I recently received the following question from an ASH member through the Society's Consult-a-Colleague program. He asked:
I have a patient with mild hemophilia and a factor VIII level of 21 percent. He is 61 years old. He has recent-onset atrial fibrillation. I have recommended acetylsalicylic acid (ASA) since he has basically lone atrial fibrillation and no evidence of any cardiac disease. I also recommended the possibility of doing a transesophageal echocardiogram and, if negative, cardioverting him without anticoagulation. He has bled with surgery in the past but never otherwise. Do you agree with these recommendations?
The Response
I agree with this approach in this patient in the context of mild hemophilia A (factor VIII level of 21 percent of normal); however, my recommendations are empirical and are derived from my own experience, since these clinical situations are too unusual to develop evidence-based guidelines. In addition, the approach to mild hemophilia A is going to be somewhat less challenging than what would be required with severe hemophilia A. In any case, a factor VIII level of 21 percent should be adequate to sustain the anti-platelet aggregation effects of ASA (81 mg daily).
The fundamental question here is whether aspirin is an effective single agent to prevent the cardioembolic complications of atrial fibrillation. Recent large clinical trials and meta-analyses provide some insights. Warfarin appears to be superior to anti-platelet agents in preventing embolic stroke in patients over the age of 75 and in those with hypertension, prosthetic heart valves, mitral stenosis, or congestive heart failure. Others with atrial fibrillation seem to do equally well with aspirin alone.1 The 2007 Cochrane Report on this topic analyzed data from eight trials including more than 9,500 non-coagulopathic individuals (mean age = 69 years) with non-valvular atrial fibrillation.2 There were mixed statistical messages delivered in this publication in that while warfarin reduced the risk of all strokes by 32 percent compared with aspirin (P=0.0007) and decreased the incidence of disabling or fatal stroke by 31 percent (P=0.06), the mortality rate in both groups was similar. Of note, warfarin increased the risk of intracranial hemorrhage by 98 percent compared with aspirin. This latter risk issue has particular relevance for an individual with hemophilia, since it is very likely that any intracranial bleed (warfarin- or aspirin-induced) would result in exaggerated morbidity and mortality.
For the patient at hand with mild hemophilia A, I would place him on aspirin (81 mg daily) and would avoid warfarin until or unless an embolic event occurs or unless it is determined that there is a mitigating circumstance contributing to the onset of atrial fibrillation (e.g., valvular disease, cardiac disease, etc.). I would not expect any significant aspirin-induced hemorrhagic events to occur at this level of factor VIII activity. If the patient has severe or moderate-severity hemophilia A (FVIII <5 percent), I would recommend the same aspirin dose, initiate a secondary prophylaxis regimen of FVIII replacement therapy with a recombinant factor VIII concentrate (25-30 IU/Kg three times weekly), maintain the trough FVIII level greater than 5 to 10 percent, and subsequently titrate the dose of FVIII concentrate to achieve a FVIII activity adequate to minimize the development of spontaneous bleeding complications. If warfarin is needed in the context of hemophilia, I would aim for a target INR of 2 to 2.5 and administer once or twice daily FVIII concentrate to maintain FVIII levels of >50 percent, titrated upward based on the propensity of the patient to experience bleeding complications.
As for cardioversion, I would replace this patient's factor VIII activity level to at least 50 percent of normal with recombinant factor VIII concentrate prior to the procedure and then proceed with cardioversion without concurrent anticoagulation. Lastly, I think that this patient should be screened for the von Willebrand disease (VWD) variant 2 Normandy, since a number of patients with mild cases of hemophilia have been shown to be phenotypically similar to and genotypically confirmed to actually have this VWD variant.3
