Atrial fibrillation can cause blood clots to form in the heart, which can then travel to the brain and result in a blocked artery, leading to an ischemic stroke that destroys nearby neurons.
Doctors commonly prescribe warfarin to thin the blood so clots won't form in the heart. But the drug, which works by deactivating clotting factors in the blood, may initially cause a "hypercoagulable" state — one in which the blood becomes stickier and more likely to form clots in blood vessels.
In a study, published online in the European Heart Journal, McGill University researchers found the stroke risk was particularly high in the first week after patients started to take warfarin, peaking at about three days. After 30 days, that risk dropped by half, compared to patients not taking the drug.
"Warfarin is a highly effective drug for the prevention of stroke, there is no question there," lead author Laurent Azoulay, an epidemiologist at Jewish General Hospital who specializes in pharmacology, said Wednesday from Montreal.
"But it is documented that for some paradoxical reason, warfarin can increase the risk of stroke in the first weeks of treatment."
The chance of having a stroke soon after starting warfarin was found to be most pronounced in those with atrial fibrillation who had experienced a previous stroke, which he described as a small subset of patients.
"For the vast majority, this is not an issue," Azoulay stressed. "And, in fact, what we found is that for patients who continued to use the drug past those 30 days, they had a very strong protection against the risk of stroke."
To conduct the study, researchers analyzed data from more than 70,000 adults who were diagnosed with atrial fibrillation between 1993 and 2008. The study was carried out using the U.K. Clinical Practice Research Datalink, the world's largest primary-care database.
The researchers tracked patients through the database for up to 16 years, until an ischemic stroke, death, end of registration with their primary-care doctor or the end of the study, whichever came first. During that time, 5,519 patients, or two per cent per year, suffered a stroke.
In the first 30 days after starting warfarin, there was a 71 per cent increased risk of having a stroke when compared to patients not taking an anti-coagulant drug, the study found.
The highest risk was in the first week of use, peaking on the third day after starting the drug, when the risk rose more than twofold. In patients with a history of previous stroke, the risk was even higher.
Azoulay said the findings should not deter physicians or patients from using warfarin, since only a small proportion of patients would be affected.
"However, the results of our study suggest that physicians should be vigilant when initiating warfarin, particularly in the first week of use."
Dr. Muhammad Mamdani, director of the Applied Health Research Centre at St. Michael's Hospital in Toronto, said he has "a hard time buying" the study's conclusions, in part because of its design.
"The problem we see with these sorts of studies is that they tend to be biased in that patients who get started on warfarin get started on it for a reason," he said. "They tend to be much sicker than those who don't get started on warfarin.
"And because of that sickness, they're going to have higher event rates — so a higher likelihood of having a stroke — because they're sicker."
While the results offer food for thought, he said more rigorous studies need to be conducted to see if the findings hold up.
"Patients with atrial fibrillation should be on warfarin," said Mamdani, who was not involved in the study. "There may possibly be a short-term risk, but more definitive data are really needed and I don't think this will really change practice."
Senior study author Samy Suissa, a professor of epidemiology, biostatistics and medicine at McGill, agreed the findings need to be confirmed through additional studies by other research groups.
"It would be imperative to also investigate whether the newer popular anticoagulants also carry this early risk," said Suissa, referring to such drugs as dabigatran, rivaroxaban and apixaban.
Unlike warfarin, those drugs don't require frequent tests to ensure blood hasn't been thinned too much or too little. And while they appear from clinical trials to work as well as warfarin, there's a suggestion that the newer medications may increase the risk of gastrointestinal and other types of bleeding.
They also have a downside: there is no antidote to quickly turn off the newer medications' blood-thinning properties in the event of a bleed. With warfarin, the effects of too much drug can be reversed with doses of vitamin K, which plays a major role in the blood's ability to clot.