibly allow earlier resumption of treatment) but would not be expected to lower the risk of major bleeding. Advanced age, decompensated heart failure, low weekly warfarin dose, and active malignancy are independent predictors of slow INR decay4; patients with these characteristics or clinical features to suggest a higher-than-average bleeding risk may benefit the most from a more rapid INR correction because they may be exposed to a higher risk of bleeding.5 Although the patient is asymptomatic, the treating physician decides to administer vitamin K. What dose and route of administration should be used? For most warfarin-treated patients who are not bleeding and whose INR is 4.0, oral vitamin K (in doses between 1 and 2.5 mg) will lower the INR to between 1.8 and 4.0 within 24 hours.6 Intravenous vitamin K can lower the INR more quickly than oral vitamin K, but at 24 hours, intravenous and oral vitamin K produce similar