Subscribe to eTOC


SAN ANTONIO — While the overall trial data from the Warfarin-Aspirin Recurrent Stroke Study (WARSS) showed no significant difference between aspirin or warfarin treatment in preventing recurrent stroke in certain patients, some interesting data are being found deep inside its substudies.

The substudies, which were reported here at the American Stroke Association Conference in February, explored these and other questions: (1) Does antibody status play a role in recurrent stroke risk?; (2) Does valvular incompetence of the foramen ovale of the heart affect the chance of adverse events in patients taking warfarin or aspirin; and (3) Is there a connection between levels of the prothrombin fragment Fl.2 and WARSS endpoints?


Robin L. Brey, MD, Professor of Neurology at University of Texas Health Science Center in San Antonio, reported that the role of antiphospholipids in recurrent stroke risk is still far from clear, even after the WARSS subset AntiPhospholipid Antibodies and Stroke Study (APASS). Dr. Brey said APASS patients who were antibody positive for lupus anticoagulant, anticardiolipin antibody, or both in the WARSS study were at no higher risk for stroke, death, or other thrombo-occlusive events compared with those who were antibody negative. Antibody status did not appear to predict any differential treatment response to either aspirin or warfarin, she said.

The 1,770 WARSS patients enrolled in APASS provided blood samples for antiphospholipid antibody testing at baseline, 12 and 24 months, and at the time of a recurrent event. Besides the WARSS endpoints, APASS endpoints included transient ischemic attack (TIA), myocardial infarction (MI), deep venous thrombosis, pulmonary embolism, peripheral arterial thrombosis, and systemic visceral arterial thrombosis.

At baseline, 41 percent of subjects overall were positive for either lupus anticoagulant or anticardiolipin antibody, while 59 percent were negative for both.

“Routine screening for antiphospholipid antibodies in this group of ischemic stroke patients does not appear to be warranted,” Dr. Brey concluded, but she said that titers, isotypes, and serial data have yet to be studied.


Another WARSS substudy – which looked at the role of patent foramen ovale (PFO) in cryptogenic stroke – found that patients with cryptogenic stroke who took warfarin had about half the event rate as those with cryptogenic stroke who took aspirin, regardless of the presence of PFO. The substudy is referred to as PICSS or the PFO in Cryptogenic Stroke Study.

Approximately 40 percent of strokes are cryptogenic with no clear cause, and researchers hypothesized that some of these may be accounted for by the presence of PFO.

“In a cohort of patients with cryptogenic stroke there was a trend toward a warfarin effect in patients with and without PFO,” said Shunichi Homma, MD, Associate Professor of Medicine at Columbia-Presbyterian Medical Center in New York City.

But the presence of PFO did not increase the chance of adverse events whether the patient was on warfarin or aspirin, he said, and there was no difference between warfarin and aspirin preventing adverse events in patients with PFO.

PICSS enrolled 601 patients from WARSS. Stroke subtypes were 42 percent cryptogenic, 39 percent lacunar, and 11 percent large artery. PFO as defined by transesophageal echocardiography was present in 34 percent of these patients, 36 percent large PFOs and 64 percent small.

Dr. Homma said there was very little difference in outcomes between the two treatment groups in the entire PICSS cohort. For those with PFO, rates were 16.5 percent for those taking warfarin and 13.2 percent for aspirin.

For those patients without PFO the rates were 13.4 percent for the warfarin arm and 17.4 percent for aspirin.

There was an 8.3 percent event rate for the warfarin cohort without PFO versus 16.3 percent for the aspirin cohort without PFO; and 9.5 percent for the warfarin cohort with PFO versus 17.9 percent for the aspirin cohort with PFO.

After his presentation, Dr. Homma was asked how PICSS data would compare with a New England Journal of Medicine (NEJM) study by Dr. Jean-Louis Mas and colleagues (NEJM 2001; 345 (24): 1740–1746), which found that the risk of recurrent stroke among patients with patent foramen ovale alone was 2.3 percent, much lower than that seen in PICSS.

