to the editor: We appreciate the recent article by Drs. Horton and Bushwick on warfarin therapy.¹ We would like to add our comments based on a recent case we shared.

A 52-year-old man presented with severe epistaxis of two hours' duration. In 1996, he had undergone surgery to replace a congenital bicuspid aortic valve. The patient's medication included warfarin, in a dosage of 12.5 mg per day. He denied taking any other medication or herbal remedies, receiving direct trauma to the nose, or having hematuria, back pain or hematochezia. On physical examination, he was hemodynamically stable and neurologically intact, although he had bleeding from the left nostril.

Laboratory studies demonstrated a prothrombin time (PT) of 24.0, an International Normalized Ratio (INR) of 4.65 and a hemoglobin level of 150 g per L (15.0 g per dL). Local measures to control the bleeding failed. Hemostasis was ultimately achieved by placing anteroposterior packing soaked in bacitracin, oxymetazoline and 4 percent cocaine solution. Anticoagulation was reversed with vitamin K, 2.5 mg subcutaneously, and 2 U of fresh frozen plasma. By the next morning, the PT had nearly normalized to 15.0 and the INR was 1.71.

The patient declined readmission for heparin anticoagulation therapy and was followed daily in our clinic. The lowest hemoglobin level was 102 g per L (10.2 g per dL). Nasal packing was removed after five days without incident, and the patient was restarted on his usual dosage of warfarin. His response was still subtherapeutic (PT = 14.8; INR = 1.67) when he returned home five days later.

We identified two major teaching points in this case. The first was that the reversal of anticoagulation with fresh frozen plasma and vitamin K was overly aggressive, since the bleeding was controlled. Patients with prosthetic valves are at high risk for systemic embolism, necessitating higher INR values (2.5 to 3.5). By administering vitamin K and plasma, coagulopathy was corrected at the cost of increasing the risk of systemic embolism. The cost of fresh frozen plasma and the risk of infection are factors to consider. A more elegant solution, as outlined by Drs. Horton and Bushwick, would have been to hold off the warfarin therapy, follow the PT and INR values, and restart warfarin therapy at a lower dosage while watching for bleeding.

The second point is that it is essential to obtain a comprehensive history to identify the cause of the elevated INR. Our patient revealed that he had been eating 1 to 2 lb of carrots a day as a substitute for smoking. He had abruptly stopped eating the carrots one week before admission, when he had flown out to visit his daughter. This history explains the high maintenance dose of warfarin, as well as the development of a supratherapeutic INR. The patient is currently doing well taking 12.5 mg of warfarin and eating 1 lb of carrots per day.

in reply: Bleeding complications related to anticoagulation are a source of concern. The case presented by Drs. Beanland and Dorsey refers to the key discussion points of reversal and dietary education regarding vitamin K.

In the case presented, the patient experienced epistaxis at an International Normalized Ratio (INR) of 4.65 that required treatment with nasal packing. Recent literature suggests that if rapid reversal were necessary, the administration of intravenous vitamin K, fresh frozen plasma and prothrombin complex concentrate (depending on urgency) would be indicated. Rapid reversal with intravenous vitamin K via slow infusion is safe, effective and more rapid than reversal with subcutaneous vitamin K; a minimum of six to 12 hours is necessary to evaluate responsiveness.¹ The physician should also consider that maximal response to vitamin K may not be noted until 24 to 48 hours after administration.

Administration of oral vitamin K (in doses of 2 to 4 mg) would have been an appropriate choice in this case, as the patient was hemodynamically stable and did not require either complete reversal to an INR of 1.0 or a rapid reduction to an INR of less than 1.5 (i.e., to go to surgery).¹ Based on our experience, we believe administering 2.5 mg of oral vitamin K and holding two doses of warfarin would have resulted in an INR of less than 2.0 in this patient (baseline INR of 4.65) within 48 hours (unpublished data). Our experience suggests that this patient might have had an INR in the 2.0 to 3.0 range at 24 hours with this treatment approach and would not have been resistant to the effects of warfarin administration. INRs greater than 9.0 require treatment with larger doses of oral vitamin K (5 mg) for adequate INR reduction by 24 to 48 hours.

The second point in this case relates to dietary intake of vitamin K. This patient's daily warfarin dose was significantly higher than that used in a typical patient between 50 and 59 years of age. The normal warfarin dose in this age group is approximately 5 mg per day.² When evaluating such patients, it is helpful to identify possible reasons for this finding so that adverse events can be avoided.³ In this case, initial education and evaluation would have identified the confounding dietary factors. Unfortunately, we often stress “green leafy vegetables” and minimize discussions relating to other sources of vitamin K (cabbage, cauliflower, tuna, etc.).

While there are no cases reported in the medical literature of the interaction between carrots and warfarin, carrots are a recognized source of vitamin K. The normal serving size of carrots is approximately 78 g, which contains 12 mg of vitamin K.⁴ The patient reported eating 1 to 2 lb of carrots daily or approximately 339 mg of vitamin K (339 mg vitamin K per 1 lb of carrots). This alone exceeds the total recommended daily allowance for men (80 mg) by a factor of 4. It also demonstrates that if any diet is maintained consistently, an appropriate dosage of warfarin can be identified and is safe, as long as the patient is warned to (1) avoid major dietary changes or (2) notify those responsible for warfarin monitoring when dietary changes are anticipated.