Rocket-AF: Predictors of ICH Include Warfarin Treatment

Susan Jeffrey

February 23, 2012

February 23, 2012 (New Orleans, Louisiana) — A subanalysis from a large clinical trial comparing warfarin with rivaroxaban (Xarelto, Bayer/Johnson & Johnson) points to a number of predictors of risk for intracranial hemorrhage (ICH) in patients with atrial fibrillation (AF).

Investigators report that factors such as renal impairment, increasing diastolic blood pressure, and decreasing albumin level were associated with higher bleeding risk and that several factors were related to lower risk, including baseline warfarin use and residence in Eastern Europe compared with other regions. In addition, randomization to rivaroxaban was associated with a 40% lower risk for ICH.

The findings with regard to rivaroxaban are consistent with those seen recently with dabigatran (Pradaxa, Boehringer Ingelheim) and apixaban (Eliquis, Pfizer/Bristol-Myers Squibb), other approved and potential warfarin alternatives, probably because unlike warfarin, none of these newer agents interact with factor VII, Graeme Hankey, MD, lead author and neurologist at the Royal Perth Hospital and University of Western Australia, told a press conference here.

Dr. Graeme Hankey

The subanalysis, from the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET-AF) was presented at the International Stroke Conference 2012. The trial was funded by Johnson & Johnson Pharmaceutical Research and Development and by Bayer HealthCare AG.

Uncommon Condition

Treatment with anticoagulation in patients with AF reduces the risk for stroke and systemic embolism, but identifying patients at risk for ICH with anticoagulation "remains a challenge for us," Dr. Hankey said. Several studies have looked at risk factors associated with ICH in this setting, he noted, "but many of the models have been unstable because this is such an uncommon condition and the number of outcomes is small."

In this report, then, the researchers aimed to determine the independent significant predictors of intracranial hemorrhage in the ROCKET-AF trial of 14,264 patients with AF receiving anticoagulation.

ROCKET-AF was a randomized, double-blind, double-dummy trial. Patients were included if they had AF and either a history of stroke, transient ischemic attack (TIA), or systemic embolism, or at least 2 of 4 risk factors for stroke (cardiac failure or ejection fraction ≤ 35%, hypertension, age 75 years or older, or diabetes mellitus), and no condition associated with increased bleeding risk.

Eligible patients were randomly assigned to receive rivaroxaban at a dosage of 20 mg daily (or 15 mg daily for those with impaired creatinine clearance) or warfarin with a target international normalized ratio (INR) of 2.0 to 3.0.

The primary outcome was stroke or systemic embolism. The final results, published in the New England Journal of Medicine, showed that rivaroxaban was noninferior to warfarin. ICH was reduced with rivaroxaban, but mortality did not differ between groups.

For this analysis, the primary outcome of interest was the rate and predictors of ICH in the entire cohort, regardless of what treatment they were randomly assigned to receive. ICH was defined as symptomatic bleeding into the skull causing focal neurologic symptoms and diagnosed by imaging or autopsy.

The final analysis included 13,833 of the patients in ROCKET-AF because of missing baseline data for some patients, Dr. Hankey noted. Over a median of 1.94 years of follow-up, there were 172 (1.2%) ICH events; most were intracerebral (n =128) but a "substantial proportion" (n = 38) were in the subdural space, he said. The actuarial rate was 0.68 per 100 patient-years of follow-up.

Multivariate analysis turned up "3 independent significant protective factors, and 6 independent significant hazardous factors for intracranial hemorrhage," Dr. Hankey said.

Protective factors included residence in Eastern Europe (possibly related to its lower population of Asian and black participants, both of whom had a higher risk for ICH, or, alternatively, to ascertainment bias), randomization to rivaroxaban, and treatment at baseline with warfarin. It’s possible these latter patients had already perhaps "survived the stress test of warfarin without an occult bleed," he speculated.

Factors associated with increased ICH hazard were reduced creatinine clearance indicating impaired renal function, a lower platelet count, low albumin, a history of stroke, baseline use of thienopyridine (such as clopidogrel), and increasing diastolic blood pressure. "This was closely correlated with systolic blood pressure and impaired creatinine clearance was closely correlated with age," Dr. Hankey noted.

Table 1. ROCKET-AF: Predictors of Intracranial Hemorrhage

Factor Hazard Ratio (95% Confidence Interval) P Value
East Europe (vs West Europe, Asia-Pacific region, North America, Latin America) 0.34 (0.22 - 0.50) <.001
Randomly assigned to rivaroxaban vs warfarin 0.60 (0.44 - 0.83) .002
Baseline vitamin K antagonist use 0.62 (0.45 - 0.85) .003
Cockcroft-Gault creatinine clearance (per 10-mL/min decrease) 1.10 (1.04 - 1.18) .002
Platelet count < 210 x 109/L 1.08 (1.03 - 1.14) .002
Albumin (per 0.5-g/dL decrease) 1.37 (1.11 - 1.69) .004
History of stroke or TIA 1.55 (1.14 - 2.11) .006
Prior thienopyridine use 2.50 (1.27 - 4.92) .008
Diastolic blood pressure (per 10-mm Hg increase) 1.21 (1.05 - 1.41) .010


Asked for comment, José Biller, MD, professor and chair, Department of Neurology, Loyola University College in Maywood, Illinois, pointed out several points to be made about these observations.