“The difference can very well be from the difference in the age of the patients,” Dr. Homma said. “Their patients were a lot younger than ours.”


The prothrombin fragment F1.2 is elevated in pro-thrombotic states, such as inherited thrombophilias and antiphospholipid antibody syndrome. The WARSS Hemostatic Activation Study (HAS) investigated the possible connection of F1.2 levels with stroke.

Acute thrombotic episodes can also elevate F1.2 levels, said Karen Furie, MD, Instructor in Neurology at Harvard Medical School and Director of Stroke Neurology at Spaulding Rehabilitation Hospital in Boston. She said in acute ischemic strokes, F1.2 levels are higher in larger cortical strokes and in strokes that are cardioembolic as opposed to lacunar in etiology. And in patients with TIAs and carotid bruits, levels of F1.2 can be predictive of MI and stroke.

But data on the 320 patients in HAS showed no significant difference at three months in F1.2 levels between patients with cryptogenic and noncryptogenic stokes treated with aspirin, or in those treated with warfarin, Dr. Furie said.

“We found an association between levels of F1.2 and stroke risk factors, and that F1.2 levels increased over time,” Dr. Furie said. But she said the investigators were unable to demonstrate a statistically significant association between WARSS endpoint events and levels of F1.2, even though it did appear that higher levels of F1.2 did correlate with a greater risk of stroke and death.

The researchers did see that warfarin reduced F1.2 levels by approximately 40 percent, and that prothrombin time as measured by the international normalization ratio levels influenced F1.2. Greater age was associated with elevated F1.2 levels, Dr. Furie said, as was male gender.

But there was no statistically significant difference in mean F1.2 levels in patients with cryptogenic versus noncryptogenic stroke. Dr. Furie said this may be due to misclassification of strokes, or to concomitant atherosclerotic disease, or it may be that there is no difference in terms of hemostatic activation between the stroke subtypes.


Principal WARSS investigator J.P. Mohr, MD, Professor of Clinical Neurology at the New York Neurological Institute of Columbia-Presbyterian Medical Center, said here that there was a concern that major hemorrhage might become an issue in the HAS trial, but it did not. With 1,103 patients in each arm, there were 44 major hemorrhagic events in the warfarin arm and 30 in the aspirin cohort, not a statistically significant difference.

Minor hemorrhagic events totaled 413 for warfarin and 259 for aspirin, with a strong statistically significant p-value. But Dr. Mohr said, “The difference which failed to be shown in major hemorrhage may serve somewhat to reassure those who give warfarin, or frighten those who give aspirin.”


WARSS was a double-blind, multicenter study of 2,206 patients, which found that after two years that there was no difference between aspirin and warfarin in the prevention of recurrent ischemic stroke or death. The hazard ratio was 1.13 against warfarin and in favor of aspirin, and the authors concluded that both drugs are reasonable therapeutic alternatives.

In a presentation here at the 27th International American Stroke Association Annual meeting, Ralph L. Sacco, MD, Associate Professor of Neurology and Public Health at Columbia University College of Physicians and Surgeons, said that further study of the 2,206 cases found no significant difference between the warfarin and aspirin based on age, gender, race, and education.

Similarly there was a trend in favor of warfarin but no significant difference with regard to patients with a history of diabetes, cardiac disease, prior transient ischemic attack (TIA) or stroke, arrival at the hospital within 24 hours, sedentary lifestyle, smoking, or obesity.

There was a significantly increased hazard for patients who had a brainstem infarct taking warfarin compared with aspirin, Dr. Sacco said, while outcomes for patients with cryptogenic infarcts were slightly better with warfarin. Dr. Sacco cautioned that these subgroup analyses are exploratory and cannot be considered evidence for treatment.


Dr. Ralph L. Sacco cautioned that the subgroup analyses of the WARSS cannot yet be considered evidence for treatment.