Author + information
- Walter Ageno, MD⁎ (, )
- David Garcia, MD,
- Mauro Silingardi, MD,
- Matteo Galli, MD and
- Mark Crowther, PhD
- ↵⁎U.O. Medicina Interna, Ospedale di Circolo, Viale Borri 57, 21100 Varese, Italy
To the Editor:A strong relationship exists between the risk of hemorrhage and the degree of prolongation of the international normalized ratio (INR) in patients receiving warfarin, particularly when the INR exceeds 6.0 (1). Simply withholding warfarin is likely to be associated with a low, but clinically important, risk of major or life-threatening hemorrhage (1,2). The oral administration of low doses of vitamin K has been shown to safely and effectively reduce the INR toward the normal range in patients presenting with INR values >4.5 (3–5). In patients with mechanical heart valves, there is a safety concern with the use of vitamin K because overcorrection of the INR may cause thrombosis. The purpose of this study was to demonstrate that the oral administration of low-dose vitamin K is an effective and safe treatment for patients with mechanical heart valves who present with excessive, warfarin-associated, anticoagulant effect.
Patients with mechanical heart valves who were receiving warfarin therapy and who presented with INR values between 6.0 and 12.0 were allocated randomly to receive either 1 mg of vitamin K administered orally (Phytomenadione, Konakion, Roche, Milan, Italy, or Phytonadione, Mephyton, Merck & Co. Inc., West Point, Pennsylvania) or no treatment. Patients were excluded in the presence of elevated INR value determined more than 12 h before screening; high risk of hemorrhage; major non-orthopedic surgery within the past week; thrombocytopenia (platelet count <50 × 109/l); indication for acute reversal of anticoagulation; termination of anticoagulation; receipt of vitamin K or plasma in the 48-h period before enrollment; allergy or sensitivity to vitamin K or plasma or other contraindication to vitamin K therapy; severe liver disease; geographic inaccessibility or inability to comply with requirements for blood work; or failure to obtain informed consent.
The day of randomization and study drug administration was defined as day 0. All patients underwent INR measurement on the day after study intervention (day 1). We determined INR values using standard local laboratory techniques. In particular, the following thromboplastin reagents were used: Innovin by Dade Behring (Milan, Italy; ISI 1.03) at study locations in Varese and in Reggio Emilia, Italy, and Neoplastine by Diagnostica Stago (Parsipanny, New Jersey; ISI 1.27) at the study location in Albuquerque, New Mexico.
Warfarin therapy was restarted once the INR was in the target range, as determined by the attending physician. Patients were clinically followed-up on day 5 ± 1 and were subsequently contacted by telephone or in person, 4 ± 1 week after enrollment to determine whether thromboembolic or hemorrhagic events had occurred.
Analysis was based on the principle of intention to treat. For the primary analysis, the proportion of patients with INR values between 2.3 and 3.7 on day 1 was determined and compared using the Fisher exact test. The primary safety analysis was an analysis of the number of patients who received vitamin K who developed an INR value of <1.8 on day 1 and also was performed using the Fisher exact test. The value of 1.8 was chosen as a “lower limit” for the INR because this represents a value below which many physicians would be concerned that there is a substantially increased risk of thromboembolism. Baseline data were compared using the Student ttest.
We calculated that we would need to enroll 48 patients per arm to reliably demonstrate that oral vitamin K was more effective than simple warfarin withdrawal. We assumed a binomial distribution and conservative “success” rates of 40% and 10% in the “oral vitamin K” and “no treatment” arms, respectively, because a sample size of 96 would be adequate to achieve power of 90% and alpha = 0.05. Institutional review board approval was obtained at all participating centers.
Between September 2002 and October 2004, 59 patients provided informed consent for the study. In October 2004, the study was prematurely stopped because a multinational study examining the utility of oral vitamin K, using clinical end points as the primary outcome measure, was initiated. Of the 59 patients, 29 were assigned randomly to 1 mg of oral vitamin K and 30 were assigned to withhold warfarin only. Baseline characteristics are summarized in Table 1,and results are summarized in Table 2.No major bleeding events occurred in either group. There was one episode of minor bleeding (epistaxis) in the vitamin K group and no episodes in the control group. One patient died in the vitamin K group because of acute pulmonary edema; no deaths were recorded in the control group.
This randomized clinical trial is the first to examine the efficacy of oral vitamin K in patients with mechanical heart valves. Compared with “no treatment,” 1 mg of oral vitamin K more commonly returned prolonged INR values to the therapeutic or near-therapeutic range within one day. Given that major bleeding can occur in as many as 4% of patients who present with an INR >6.0 who are treated with simple warfarin withdrawal (1), this finding may be clinically relevant. In our study, only 3.4% of patients treated with low-dose oral vitamin K had INR values >5.0 at 24 h after administration of the study drug. In contrast, 43.3% of the patients who simply withheld warfarin had INR values >5.0 at 24 h after study enrollment (p < 0.001).
The administration of vitamin K produced what we defined as an over-reversal in approximately 10% of patients. The difference between the two groups in the proportion of patients with INR values <1.8 was not statistically significant, and none of the patients in this study had thrombotic events during follow-up. Nevertheless, such a high proportion of patients with an INR less than the therapeutic range raises the concern that low-dose oral vitamin K may increase the risk of thromboembolism in this population; we anticipate that the results of our ongoing multinational study will determine whether oral vitamin K, administered in this setting, is associated with thromboembolism.
Our study had two important limitations. First, we used the INR, rather than clinical outcomes, as the primary outcome measure. Although the INR is a validated surrogate marker for bleeding and thrombotic complications in stably anticoagulated patients, its association with these complications is less clear when the INR is changing rapidly (for example, during warfarin initiation or after administration of oral vitamin K). Second, this study was terminated prematurely, and the enrollment target was not reached. Nonetheless, the observed difference in the primary end point was statistically significant. Even if target enrollment had been reached, it is unlikely that our findings would have had sufficient power either to detect differences in clinical outcomes or to demonstrate a statistically rigorous association between oral vitamin K administration and an INR <1.8 on day 1.
In summary, our study provides additional support for the hypothesis that the administration of 1 mg of oral vitamin K is very effective for the treatment of asymptomatic warfarin-associated coagulopathy in patients presenting with INR values >6.0. Our enthusiasm for this treatment is, however, tempered by the observation that about one in 10 patients who receive vitamin K will have an INR of <1.8 on the day following vitamin K administration; it is possible that these patients are at risk for avoidable thrombosis.
Please note: Dr. Crowther is a Career Investigator of the Heart and Stroke Foundation of Canada.
- American College of Cardiology Foundation