ARTICLE IN BRIEF
Investigators reported that patients on long-term, low-dose warfarin therapy may not be at increased risk of intracranial hemorrhage after receiving tissue plasminogen activator following an acute ischemic stroke.
Many patients on long-term, low-dose warfarin therapy may not be at increased risk of intracranial hemorrhage (ICH) after receiving tissue plasminogen activator (tPA) following an acute ischemic stroke, according to an observational study of cases from the American Heart Association's Get With The Guidelines–Stroke Registry.
Reporting in the June 27 edition of the Journal of the American Medical Association (JAMA), a team of doctors led by Eric D. Peterson, MD, MPH, professor of medicine in the Division of Cardiology and director of the Duke University Clinical Research in Durham, NC, reviewed data on 23,437 patients with an international normalized ratio (INR) of 1.7 or lower at 1,203 registry hospitals between April 2009 and June 2011. Of these patients, 1,802 were taking warfarin at the time of their stroke. Their median INR was 1.2, although the median quartile ranged from 1.07 to 1.4. [An INR value of 1 indicates normal bleeding time and 2–3 indicates a target, safe range with little or no bleeding risk.]
It is the largest clinical study to assess the safety of intravenous tPA in warfarin-treated patients meeting clinical guideline eligibility criteria. The current AHA guidelines endorse intravenous tPA in warfarin patients if their INR is 1.7 or lower, however these were based on limited data on its safety in clinical practice.
“When the AHA guidelines were developed there were no large trials of tPA in patients on blood thinners like warfarin. In fact, such patients were usually excluded from the clinical trials upon which the guidelines were based,” Dr. Peterson told Neurology Today in a telephone interview. “So the recommendations for use in low-dose warfarin patients were basically shot-gunned into the recommendations without any real supporting data. But what we found is that the evidence does in fact support inclusion, and this was a pleasant surprise.”
Although patients in the new study were older and had other comorbid conditions, and systemic ICH levels were higher than in patients not taking the blood thinner, the differences were not significant after adjustment for baseline clinical issues. Moreover, there were no significant differences for serious systemic hemorrhage, tPA complications, or in-hospital mortality; nor was the degree of anticoagulation significantly linked to systemic ICH risk among patients with INRs of 1.7 or lower.
That baseline INR values in the patients were below the therapeutic range (2.0–3.0 or 2.5–3.5) might explain the low incidence of ICH, the authors noted.
The researchers also found that about half of all warfarin-treated patients who otherwise would have been eligible for tPA were not treated, underscoring the need for further outreach efforts among clinicians, Dr. Peterson noted.
“For clinicians, the findings support giving tPA to patients with INRs of 1.7 or lower,” he said.
“This is an important study because it was large and well done,” said Mark J. Alberts, MD, professor of neurology at Northwestern University Feinberg School of Medicine, and director of the Stroke Program at Northwestern Memorial Hospital in Chicago.
In an accompanying editorial in JAMA, Dr. Alberts noted that more than 15 years after tPA was approved, it remains “vastly underutilized,” estimated to be used in only 2–8 percent of all acute ischemic stroke patients in the US.
He told Neurology Today in a telephone interview that the findings help demonstrate that long-term warfarin patients with an INR of 1.7 or less should receive the same critical emergency care that others not taking the blood thinner would after an acute ischemic stroke. Moreover, with exclusive inclusion and exclusion criteria, the incidence of ICH is low (4–5 percent) among treated patients, according to registries in the US, Canada, and Europe.
“We need to look for more ways to treat patients with tPA, not less — and this study indicates that tPA is safe for warfarin patients with subtherapeutic INR levels,” he said. “I think this paper shows that patients with lower INRs are not fully anticoagulated, so this is safe, even in patients with moderately elevated INRs. It is the reasonable thing to do.”
NOT SO FAST?
The debate is far from over, however. Another study, published July 3 in Neurology, found treating such patients with tPA may not be safe. The researchers performed a prospective trial in 548 stroke patients, including 15 who were taking subtherapeutic warfarin (INRs of 1.7 or less), and found a four-fold risk for intracranial hemorrhage (20 percent versus 5.6 percent). They also included a meta-analysis of other studies pointing to similar rates.
