The trusted source for
healthcare information and
Abstract & Commentary
Synopsis: A warfarin initiation nomogram that uses 10 mg as a starting dose achieves a therapeutic INR more rapidly than one using a 5-mg starting dose.
Source: Kovacs MJ, et al. Ann Intern Med. 2003;138:714-719.
A patient presenting with a deep-vein thrombosis (DVT) or pulmonary embolism (PE) commonly is started on heparin, either regular or low molecular weight, by subcutaneous injection or intravenous administration, usually for 5 days or so. Warfarin is added and the dose adjusted periodically to achieve the desired international normalized ratio (INR). Once achieved, the heparin is stopped, the INR monitored every so often, and the warfarin dose adjusted to keep it in a predetermined therapeutic range. Since heparin is more costly and more difficult to administer, getting to a therapeutic INR quickly has appeal. The issue of what dose of warfarin to start with has been argued for several years. Too high a starting dose could result in overanticoagulation and hemorrhage. Too low a dose could result in a prolonged initiation and extension of the thrombosis. Many physicians prefer the 5-mg starting dose and others the 10-mg dose. A study by Crowther concluded that the 5-mg dose was safer and as swift as the 10-mg dose.1
Building on an earlier trial that favored the 10-mg dose,2 Kovacs and associates conducted a randomized, double-blind, controlled study of 2 warfarin initiation nomograms (5 mg vs 10 mg) in 4 Canadian thrombosis clinics. They randomized 210 consecutive outpatients with either a DVT or a PE. The diagnostic criteria, while not specified, were described as objective. Patients were excluded if their baseline INR > 1.4; if they were thrombocytopenic; if age was < 18 years; if they required hospitalization; if they had received warfarin in the last 2 weeks; or if they were judged to be at high risk for hemorrhage. Nine patients were excluded; the remaining 201 were randomized into 2 groups. There were 104 in the 10-mg dose group and 97 in the 5-mg group. The 2 groups were statistically similar in age (about 55.5 years), proportion with cancer (25% vs 23 %), and weight (about 83.5 kg). There were proportionately more men in the 10-mg group (62.5% vs 48.4%). All patients received at least 5 days of low-molecular-weight heparin). Blood was drawn daily for INR, beginning on day 3. Subsequent doses of warfarin in both nomograms were administered based on the INR values; however, the 10-mg nomogram did not rely on the day 4 value. The patients were followed for 90 days.
Data analysis was by intention to treat. The primary end point was time to a therapeutic INR, which Kovacs et al defined as >1.9. Patients in the 10-mg group were at end point at 4.2 days compared to 5.6 days in the 5-mg group. More patients in the 10-mg group were at end point at day 5 (83% vs 46%). These findings were statistically significant. Three patients in the 10-mg group experienced venous thromboembolism during follow-up. No patients in the 5-mg did, but this difference did not reach statistical significance. One patient in each group had a major bleeding episode, and 1 patient in the 5-mg group died during the study period. Neither of these was statistically significant. In the first 4 weeks, the patients in the 10-mg group had an average of 8.1 INR tests compared to 9.1 in the INR group; this result just barely made statistical significance. Both groups had equivalent numbers of patients who were overanticoagulated during the first 4 weeks. The researchers defined 3 levels of overanticoagulation: INR = 3.0, INR = 4.0, and INR = 5.0. Many patients in both groups were overanticoagulated by the first definition sometime during the first 4 weeks (78% in the 10-mg group vs 87% in the 5-mg group). The percentages for the second and third definitions were 9% vs 11% and 5% vs 6%, respectively.
Comment by Allan J. Wilke, MD
Outpatient treatment of venous thromboembolism is attractive for a variety of reasons, not the least of which are economic and patient satisfaction. These results should only be applied to the outpatient setting; Kovacs et al’s previous study2 suggested that inpatients are more sensitive to warfarin. The main finding of this study is that a nomogram that uses 10 mg of warfarin as its starting dose is safe and results in speedier anticoagulation than a nomogram that uses 5 mg as a starting dose. The finding is statistically significant, but is it clinically significant? From a patient’s perspective, one less blood draw and fewer heparin injections would be welcomed. However, a patient with a venous thromboembolism can expect to be treated for 3 months or more, making the reduction in blood draws and heparin injections small indeed. From the physician’s perspective, there may be one less day of laboratory review and decision-making, but, again, in the long run this appears to be a modest improvement. Both nomograms are safe, but not simple, and would require having them in front of you as you make dosing decisions. Ideally, someone will write a program for a handheld computer that would make the decision-making automatic.
Dr. Wilke, Assistant Professor of Family Medicine, Medical College of Ohio, Toledo, OH, is Associate Editor of Internal Medicine Alert.
1. Crowther MA, et al. Arch Intern Med. 1999;159:46-48.
2. Kovacs MJ, et al. Haemostasis. 1998;28:62-69.