The dilemma of resuming anticoagulation after intracranial hemorrhage: little evidence facing big fears.

Intracranial hemorrhage (ICH) is the most feared and devastating complication of anticoagulant treatment, leading to death or disability in two thirds of cases. Once ICH occurs, the decision of whether to resume anticoagulation is a true therapeutic dilemma that requires balancing the competing risks of hematoma growth or recurrent ICH and disabling thromboembolic events. Although the risk of thromboembolism in patients off anticoagulation is higher than the overall risk of ICH recurrence, there is a marked paucity of prospective large population-based data on the real risk of ICH recurrence on warfarin. The lack of randomized controlled trials probably reflects the ethical challenge of prescribing patients a medication to which they have an apparent contraindication. Therefore, in clinical practice, the risk is usually, and inappropriately, extrapolated from the overall risk of major bleeding on warfarin (approximately 3%), in which older age and elevated international normalized ratio are factors associated with an increased risk. The little evidence available on resuming oral anticoagulation after ICH comes from either expert opinions or few nonrandomized mainly retrospective studies.1,2 These studies included highly selected high-risk patients and showed nonconclusive and even discrepant results. This limited and weak evidence along with our own experience and common sense are the weapons that our protagonists use for facing the physicians’ fears and uncertainties of increasing the risk of a devastating recurrent ICH or leaving the patient unprotected from thromboembolic complications. Dr Shulman’s argument is based on the high risk of recurrent ICH on warfarin after ICH at any location and gives a broad recommendation of abstaining from resuming warfarin. He argues that in our case, warfarin should not be resumed because the risk of recurrent ICH (15%) is more than twice as high as the risk of ischemic stroke (6%). Dr Steiner delineates a more restrictive scenario, in which the risk of thromboembolism outweighs the risk of ICH recurrence in a hypertensive-related, nonlobar ICH. He recommends that warfarin should be restarted if blood pressure and other risk factors are adequately controlled. Among all factors associated with an increased risk of recurrent ICH on warfarin, ICH location and documented history of thromboembolism seem to be the key factors that tilt the risk/benefit balance of restarting anticoagulation after ICH. In 1 of the only 2 published epidemiological studies, the risk of recurrent ICH on warfarin was 5-fold higher in lobar compared with deep ICH, although the rate of survival among patients with deep ICH was low. The topographical location of ICH may reflect the underlying microvascular pathology. Lobar ICH in the aged population is associated with cerebral amyloid angiopathy and an inherent high risk of recurrence. Deep ICH is often hypertension-related. Although improved management of hypertension can reduce the risk of recurrent deep ICH, there is limited room for improving management in lobar ICH if blood pressure is well controlled. On the other hand, in an analysis of 52 patients, thromboembolic events occurred in 48% of patients in whom warfarin was not restarted, all of them were being treated for a previous event, suggesting that secondary rather than primary prevention is a stronger indication for resuming anticoagulation.2 The dilemma of restarting oral anticoagulation in the long-term management of ICH may be better addressed by considering other factors, including the underlying reason for which the patient was originally started on anticoagulation, difficulties in controlling the international normalized ratio, the risk of thromboembolic stroke based on the CHADS2 score, and the presence and extent of microbleeds on gradient-echo MRI. Deep ICH, secondary prevention, high CHADS2 score, mechanical valve, or hypercoagulable state are factors arguing in favor of resumption of anticoagulation. Conversely, lobar ICH, presence of multiple microbleeds on MRI, low CHADS2 score, and difficulties controlling international normalized ratio configure an unfavorable risk/ benefit profile. In addition to these factors, the decision of whether to resume anticoagulation must take into consideration the underlying cause of ICH. For example, treatable