In an e-mail to Neurology Today, study co-author Stefan Kiechl, MD, a neurologist with the Innsbruck University Hospital in Innsbruck, Austria, said the study was characterized by a high-quality assessment of warfarin, “which is far from trivial,” as well as complete data assessment. “There were no missing data,” he noted.
“The main disadvantage is the small sample size (with a risk of chance findings) which is partly overcome by a meta-analysis of similar evaluations.”
On the other hand, he wrote, the JAMA study using the Get With The Guidelines (GWTG) data has the advantage of a larger sample size, but with the limitations inherent in using different registries — missing data, lack of detailed information, no systematic confirmation of responses on warfarin treatment, huge number of persons feeding data into the registry, etc.
“Inaccuracies in data assessment in GWTG may well have introduced a bias towards a null-association,” he said.
“In our study, there was a uniform procedure for all patients on warfarin and an INR of <1.7 whereas in GWTG only about 50 percent of these patients were treated. Treatment selection is a matter of concern and it cannot be judged for sure whether or not this has caused an underestimation of the true bleeding risk in warfarin-treated patients receiving rtPA.”
More importantly, he continued, the researchers tested the hypothesis that patients taking warfarin up to the manifestation of stroke had a heightened bleeding risk. GWTG included all patients that were taking warfarin within seven days of stroke onset — that is, patients with active treatment and all those who have stopped warfarin therapy (due to surgical interventions) and subsequently experienced stroke.
“This latter group had no heightened bleeding risk in our study,” Dr. Kiechl said. “Given the potential of a hypercoagulant state after warfarin withdrawal even the opposite may be true. There is no doubt that inclusion of such subjects in GWTG may have introduced a bias towards null and also no doubt that the research question tested in our study and in GWTG is not the same. Controversies surrounding the issue will go on and more data are required to learn the truth.”
In an editorial accompanying the Neurology study, Judith U. Harrer, MD, a neurologist at the Caritas Medical Center in Saarbrücken, Germany, said that because the results so closely correspond to rates in similar small trials, “the results … are a cause for concern.”
“This is not just another single-center study, given that this report provides the first meta-analysis on this issue,” she and coauthor Raymond C.S. Seet, MD, formerly with the Mayo Clinic of Rochester, MN, wrote, noting that the authors took special care to determine warfarin intake in the patients and used different analyses, yet reached the same conclusion.
Still the study's patient sample, like earlier studies, was small.
“This drawback — despite what the authors stated in their discussion — cannot simply be overcome by the addition of a meta-analysis,” they commented.
RESOLVING THE DEBATE?
According to Margaret Fang, MD, MPH, whose research has focused on outcomes associated with anticoagulation therapy, the JAMA study should help resolve some of the conflicting findings in earlier studies.
“I think this is a very important study in clinical terms because it answers the debate about whether or not tPA is safe and appropriate in these patients,” said Dr. Fang, associate professor in residence at the University of California, San Francisco (UCSF) School of Medicine, who is medical director of the UCSF Anticoagulation Clinic.
“Other studies have been much smaller and have had conflicting results, with some even showing a higher risk of ICH. Even though this was a retrospective observational study, the large number of patients and its rigorous sensitivity is pretty persuasive. I think it trumps all these earlier studies.
“Patients on warfarin with an INR of 1.2 are pretty close to normal, and this study shows that up to 1.7 appears to be safe and effective. But as INR values went up, so did the risk of ICH. I think that there is probably a threshold, but for patients with an INR of 1.7 or below, I think this shows that giving them tPA is pretty safe.”
One question, however, is whether any of the patients were given any warfarin reversal agent(s) prior to receiving tPA. Another is whether they used newer oral anticoagulants, which do not permit a patient to be given an INR rating.
“Clearance rates are quicker with the oral drugs, but different for each patient. And for those with compromised renal capability, the issue is even more critical,” Dr. Fang said.