The higher rate of ICH seen in Asian participants in this analysis suggests that a lower dose of rivaroxaban should perhaps be considered for this group, Dr. Biller told Medscape Medical News. Some of the other findings (for example, higher risk with lower creatinine clearance, albumin, or platelets, higher diastolic blood pressure, or dual antiplatelet therapy) were not surprising, he added.

Dr. José Biller

"So many of these observations are intuitive, but they are very important in properly selecting patients as to what would be appropriate treatment in patients who have high risk nonvalvular atrial fibrillation," Dr. Biller said.

Importantly, in ROCKET-AF, 90% of patients had a CHADS2 score of 3 or greater, "so they were high-risk patients," he noted.

"We recognize that there are a number of disadvantages to warfarin use, so you have to do a careful analysis of pros and cons," he said. Warfarin is cheaper, but keeping patients in therapeutic range is more difficult. Rivaroxaban has no antidote to date, and if a stroke does occur, thrombolysis may not be possible.

"I am very intrigued and excited about these new agents, and more and more patients are being given these medications," Dr. Biller concluded, "but time will tell us about their advantage over warfarin."

Safe for the Elderly?

In a separate presentation here, Dr. Hankey reported on a second subanalysis from the ROCKET-AF trial on behalf of Jonathan L Halperin, MD, from Mount Sinai School of Medicine in New York, who was unable to attend.

This analysis aimed to see whether the safety and efficacy of rivaroxaban vs warfarin in patients with AF who were 75 years of age and older were consistent with the overall trial results. Although AF is more common among older patients, warfarin is known to be widely underused in this group because of concerns about risk, Dr. Hankey noted.

Of the overall ROCKET-AF population, 6154 patients, or about 43%, were 75 years of age or older. Mean time with an INR in the therapeutic range in the warfarin group was 53.9% for those under 75 and 56.9% for those 75 and older.

The average rate of stroke and systemic embolism was higher in the older patients, Dr. Hankey noted, but "the relative treatment effect of rivaroxaban against warfarin was reasonably consistent amongst older patients compared to younger patients; just their absolute rates were higher."

Table 2. ROCKET-AF: Stroke and Systemic Embolism Rate by Age

Age Group Rivaroxaban (% per year) Warfarin (% per year) Hazard Ratio (95% Confidence Interval) P Value for Interaction
<75 y 2.00 2.10 0.95 (0.76 - 1.19) .3131
≥75 y 2.29 2.85 0.80 (0.63 - 1.02)  


Similarly, major bleeding was higher in the older patients, at around 4.6% in older patients vs 2.7% in younger patients. "There was a slightly higher rate of major bleeding in patients assigned rivaroxaban compared to warfarin who were elderly; only an 11% relative excess but consistent with an 8% reduction or a 34% excess," Dr. Hankey noted.

Major bleeding did not differ between groups for younger patients, and no significant interaction was seen between the 2 proportional reductions.

For hemorrhagic stroke, there was no absolute difference in stroke rates between younger and older patients, he noted. "What was consistent in this trial though, and with the other newer anticoagulants, is there is a lower intracranial hemorrhage rate amongst patients assigned the new anticoagulants compared to warfarin."

Table 3. ROCKET-AF: Bleeding Events by Age

Endpoint Rivaroxaban (Events per 100 Patient-Years) Warfarin (Events per 100 Patient-Years) P Value for Interaction
Major bleeding      
   <75 y 2.69 2.79 .336
   ≥75 years 4.86 4.40  
Hemorrhagic stroke      
  <75 y 0.20 0.34 .365
  ≥75 y 0.41 0.49  


In the ROCKET-AF trial, elderly participants had a higher absolute rate of stroke and major bleeding and a similar absolute rate of hemorrhagic stroke as younger patients, Dr. Hankey concluded.

"But the overall relative effects of rivaroxaban compared to warfarin were consistent among the elderly patients for efficacy and safety, and so the overall trial results can be applied confidently to the elderly population," he said.

ROCKET-AF was funded by Johnson & Johnson Pharmaceutical Research and Development and by Bayer HealthCare AG. Dr. Hankey reports that he received honoraria as a member of the ROCKET-AF executive committee. Dr. Biller has disclosed no relevant financial relationships.

International Stroke Conference 2012: Abstracts #152, 148. Presented February 2, 2012.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: