Chief of Service Rounds: Should You Always Stop Anticoagulating a Bleeding Patient?

January 20, 2010

bellevueChief-of-service rounds is a new feature of Clinical Correlations.  Here we summarize Bellevue Hospital’s Chief of Service Rounds moderated by the Chief of Medicine, Nate Link, MD.  This multidisciplinary bimonthly conference focuses on a case that presents a diagnostic or treatment challenge.  A clinical question is posed at the end of the case and then answered using the principles of evidence based medicine.

Daria Crittenden , MD

Moderator: Nate Link, MD Associate Professor of Medicine, . GI consultant:  Gerry Villanueva, MD Clinical Assistant Professor of Medicine (Gastroenterology)

Case-in-brief: A 91 year old female with a medical history significant for hypertension, a remote history of congestive heart failure, and atrial fibrillation treated with warfarin, presented to the emergency room complaining of bright red blood per rectum for one day. Laboratory testing revealed an INR of 5.3 and a hemoglobin concentration of 6.8 gm/dL. She received four units of packed red blood cells and her hemoglobin subsequently normalized and remained stable. Upper endoscopy showed a normal esophagus and stomach. Colonoscopy revealed severe diverticulosis, but did not find an obvious source of bleeding.

Clinical question: For patients taking warfarin for stroke risk-reduction, should oral anti-coagulation be re-initiated following an episode of significant bleeding? Should an alternative treatment with an anti-platelet agent, such as aspirin, be used instead?

Response: The clinician must weigh the risk of stroke against the risk of bleeding when deciding whether or not to re-initiate an anti-coagulant or anti-platelet agent following an episode of significant bleeding. In this case,  he or she must also consider the advanced age of the patient.

Risk of stroke and amount of stroke reduction with warfarin: The commonly used CHADS2 scoring system is one of the best-validated clinical prediction rules for determining the risk of stroke in patients with atrial fibrillation. CHADS2 assigns one point each for the presence of recent congestive heart failure, hypertension, diabetes mellitus, and age greater than 75 years old. It assigns two points for a history of stroke or transient ischemic attack. A validation study published in JAMA in 2001 followed 1733 Medicare beneficiaries ages 65 to 95 with non-rheumatic atrial fibrillation [1]. None of the patients had been prescribed warfarin for the prevention of ischemic stroke. The study found that the annual ischemic stroke rate in this population increased with increasing CHADS2 scores. The scoring system quantified the increase in risk, such that the risk of stroke per 100-patients not receiving anti-thrombotic therapy increased by a factor of 1.5 for each 1-point increase in the CHADS2 score. For example, a CHADS2 score of 2 gives an annual ischemic stroke rate is 4.0%; for a score of 3 it is 5.9%, and so on up.

How much lower is this risk with warfarin? The cumulative evidence in the literature shows that administration of warfarin in patients with atrial fibrillation results in a relative risk reduction of ischemic stroke of approximately two-thirds [2, 3, 4]. This means that a patient with a 4% baseline risk of stroke will reduce his risk to 1.3% if he is prescribed warfarin.

Aspirin vs warfarin for stroke prevention and bleeding risk in the elderly: Mant and colleagues recently investigated the efficacy of warfarin compared to aspirin for stroke prevention in patients older than 75 years with atrial fibrillation [5]. Investigators randomized 973 elderly patients in a primary care setting with atrial fibrillation to receive either warfarin (goal INR 2-3) or aspirin 75 mg daily and followed both groups for an average of 2.7 years. Primary endpoints were first occurrence of fatal or disabling stroke (ischemic or hemorrhagic), intracranial hemorrhage, or clinically significant arterial embolism. Patients receiving warfarin had an annual risk of ischemic stroke of 0.8%. In contrast, patients receiving aspirin had a much higher annual risk of 2.5%. The risk of hemorrhagic stroke was remarkably similar in the two groups (0.5% in the warfarin group vs 0.4% in the aspirin group). Other intracranial hemorrhage and extracranial bleeding also occurred in similar rates in both groups. Overall, investigators determined that administration of warfarin resulted in roughly a 2% absolute risk reduction reduction in annual ischemic stroke compared to the administration of aspirin. And, the risk reduction is cumulative each year. For example, patients taking warfarin for five years will benefit from a 10% absolute risk reduction of stroke compared to patients taking aspirin, without significantly increasing their risk of bleeding.

