Small Study Finds Spike in Hemorrhages Among Warfarin Patients Given tPA After Stroke

Neurology Today
1 July 2010; Volume 10(13); pp 1,26

SAMSON, KURT

Outline

article in brief
study limitations
earlier indications?
references

ARTICLE IN BRIEF

A small, retrospective study found that older patients taking warfarin had a ten-fold increased risk of intracerebral hemorrhage (ICH) when treated with intravenous tPA within the approved three-hour treatment window after an acute ischemic stroke.

Despite general agreement that patients taking the blood thinner warfarin can safely be given tissue plasminogen activator (tPA) after an acute ischemic stroke if their clotting levels fall within the accepted “normal” range, new evidence suggests this might not be the case.

The new study, reported in the May Archives of Neurology, raises questions about the widely accepted emergency procedure in such patients, but larger studies are needed, according to the report.

Intravenous tPA is a rapid clot-dissolving agent that can improve outcomes after acute stroke, although the risk of hemorrhage is increased in some populations, including older adults and those with more severe strokes, high blood glucose levels, lower platelet counts, and high blood pressure. Guidelines recommend that patients taking warfarin can be given tPA if blood clotting test results fall within the international normalized ratio (INR) of 1.7 or less. It is considered safe for all patients with INR scores of less than 1.7, even in those as young as 18 years.

Figure. OLDER PATIENTS taking warfarin had a 10-fold increased risk of intracerebral hemorrhage when treated with intravenous tPA.

[ Click here to enlarge ]

The authors considered such factors as age, NIH Stroke Scale score, atrial fibrillation, and initial INR; baseline warfarin use remained strongly associated with symptomatic intracerebral hemorrage (p=.004). The average age of warfarin users in the study was 80.6 years versus 67.7 years for those not taking the drug, and the older patients were more likely to present with atrial fibrillation — 69.2 percent versus 19.1 percent — and higher initial INR (an average 1.21 versus 1.03).

Although the US FDA has approved tPA within three hours of symptom onset, several major studies have found it effective if administered within four and a half hours. The American Heart/American Stroke Association and European guidelines now accept this wider treatment window, as do many hospitals and stroke centers in the US.

STUDY LIMITATIONS

Shyam Prabhakaran, MD, an assistant professor of neurological sciences at Rush University Medical Center in Chicago, and colleagues at the Neurological Institute at Columbia University Medical Center in New York, and St. Luke's-Roosevelt Hospital Center in New York, studied the records of 107 stroke patients.

They found 13 (12.1 percent) were taking warfarin with INR values of less than 1.7 when administered tPA after their stroke. Their NIHSS score was 14, and the median time from symptom onset to treatment time was 140 minutes.

Nonetheless, symptomatic hemorrhages occurred in seven of these patients (30.8 percent; 95 percent confidence interval), compared with only two (3.2 percent) of the others, and remained strongly associated with symptomatic ICH after adjusting for age, atrial fibrillation, NIHSS scores, or INR values. In addition to these patients, there were three asymptomatic intracerebral hemorrhages (2.8 percent) within 36 hours of tPA administration.

Several mechanisms may explain this association, the authors noted. The clot-dissolving effects of tPA may be enhanced by the clot-preventing effects of warfarin, even at low levels, and the effects of warfarin last for an average of three days after the last dose, so the INR may continue to increase following treatment with tPA, they noted.

Given the small, retrospective study design, as well as other limitations of the study, it should “serve as a hypothesis-generating report that requires confirmation in larger cohorts,” the authors concluded. “Further analysis including more extensive adjustment for confounding variables in larger data sets may prove useful.”

Other study limitations included a referral bias toward sicker patients seen at stroke centers versus community hospitals. In addition, initial CT scans were not available for all patients and serial INR measurements and angiography were not performed in every patient's case.

EARLIER INDICATIONS?

The results raise potential safety concerns regarding the currently accepted guidelines for tPA use in patents taking warfarin and other blood thinners, according to Glenn D. Graham, MD, associate professor of adult neurology at the University of New Mexico School of Medicine, and director of the Cerebrovascular Disorders Program at the US Department of Veterans Affairs Hospital in Albuquerque.

“This is an important paper because it generates a hypothesis that these treatment guidelines, which were assumed to be correct for patients taking warfarin who receive tPA within three hours, may not be sufficient,” he told Neurology Today in a telephone interview.

Patients taking anticoagulants may have been underrepresented or excluded in the original studies on which the FDA-approved guidelines are based, he noted.

“My concern is that the INRs may not have been normal in the original guideline studies. My sense is that there might have been a little stretch in the earlier guidelines with regard to warfarin patients, and perhaps it was not justified. The patients in this study were also older and had atrial fibrillation, so we don't know exactly what was going on or why.”

The next logical step would be to check the registries at other stroke centers for any similar trends, he said. “It would be premature to recommend any changes in the current treatment guidelines. At this stage, this is just something to give us pause.”

Figure. DR. GLENN D. GRAHAM said the study results raise potential safety concerns regarding the currently accepted guidelines for tPA use in patents taking warfarin and other blood thinners.

[ Click here to enlarge ]

J. Claude Hemphill III, MD, associate professor of clinical neurology and the Kenneth Rainin Chair in Neurocritical Care at the University of California-San Francisco, agreed that the data in the earlier, larger trials need to be revisited.

“I think this study is a good piece of work,” he told Neurology Today in a telephone interview. “We have long been concerned about giving tPA to patients with significant anticoagulant levels with INRs above 1.7, and some people have treated patients with INRs higher than that, but what we have not known is whether there is a relative increase in hemorrhagic transformation in patients with INRs that are 1.7 or below. This is really the first time someone has looked into this,” said Dr. Hemphill, who also directs San Francisco General Hospital's Neurocritical Care Service.

The European Co-operative Acute Stroke Study (ECASS) studies specifically excluded patients taking anticoagulants, as did the original 1995 and 1997 studies by the National Institute of Neurological Disorders and Stroke rt-PA Stroke Study Group. The Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke study and the phase 4 post-marketing study of tPA (Standard Treatment With Alteplase to Reverse Stroke Study) included warfarin-treated patients with INRs less than 1.7, however, the risk of hemorrhagic complications was not reported.

“What I have not seen before are specific data reported like this. Someone should go back and look at the earlier studies — the data are already there,” Dr. Hemphill continued.

Even given the study's small sample size, a 31 percent ICH rate seems high, according to Dr. Hemphill. “This is such a small study that the confidence intervals are really wide. The truth is somewhere between 9 percent and 61 percent, and 31 percent is extremely high. But what they have identified is something that really needs to be studied further.”

And while the findings are disconcerting, he said he doubts they will change which patients receive tPA, at least in the near future.

“It is something to be concerned about, though, and should be mentioned in informed consent discussions with patients and families. Clinicians need to be aware of the new data, but I wouldn't rely on it to deny treatment.”