Neurocritical Care

2. Reversal of Anticoagulation After Cerebral Hemorrhage

A 70-year-old man with a metallic aortic valve treated with warfarin and aspirin developed a sudden speech difficulty, right-arm weakness and numbness. On arrival to an outside emergency department he was alert, but there was a severe dysarthria and right-sided hemiparesis involving arm and leg. CT scan showed a localized thalamic hematoma without ventricular extension, but with some mass effect (Figure 2.1). International normalized ratio (INR) was 4.3. He received fresh frozen plasma (2 units) and intravenous vitamin K (5 mg). The neurosurgeon at the outside hospital felt there was no surgical option and would only intervene after INR was corrected.

After transfer to our emergency department, there was clear evidence of further neurologic deterioration, mainly decline in responsiveness with eye opening to voice only, impaired vertical eye movements, marked dysarthria, and flaccid hemiparesis. He is still able to protect his airway, but he has developed Cheyne-Stokes breathing. Repeat CT scan of the brain shows enlargement of the thalamic hematoma and rupture into the third ventricle. His systolic blood pressure has climbed to 200 mmHg. INR is still 4.0.

What do you do now?


FIGURE 2.1 Serial CT scans show expansion of the thalamic hematoma and development of acute hydrocephalus.

Ongoing anticoagulation in a patient with a cerebral hematoma is a serious concern. As expected, warfarin causes larger cerebral hemorrhages and increases the chance of poor outcome. Avoiding this expansion—and even reduction of size by 2 teaspoons could make a difference—may reduce the morbidity from additional brain tissue destruction or prevent brainstem injury from displacement. Everyone would agree that a first step would be to quickly correct the INR to a normal value (INR < 1.5). However, it is not well established that rapid reversal of the anticoagulant effects of warfarin effectively reduces enlargement of the hematoma.

The approach is to reverse the antagonistic effect of warfarin on vitamin K, and vitamin K will then reactivate factors II, VII, IX, and X. Using both intravenous vitamin K and fresh frozen plasma (FFP) accomplishes that, but only after several hours. Moreover, vitamin K alone is not sufficient and may even take 6–24 hours to take full effect; therefore by itself vitamin K has no substantial effect on expansion of the hematoma occurring usually in the first hours after the initial hemorrhage. FFP further replaces the depleted coagulation factors, but multiple studies have shown that target INR is not reached within 2–4 hours in the majority of patients (compatibility testing and thawing of plasma lasting 30–60 minutes adds to the delay). Equally problematic is when INR is not corrected rapidly with FFP, since it may lead physicians to infuse more units of FFP, since leading to transfusion-associated circulatory overload, pulmonary edema, and in the worst-case scenario endotracheal intubation and mechanical ventilation. There is no consensus on the number of FFP units needed, although weight-based calculation may reduce complications. As a general rule, a dose of 10–20 mL of FFP/kg of body weight will produce a sufficient 10% increase in coagulation factors. A typical unit is 250 cc, thus 3 to 4 units are often needed. Finally and most concerning, one should not be surprised to find out that some emergency departments may not have fresh frozen plasma readily available. This delay in treatment is obviously unacceptable.

The best alternative options for correction of warfarin are prothrombin complex concentrate (PCC) or recombinant activated factor VII (rFVIIa). PCCs contain human derived clotting factors and rFVIIa is bioengineered. PCCs are basically a concentrate of factor IX and smaller amounts of II, X, and VII. Neither PCC nor rFVIIa is universally available.

There is a tendency to prefer PCC (Table 2.1). There are several reasons for that: it is easy to use and quickly prepared, there is a minimal infused volume, it nearly completely replaces clotting factors, and the most convincing argument for some physicians is that it lasts longer than rFVIIa and less additional FFP may be needed. But thrombotic events using PCC may not be different from rFVIIa, and there are very few studies that have assessed this risk with PCC. The risk of arterial and venous occlusive events in rFVIIa was 26% with a low dose (20 mcg/kg) and almost 50% with higher dose (80 mcg/kg) in the largest cerebral hematoma trial although most of these events were inconsequential. This is consistent with our experience in daily practice. A recent detailed analysis of several clinical trials in multiple conditions found these increased risks, particularly in patients over 65 years of age, often resulting in venous occlusions.

Patients with severe thrombocytopenia need platelet transfusions. An unresolved issue is whether platelet infusion in a patient with prior use of antiplatelet agents reduces hematoma expansion or improves outcome. Clinical trials are underway, and there is yet no definite evidence that platelet infusion can impact on progression of the hemorrhage or even outcome.

A new problem will be introduced when the use of dabigatran (a thrombin or factor II inhibitor) or apixaban (a factor Xa inhibitor) becomes more commonplace, because no reversal strategy is available other than stopping the drug. The anticoagulation effect of these newer drugs reverses many hours (half life 12–15 hours) after discontinuation and thus not soon enough. FFP would not have any major effect, but PCC or rFVIIa may help. Research on antibodies against these drugs is ongoing. Finding an antidote will be important because these new expensive drugs may replace warfarin in the long run.