Bleeding risk for patients on warfarin: Just as the CHADS2 score predicts the risk of stroke in patients with atrial fibrillation, other clinical prediction rules have been developed to estimate the risk of a major bleeding episode. The journal Chest recently published a model specifically designed to predict the risk of a subsequent bleeding episode within 90 days following hospitalization for an episode of gastrointestinal or intracranial hemorrhage in elderly patients receiving warfarin for atrial fibrillation [6]. Variables included in the model were age greater than or equal to 70, female gender, history of a remote bleeding event, a recent bleeding event, alcohol or drug abuse, diabetes, anemia, and use of an antiplatelet drug. Each variable was weighed in the model according to its relative impact on risk. The final score was then calculated as follows,

Risk score = 0.49*Xage 70+ + 0.32*Xfemale + 0.58*Xremote bleed + 0.62*Xrecent bleed

+ 0.71*XEtOH or drug abuse + 0.27*Xdiabetes + 0.86 Xanemia + 0.32*Xantiplatelet

where Xj equals 1 when the specific variable is present and 0 otherwise. A final score of 1.07 or less predicts a low-risk patient with a 0.9% risk of a major bleeding event. A score between 1.07 and 2.19 predicts a moderate-risk patient with a 2.0% risk. A score of 2.19 or greater predicts a high-risk patient with a 5.4% risk of a major bleed.

Discussion: Our patient has a CHADS2 score of 2, which corresponds to an annual stroke risk of 4% [1]. The use of warfarin reduces her risk of stroke to 1.3% each year [2, 3, 4]. This absolute risk reduction of 2.7% approximates the results of the study performed by Mant and colleagues, in which they found an absolute risk reduction of roughly 2% in patients older than 75 years of age [5]. On the other hand, the bleeding risk model published in Chest above gives her a score of 2.3, which predicts a risk of 5.4% of having another major bleed within 90 days of discharge from her current hospitalization, if she continues to take warfarin [6]. The literature suggests that aspirin provides a less effective alternative compared to warfarin for the prevention of stroke, and furthermore, does not significantly reduce the risk of bleeding [5]. Therefore, use of aspirin does not appear to be a reasonable alternative.

Dr. Villanueva noted that one-third of all patients with a diverticular bleed will suffer a second episode of bleeding at some later time. Certainly, this patient is at a significant risk for recurrent bleeding. However, many would argue that a stroke can cause greater morbidity, and thus is a worse outcome. Dr. Mints pointed out that recent literature shows decreasing rates of stroke in patients with atrial fibrillation over the past 10 years. He suggested that our improvements in the treatment of other risk factors may have had an impact on the declining risk of stroke in these patients.The question still remained-should the patient be restarted on warfarin for stroke prevention following her episode of major bleeding? The audience concluded that the patient would probably be best served by permanently withholding anti-coagulation, given her significant risk of a major bleeding episode. A conversation with the patient will be necessary to ascertain what risks and benefits are most in line with her own priorities. If ultimately it is decided to restart oral anti-coagulation with warfarin, closer monitoring of her INR should be undertaken, as bleeding complications are strongly associated with INRs greater than 3.


1. Gage BF et al. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA. 2001 Jun 13;285(22):2864-70.

2. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation. Arch Intern Med 1994;154(13)1449-1457.

3. Parkash R et al. The impact of warfarin on clinical outcomes in atrial fibrillation: a population based study. Can J Cardiol. 2007 May 1;23(6):457-61.

4. Go AS. Anticoagulation therapy for stroke prevention in atrial fibrillation: how well do randomized trials translate into clinical practice? JAMA. 2003 Nov 26;290(20)2685-92.

5. Mant J et al. Warfarin versus aspirin for stroke prevention in an elderly community population with atrial fibrillation (the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomized controlled trial. The Lancet. Vol 370 Aug 1, 2007.

6. Shireman TI et al. Development of a contemporary bleeding risk model for elderly warfarin recipients. Chest. 2006; 130:1390-1396.

Daria Crittenden is a 3rd year resident at NYU Medical Center