We tested the safety of IA versus IV rt-PA administration in a rat model of MCAo in an extended ischemic time window. The main finding of the study was that rt-PA administered by IA infusion after 6 hours of MCAo had significantly more intracerebral bleeding than did the same dosages administered IV. Indeed, even 1 mg/kg rt-PA IA caused more severe bleeding than 10 times that dose IV (10 mg/kg). The increased bleeding did not translate into a significant increase in infarct volume. The highest rt-PA dosage by both the IA and IV routes resulted in increased neurological dysfunction.
The increased bleeding with IA administration probably relates to higher concentrations of rt-PA in the local milieu. Also, rt-PA is metabolized rapidly by the liver resulting in a biological plasma half life in the rat of only 1-2 minutes , further exaggerating the difference in concentration after IV versus IA administration.
Cerebral ischemia predisposes the vasculature and surrounding parenchyma to hemorrhage upon vascular recanalization  and the vulnerability of the downstream brain parenchyma increases with longer periods of ischemia. Several pathological mechanisms underlie the ICH caused by rt-PA including, activation of MMP-9, interaction with NMDA receptors, endogenous proteolytic activity, injury to the endothelial junctional proteins contributing to the blood-brain barrier dysfunction, and activation of other proteolytic enzymes such as plasmin . Taken together, it is not surprising that our results support the conclusion that similar dosages of rt-PA delivered IA carry greater ICH risk than when delivered IV.
The duration of the ischemic event may modify the toxicity of rt-PA. Increased bleeding did not occur with IV rt-PA doses as high as 20 mg/kg if administered within 1 hour ischemic onset , suggesting a therapeutic index (TI = Toxic dose ÷ Therapeutic dose) of greater than 2. Whereas doses of rt-PA administered after a 3 hour delay resulted in a dose-dependent (0.9, 9 and 18 mg/kg) increase in hemoglobin extravasation . By extension, in our study, 30 mg/kg rt-PA IV resulted in significant bleeding following 6 hours of ischemia. Taken together, our results and the data of other groups supports the conclusion that bleeding is not induced by substantial doses of rt-PA following short periods (< 3 hours) of ischemia, but that the risk of bleeding increases significantly with longer intervals of occlusion.
Because of the implications of high concentrations of rt-PA when administered IA, we closely mimicked such conditions by use of a snare ligature model, which enabled us to initiate IA rt-PA immediately prior to vascular recanalization. Furthermore, as fibrin degradation products may contribute to rt-PA-associated ischemic damage [34, 35], the snare ligature model allowed us to dissociate ischemic damage caused by rt-PA from that caused by fibrinolytic products.
The clinical therapeutic dose of rt-PA IV for ischemic stroke has been established at 0.9 mg/kg. The dose for rt-PA IA as a stand alone treatment is more variable and ranges from 20 to 60 mg . Our data suggest that to avoid increased ICH risk by the IA route compared to IV, the dose would need to be considerably less, although such conclusions must be tempered by the understanding that our model may not entirely reflect the clinical state in human disease.
The 30 mg/kg rt-PA IA and IV dose groups were included in the study to provide an estimate of the therapeutic index and to exaggerate potential bleeding complications. In 3 of 5 rats, 30 mg/kg IA induced spontaneous bleeding in the brain stem that was not directly affected by the vascular ligation. All 3 rats had damage to the hypothalamus and two of these had seizure activity necessitating euthanasia. As the arterial perforators to the brain stem come off of the Circle of Willis prior to the branching of the MCA, significant blood levels of rt-PA would be present in these arteries following IA administration resulting in brain tissue exposure to high concentrations of rt-PA over the extended duration of administration (60 minutes). Hemorrhagic damage to the brain stem, hypothalamic damage and seizure activity did not occur in any animal dosed with 30 mg/kg rt-PA IV.
In our study, the Modified Bederson Score was used to assess the functional recovery of the rats. Our data suggests that the final behavior of the rats was significantly affected by both dose and route of rt-PA administration and the major driving force was bleeding. The infarct volume was not a major contributor to behavioral function after 6 hours of ischemia.
A limitation to this study was the number of rats (4 out of 5) euthanized in the 30 mg/kg IV group because of unresponsiveness to external stimuli for over 3 hours following completion of the rt-PA infusion. The Bleeding Score for these 4 rats were determined in brain sections after < 3 hours maturation following reflow, which may have underestimated the true extent of bleeding. However, 2 of the 4 rats received a score of 4 because of large space occupying hemorrhages. The other 2 rats had Bleeding Scores of 3 and 2 which were most likely accurate despite the shorter maturation time. This suggests that the premature termination probably had little effect on the Bleeding Score results.
The number of rats prematurely terminated in the both of the 30 mg/kg rt-PA dose groups may have had an impact on the infarct volume data. Of the 10 animals in these 2 groups, 7 were euthanized within 3 hours of dosing, thereby reducing the maturation of the infarct following reflow to less than 3 hours as opposed to ~18 hours. In addition, one rat dosed 30 mg/kg rt-PA IA died overnight. The lack of maturation of the infarct volume may have resulted in an underestimate of the true infarct volume in these rats.
The early termination of the rats in the 30 mg/kg IV group may have lead to an under estimation of the functional recovery of these rats. Two of the rats had surprisingly little bleeding in relation to behavior and thus, may have survived to scheduled termination. Subsequent experience suggests that survival would have been likely. The pronounced depressed behavior of these rats post dosing suggests that rt-PA may have severe intrinsic CNS depression activity at high doses.