Stroke treatment took a giant step forward in 1995 with publication of the National Institute for Neurological Disorders and Stroke (NINDS) study, which established intravenous tPA as an effective treatment for ischemic stroke.1 However, full recanalization rates with intravenous tPA are "terrible," Broderick said. In an attempt to address residual clot more directly, researchers undertook the Prolyse in Acute Thromboembolism 2 (PROACT-2) study and showed that introduction of the thrombolytic, in that case prourokinase, directly into the clot could successfully open large occlusions in the middle cerebral artery, some of the most catastrophic strokes, and lengthen the time in which treatment could be safely offered from 3 hours with IV tPA to 6 hours with IA thrombolysis.2

In the IMS study, investigatorsmany of whom participated in the original NINDS trialaimed to combine these 2 approaches in an attempt to improve canalization of the arteries, thereby hoping to improve outcome. The primary outcome of interest was safety, but they also intended to look at outcomes to determine whether a randomized trial would be futile, by comparing these with outcomes in tPA-treated patients in NINDS.

A total of 80 patients with moderate to severe acute ischemic stroke (NIH Stroke Scale scores of >10) presenting within 3 hours of symptom onset were enrolled from 17 centers in the US and Canada. Each received IV tPA in a dose of 0.6 mg/kg. After 30 minutes, patients were taken to the cath lab: if angiogram showed residual clot, then patients were treated with IA tPA, 2 mg distal to the clot and 9 mg/hour for 2 hours, for a 20-mg maximum dose. Investigators were also allowed to manipulate the clot using the catheter guidewire.

Primary safety measures presented here were 3-month mortality, bleeding complications, procedure-related complications, new stroke, and other serious adverse events. Bleeding complications were defined as symptomatic intracerebral hemorrhage (ICH) and severe systemic bleeding that required 3 or more units of packed red blood cells or major surgical intervention. Outcomes were compared against historical controls taken from the NINDS study, both placebo- and tPA-treated patients.

Broderick reported that mortality at 3 months was nonsignificantly lower in patients treated with the combination compared with both NINDS placebo patients and those receiving tPA in that trial. Symptomatic ICH, the most dreaded of bleeding complications, was similar in IMS compared with NINDS tPA-treated patients, although, not surprisingly, significantly higher than in placebo-treated NINDS patients. Serious bleeding was also nonsignificantly higher in IMS compared with both NINDS groups.

While symptomatic ICH was quite similar between patients in IMS and NINDS who received tPA, asymptomatic ICH was significantly higher with the combination of IV-IA tPA. However, in the PROACT-2 study, where patients were also treated percutaneously, the rate of symptomatic ICH was higher than in IMS, as was the rate of asymptomatic ICH.

The higher rates of asymptomatic ICH may relate to a greater severity of initial brain injury in these patients compared with the IV-tPA-treated NINDS patients, a higher quality of CT imaging 10 years later, use of heparin in the intra-arterial protocol, and "probably most important," may reflect the early evolution of hemorrhagic changes commonly seen in large cerebral infarctions, he said.

Other safety end points: 2 patients had retroperitoneal hematoma requiring greater than 3 units transfusion; 1 had groin and multiple areas of bruising requiring 2 units of red blood cells; and 1 had severe oozing, although transfusion was not required.

Intracranial perforation during the procedure occurred in 2 patients, seen by contrast dye extravasation, but the patients showed no clinical change. There were 5 cases of severe edema, resulting in 2 of the deaths. New strokes occurred in 4 patients; 2 were recurrent ischemic strokes due to a cardiac source; and 2 patients had delayed intracerebral hemorrhage: 1 a postperfusion hemorrhage related to carotid endarterectomy at day 21, and 1 occurred at about 2 months, related to an INR of 6.