A new study presented at the American Heart Association meeting in Orlando finds that more than one in five elderly US veterans are hospitalized for bleeding after starting warfarin. This high rate of serious bleeding complications in a real world setting surprised and even disturbed the study authors.
Anticoagulants are the proverbial double-edged sword. Striking the right balance between benefit and harm is always difficult. The major danger with both warfarin and the new oral anticoagulants (NOACs) is, of course, bleeding. But although bleeding has been very carefully examined in clinical trials, far less is known about its effects in the real world outside the idealized setting of these trials.
Taking advantage of the comprehensive data contained in the VA system, John Dodson (New York University) and colleagues studied 31,951 veterans with atrial fibrillation (AF) who were at least 75 years of age. After the initiation of warfarin therapy 22.8% (7,288) were hospitalized for bleeding, with a median time to the bleeding event of 1.6 years.
The rate of major bleeding in clinical trials for both warfarin and NOACs has been much lower. It should be noted, though, that the findings are not directly comparable, since the clinical trials used very rigorous criteria to define bleeding while the new study relied on available VA data. Also, Dodson said in an interview, “this type of patient typically doesn’t make it into the trials.” In general VA patients are older and have more comorbidities than the patients in the clinical trials.
Dodson said the findings are concerning. Even though the criteria for bleeding complications are not as rigorous as in the clinical trials, he emphasized that if the bleeding episodes “are serious enough to warrant hospitalization, they are meaningful for patients.”
Many patients had more than one bleeding episode. Gastrointestinal bleeds accounted for more than half the bleeds, followed by genitourinary bleeds in nearly 20% and intracranial bleeds in nearly 15%. 5% of patients were hospitalized for stroke over the study period.
Some of the factors associated with increased bleeding risk were being nonwhite or having high blood pressure, diabetes, heart failure, coronary disease, COPD, or chronic kidney disease. People taking 4 or more drugs were also at elevated risk.
The authors noted that the study was performed before the widespread use of the new oral anticoagulants; it is unclear whether NOACs have diminished the seriousness of the real world bleeding found in the study.
Dodson said it was likely that some of the bleeding associated with warfarin use might be prevented with a NOAC, since some of the bleeding was linked to the warfarin-specific problem of labile INR. Dodson said that he is hoping to perform a similar study to see how these real world patients do after initiation of a NOAC.
Dodson said that treating elderly people with AF is a common clinical problem, since they are the group most likely to develop AF. “We do a good job of defining the benefit but we haven’t really understood the risk side of the equation.”