t-PA Use Safe in EDs Without Specialized Stroke Teams

SAN FRANCISCO—Emergency physicians at community hospitals can use tissue plasminogen activator (t-PA) with reasonable safety, according to Phillip A. Scott, MD, and colleagues at the University of Michigan Health System. Their retrospective observational study, Safety of Community Use of tPA (SCUT), was presented at the 2007 International Stroke Conference. Dr. Scott told Neurology Reviews that the SCUT study demonstrated on multiple measures that “t-PA given with routine specialist support, but without an acute stroke team, is as safe as in the NINDS trial.”

Putting t-PA to the Test

In the SCUT study, 273 patients were treated with t-PA for acute stroke at one urban, one rural, and two suburban Michigan hospitals from January 1996 to January 2005. Mortality at one year, the primary end point, was 28% and was not statistically different from that in the randomized, placebo-controlled NINDS rt-PA Stroke Study, which was 24% with t-PA and 28% in the placebo group.

The SCUT study looked at the rate of intracerebral hemorrhage (ICH) as a secondary outcome measure and found that 10% of patients had experienced any ICH within 36 hours of t-PA, and 6.6% had symptomatic ICH or significant ICH on CT. The rate of total ICH was slightly lower than the rate reported in the NINDS study (relative risk, 0.94), but the investigators noted that “this may be an artifact resulting from mandatory repeat CT scanning in the NINDS study but not in clinical practice.” The rates of symptomatic ICH, which they considered more important, were comparable between the SCUT study (5% to 7%) and the NINDS rt-PA Stroke Study (6% to 7%).

The one finding of possible concern regarding community use of t-PA was the identification of protocol violations in 26% of cases in the SCUT trial. Of the 71 patients who represented protocol violations, 61% were treated more than three hours after stroke onset, while the rest of the patients received treatment despite having met exclusion criteria for hypertension (11%), thrombocytopenia (7%), hypoglycemia (7%), and seizure (7%)

It’s All in the Timing

According to Dr. Scott, one of the greatest obstacles to using t-PA in community emergency departments is that the time from stroke onset to arrival at the hospital is often too long. “Poor patient re­c­og­nition of the signs and symptoms of stroke remains a daunting challenge. A stroke doesn’t hurt, unlike chest pain and heart attacks, so people wait to see if it will go away,” he said. In the SCUT study, time from stroke onset to treatment was a mean of 154 minutes, and use of t-PA more than three hours after stroke onset accounted for more than half of the protocol violations. Dr. Scott explained, “These time violations generally were of small order and included documentation of permission by the family.” He further noted that the exact time of stroke onset is often difficult to determine, so the SCUT study used the most conservative estimates.

In incidences when time of onset is not clearly defined and patients present close to the three-hour cutoff, physicians may be tempted to try t-PA, particularly in light of the ATLANTIS study, which found that the risk of ICH  when t-PA is given three to five hours after stroke onset is about the same as when it is given before three hours (although there was no benefit associated with the later administration).

Nonetheless, Dr. Scott emphasized that he is not in favor of flexing the criteria for use of t-PA. In fact, he said, “It is important that guidelines be strictly followed,” and he does not recommend “even small and well-documented deviations” until further data on safety emerge to allow such expansion of the criteria. Dr. Scott stressed that even when t-PA is used properly in patients who meet accepted criteria, there is still a 6% risk of ICH. He is concerned that physicians who see ICH after using t-PA for patients who fall outside of recommended guidelines will become reluctant to use t-PA even under the more appropriate circumstances.

The focus should be on putting systems in place for using t-PA correctly, Dr. Scott believes. He emphasized that “earlier initiation of t-PA is better, and that comes with continually improving your systems.” He added, “Notoriously few hospitals are ready day-to-day to handle acute stroke with t-PA …, and it is difficult to do on an ad hoc basis.”

Whereas the NINDS study provided prospective randomized data confirming the efficacy of t-PA, it was not designed to evaluate the safety of t-PA in community settings. Dr. Scott concluded that results of the SCUT study “go a long way toward reassuring emergency physicians that with defined protocols and routine neurologic consultation arrangements, t-PA can be given safely.”

NR