TABLE 2.1 Reversal of Warfarin


† = commonly used

†† = preferred, if available, there may be substantial differences in costs between products

* Much lower doses (5–10 mcg/kg) may be sufficient in our experience

So what should you do? Table 2.2 summarizes the initial priorities. The initial management must remain focused on rapid correction of the INR. Equally important is aggressive control of blood pressure using labetalol, hydralazine, or intravenous infusion with nicardipine with the assumption that keeping the blood pressure under control additionally reduces further expansion.

Another potential complication in patients with a thalamic hemorrhage is the development of hydrocephalus due to trapping of CSF outflow at the foramen of Monro. A ventriculostomy is readily placed by neurosurgeons—after INR correction—in patients with significant intraventricular clot, but there is serious doubt if this intervention can change outcome or even result in a noticeable improvement. (In some patients the diencephalic destruction may leave the patient in a prolonged stuporous state and a ventriculostomy does not help.) Acute hydrocephalus may be a reflection of a major hemorrhage and not necessarily a treatable complication. Moreover, keeping the ventriculostomy patent and draining has always been the limiting factor. In patients with a lobar hematoma and worsening neurological findings, surgical evacuation is the only available option for survival.

TABLE 2.2 Emergency Management of Warfarin Associated Cerebral Hemorrhage

Aggressively lower INR to normal (INR < 1.5)

Aggressively control blood pressure (SBP < 160 mmHg)

Consider ventriculostomy when INR < 1.5 and hemoventricle with hydrocephalus

Monitor EKG/troponin if rFVIIa has been administered

Monitor X-Ray of the chest for pulmonary infiltrates if FFP has been used and consider diuretics

Outcome therefore will be determined by 1) whether the appropriate interventions are pursued in a salvageable patient and 2) sufficient time for recovery is allowed.

But what of resumption of anticoagulation after the patient recovers and still needs protection against future emboli? The risk of future hemorrhagic complications after resuming warfarin is 10–20%. Early thromboembolic complications have been estimated at 5%. There is no clear consensus on what to do and the trade-off depends on the individual risk of thromboembolism (higher with prosthetic valves) and recurrent hemorrhage (higher with suspected cerebral amyloid angiopathy). One retrospective analysis—mostly in patients with atrial fibrillation alone—found that waiting 1–2 months was justified. Others have found that even waiting with restarting anticoagulation for 7–10 days has resulted in increase in thrombotic complications and have recommended that one should consider restarting low intensity warfarin 3–4 days after the patient has a stable hematoma and no neurosurgical intervention has been performed. While taking the risk of restarting warfarin so promptly may be unnecessary in most cases, early resumption of warfarin may be considered in patients with prior TIAs, prosthetic valves, or echocardiographic finding of an atrial thrombus. We generally wait 7–10 days before restarting.


· PCC or rFVIIa may be a more effective way to reverse warfarin.

· Vitamin K and fresh frozen plasma may be the only available option, but INR is corrected in only 1/3 of the patients within 12 hours.

· Control of blood pressure is equally important to reduce expansion of the hematoma.

· Ventriculostomy may be needed in patients with significant intraventricular blood, but only when INR is less than 1.5.

· Neurosurgical evacuation of hematoma should remain an option, but only when INR is less than 1.5.

Further Reading

Aguilar MI, Hart RG, Kase CS, Freeman WD et al. Treatment of warfarin-associated intracerebral hemorrhage: literature review of expert opinion. Mayo Clin Proc 2007; 82:82–92.

Bershad EM, Suarez JI. Prothrombin complex concentrates for oral anticoagulant therapy–related intracranial hemorrhage: a review of the literature. Neurocrit Care 2010; 12:403–413.

Claassen DO, Kazemi N, Zubkov AY, Wijdicks EFM, Rabinstein AA. Restarting anticoagulation therapy after warfarin-associated intracerebral hemorrhage. Arch Neurol 2008; 65:1313–1318.

Diringer MN, Skolnick BE, Mayer SA, Steiner T, Davis SM, Brun NC, Broderick JP. Thromboembolic events with recombinant activated factor VII in spontaneous intracerebral hemorrhage: results from the Factor Seven for Acute Hemorrhagic Stroke (FAST) trial. Stroke 2010; 41:48–53.

Elliott J, Smith M. The acute management of intracerebral hemorrhage: a clinical review. Anesth Analg 2010; 110:1419–1427.

Imberti D, Barillari G, Baisoli C et al. Emergency reversal of anticoagulation with a three-factor prothrombin complex concentrate in patients with intracranial hemorrhage. Blood Transfus 2011;9:148–155.

Lee SB, Manno EM, Layton KF, Wijdicks EFM. Progression of warfarin-associated intracerebral hemorrhage after INR normalization with FFP. Neurology 2006; 67:1272–1274.

Levi M, Levy JH, Andersen HF, Truloff VM. Safety of recombinant activated factor VII in randomized clinical trials. N Engl J Med 2010; 363:1791–1800.

Majeed A, Kim Y-K, Roberts RS. Optimal timing of resumption of warfarin after intracranial hemorrhage. Stroke 2010; 41:2860–2866.

Robinson MT, Rabinstein AA, Meschia JF et al. Safety of recombinant activated factor VII in patients with warfarin-associated hemorrhages of the central nervous system. Stroke 2010; 41:1459–1463.

Steiner T, Bosel J. Options to restrict hematoma expansion after spontaneous intracerebral hemorrhage. Stroke 2010; 41:402–409.