Why Is Warfarin Such a Dangerous Drug?

Despite having a propensity to cause bleeding, vitamin K antagonists (VKAs) such as warfarin have been used widely to prevent thrombosis in patients with atrial fibrillation, venous thromboembolism, or cardiovascular disease. This practice is supported by the results of many randomized, controlled trials showing that benefits of thrombus prevention trump risks for bleeding. However, many patients who receive VKAs experience major hemorrhages, an outcome that has led investigators to question whether the populations in these pivotal trials accurately represent patients usually seen in practice.

To determine whether the type of patient seen in ordinary practice was excluded from clinical trials of VKA because of comorbidities or concurrent therapies and whether patient-exclusion criteria might have inadvertently skewed the risk-benefit ratios to favor VKA usage, Dutch researchers performed a 4-year case-control study. Cases consisted of approximately 1000 patients (mean age, 67; 57% male) who were admitted to a university hospital with VKA-associated bleeding (mean duration of VKA therapy, 3.2 years). Bleeding was categorized as gastrointestinal (60%), intracranial (15%), urogenital (15%), and other (10%). Controls, matched by age, sex, and VKA-therapy duration, had no hemorrhages and were hospitalized for respiratory or urinary tract infections. Compared with the control group, the case group had a significantly higher mean international normalized ratio (INR; 4.1 vs. 2.9), a higher incidence of INRs >5 (30% vs. 4.5%), higher mean creatinine levels (5.9% vs. 2.7%), and a higher incidence of liver failure (2.4% vs. 1.0%).

A greater proportion of cases than controls (40% vs. 23%) met at least one exclusion criterion: Unsuitability for anticoagulation therapy and use of restricted medications (such as aspirin, nonsteroidal anti-inflammatory drugs, and clopidogrel) were most common, followed by cardiac comorbidity, previous stroke, cancer, and risk factors for thrombosis. As such, these patients likely would not have been eligible for the trials of the pertinent indication for VKA therapy. Moreover, compared with patients who met no exclusion criteria, those who met one exclusion criterion had 3-fold higher risk for bleeding, and those who met three or more criteria had 15-fold higher risk. The authors concluded that, for patients who deviated from those in the pivotal VKA trials, the balance between risks and benefits of VKA therapy was shifted toward a less-favorable safety profile.