Can tPA be Used Safely for Acute Ischemic Stroke in Your Emergency Room?

abstracts & commentary

Sources: Albers GW, et al. Intravenous tissue-type plasminogen activator for treatment of acute stroke: The Standard Treatment with Alteplase to Reverse Stroke (STARS) study. JAMA 2000;283:1145-1150; Katzan IL, et al. Use of tissue-type plasminogen activator for acute ischemic stroke: The Cleveland area experience. JAMA 2000;283:1151-1158.

With fda approval of intravenous (iv) tissue plasminogen activator (tPA) in 1996, therapy for acute ischemic stroke moved out of the era of therapeutic nihilism and potentially became a treatable medical emergency. However, with limitations such as a three-hour time window and 10-fold increased risk of intracerebral hemorrhage (ICH) over placebo, this drug has been met with much skepticism and reluctance by the broad neurological community. Among all stroke patients presenting to emergency rooms (ERs) in the United States, only about 2-6% of these receive tPA. This figure rises to only about 10% among the minority of patients who present within three hours. tPA is more likely to be used in academic "stroke centers" than in the community. Recent reports published in the Journal of the American Medical Association (JAMA) suggest that limiting tPA use in this manner may be justified.

Since publication of the National Institute of Neurological Disorders and Stroke (NINDS) trial (N Engl J Med 1995;333:1581-1587), several groups have reported on the routine use of tPA in practice and have shown comparable outcomes and intracerebral hemorrhage (ICH) rates. Most of these studies, however, involved doctors with prior or current experience in acute stroke clinical trials. Albers and colleagues report on one such study. All 83 centers in the STARS study were simultaneously participating in ATLANTIS, a study of tPA in the three- to five-hour time window (which, of note, showed lack of efficacy). In STARS, of the 389 patients enrolled, there was a symptomatic ICH rate of 3.3%, half that of NINDS (6.4%). Protocol violations (previously shown to increase the risks of tPA by as much as fourfold) occurred in 34.7% of patients and were not associated with more hemorrhages. Among 13 patients with symptomatic ICH, five hemorrhages (3.9%) occurred in patients with protocol violations compared with eight (3.1%) among patients without violations. The most common protocol violations were treatment beyond the three-hour window and the use of additional anticoagulants within the first 24 hours. Interestingly, even among these experienced investigators, the "door to needle" time was inversely proportional to the time from symptom onset. Patients presenting close to the three-hour limit were given tPA within minutes. By comparison, those presenting within the first hour were treated up to two hours later, perhaps reflecting a feeling of a luxury of time. Such a luxury does not exist. Subgroup analysis of the NINDS data has shown that the majority of benefit of tPA was seen among patients treated in the 0- to 90-minute time range.

Katzan and colleagues report on data from 3948 strokes at 29 hospitals in Cleveland, reflecting heterogeneous experience from a moderate-sized metropolitan area. Among 70 patients who received tPA, there was a symptomatic hemorrhage rate of 15.7% (nearly 3 times that of NINDS and 5 times that of STARS). In-hospital mortality was 15%, significantly higher than that of matched patients seen within three hours who did not receive tPA (7.2%). Protocol violations occurred in half of the tPA-treated cases, but, as in STARS, these did not significantly affect ICH rates. The overall use of tPA was low. It was given in only 1.8% of all strokes presenting during the one-year study period (10.4% for patients arriving within three hours). It is not clear from Katzan et al’s data why so many potentially eligible patients were not treated.


What explains the unexpectedly high hemorrhage rates in the Cleveland community study? Protocol violations common to less experienced operators (such as excessively dosing tPA in cardiac quantities) do not appear to be the explanation. Unfortunately, the nature of Katzan et al’s data limits further insights. Baseline stroke severity (e.g., in the form of NIH Stroke Scale measurements) is known to affect ICH risk but these data are only available in a minority of the Cleveland patients. Early computerized tomography (CT) changes (such as hypodensity affecting more than one-third of the middle cerebral artery [MCA] territory) are also associated with ICH risk, but no radiological data are presented. Also, because these data analyze only end points such as ICH and in-hospital mortality, we may be missing important clinical benefits such as neurological improvement or reduced long-term disability. Finally, the Cleveland numbers are small, reflecting only 70 tPA-treated cases with 11 hemorrhages, compared with 389 treated patients in STARS and 312 in the NINDS trial.

These data have led to a call by some for centralization of stroke care in major centers akin to a trauma-care network. Time is of the essence, however, and travel can be cumbersome. Perhaps better education of local community practitioners would allow safe administration of this efficacious drug. —azs

The most common protocol violations are:

a. treatment beyond the three-hour window.

b. use of additional anticoagulants within the first 24 hours.

c. not using anticoagulants.

d. a and b

e. None of the above