The natural history of basilar occlusion is generally poor with mortality ranging from 83 to 91 % [40]. Because of this poor natural history, IA thrombolysis has been preferred in patients with acute basilar artery occlusion. Approximately 278 cases have been reported with an overall basilar artery recanalization rate of 60 % [40]. Basilar artery occlusions are usually due to atherothrombosis. There is a high incidence of residual stenosis after basilar artery recanalization which often requires adjuvant therapies including angioplasty and/or stenting, antithrombotic, and antiplatelet agents.

In a compilation of reported cases of vertebrobasilar IA thrombolysis, the mortality in patients failing recanalization was 90 % compared to 31 % mortality in patients achieving at least partial reperfusion [41]. Good outcomes are strongly associated with recanalization after thrombolytic therapy. Hacke et al. [42] described 65 consecutive patients with vertebrobasilar occlusion treated either with local IA UK, or IA SK plus heparin (n = 43), or conventional antiplatelet/anticoagulation therapy (n = 22). The recanalization rate among thrombolysis cases was 44 % (19 of 43). All patients without recanalization died, while 14 of 19 with recanalization survived, 10 with a favorable outcome. The mortality rate with conventional therapy was 86 % compared to 67 % with thrombolysis. The rate of brain hemorrhage with clinical neurological deterioration was 7 % in thrombolysis patients. Schumacher et al. [43] reported 29 patients with vertebrobasilar occlusion of <6 h duration treated with up to 1,500,000 IU local UK plus IV heparin. Recanalization was achieved in 66 %. There was no or minimal deficit in 45 %. The mortality rate was 45 %. Eckert et al. [13] performed IA thrombolysis using 30 mg of rt-PA given over 2 h in combination with 12 h of IV abciximab therapy (0.25 mg/kg bolus followed by 0.125 mcg/kg/min infusion) in 3 patients. Recanalization occurred in 2 of 3 patients, both of which achieved clinical independence at 3 months post-thrombolysis.

Recanalization rates depend upon the location of the vertebrobasilar occlusion. Distal basilar occlusions have higher recanalization rates than proximal occlusions. Emboli often lodge in the distal basilar artery and are easier to lyse than atherosclerosis related thrombi, the usual cause of proximal basilar occlusions [44]. Short segment occlusions are easier to lyse than longer segment occlusions [45]. Younger patients have higher recanalization rates [46], probably due to the increased incidence of embolic occlusions seen in this age group.

The timing of IA vertebrobasilar thrombolysis is often a difficult decision. The presence of coma or tetraparesis for several hours portends a poor prognosis, despite recanalization [42, 44, 46]. Such symptoms do not preclude survival, however, and recovery has been documented after successful recanalization in such patients [44, 45, 47].

The time window for thrombolysis may be longer in the vertebrobasilar circulation. Many series have included patients up 24 h [42], 48 h [48], and even 72 h [45], after symptom onset in patients with stuttering courses. An association between time to treatment and outcome has been suggested [49] but many series do not support this [45, 50, 51]. In fact, in some studies the time to treatment was actually longer in patients who survived or had good outcomes [45, 50]. A longer time window may be due to a higher ischemic tolerance or improved collateralization in the posterior circulation. Cross and colleagues [45], reporting on 20 patients with basilar artery thrombosis who received IA thrombolysis, found that better collateral blood flow was correlated with improved responses to thrombolysis and with longer tolerance of ischemia. Patients with proximal basilar artery thrombosis did not seem to have the same benefit.

Patients with vertebrobasilar ischemia often have chronic atherosclerotic disease which allows collaterals to develop over time. As hypothesized by Cross et al. [45] there may be two distinct populations of patients with vertebrobasilar occlusion. Patients with a progressive stuttering course may have better collateral circulation and have better outcomes despite later treatment than patients with the sudden onset of severe deficits due to poor collaterals who may be brought to treatment earlier.

While some authors believe that patients with brainstem infarction on CT are not candidates for thrombolytic therapy [42, 44], others have found no correlation with neurologic outcome [45, 50]. In two separate series, none of the patients who had CT evidence for brainstem ischemia developed a hemorrhage. However, because of the experience in the anterior circulation, caution should be used when considering thrombolysis in patients with early infarct signs.

The only randomized, controlled, multicenter trials of IA thrombolysis in acute ischemic stroke are PROACT I [10] and PROACT II [1]. In PROACT I, the safety and recanalization efficacy of 6 mg recombinant pro-urokinase (r-proUK) was examined in 40 patients with acute ischemic stroke of less than 6 h duration due to occlusion of the MCA. The control group received IA saline. The recanalization rate was 57.7 % in the r-proUK group and only 14.3 % in the placebo group. Two doses of heparin were used in PROACT I. In the high heparin group (5,000 U bolus followed by 1,000 U per hour infusion) the recanalization rate was 80 %, but the symptomatic ICH rate was 27 %. In the low heparin group, the recanalization rate was 47 % but the ICH rate was decreased to 6 %. Although not a clinical efficacy trial, there appeared to be a 10–12 % increase in excellent outcomes in the IA r-proUK group.

The follow-up clinical efficacy trial, PROACT II [1], was launched in February, 1996 and completed in August, 1998. The results were first reported in February, 1999. PROACT II used an open design with blinded neurological follow-up. The vast majority of patients in PROACT II were IV rt-PA ineligible due to arrival beyond 3 h. Also, mechanical clot disruption was prohibited. Patients were screened with conventional angiography for occlusion of the MCA and had to have a NIHSS score between 4 and 30. The patients in PROACT II had a very high baseline stroke severity score with a median NIHSS of 17. Patients with early signs of an infarct in greater than one third of the MCA territory (ECASS criteria) on the baseline CT scan were excluded from the study. One hundred and eighty (180) patients were then randomized to receive either 9 mg of IA r-proUK plus low dose IV heparin or low dose IV heparin alone. The primary outcome measure was the percent of patients who achieved a modified Rankin score of ≤2 at 90 days, which signified slight or no neurological disability. Secondary measures included the percentage of patients who had a NIHSS ≤1 at 90 days, angiographic recanalization, symptomatic intracerebral hemorrhage, and mortality. The median time from onset of symptoms to initiation of IA thrombolysis was 5.3 h.

In the r-proUK treated group there was a 15 % absolute benefit in the number of patients who achieved a modified Rankin score of ≤2 at 90 days (P = 0.043). Therefore, on average, 7 patients with MCA occlusion would require IA r-proUK for one to benefit. The benefit was most noticeable in patients with a baseline NIHSS between 11 and 20. Recanalization rates were 66 % at 2 h for the treatment group and 18 % for the placebo group (p<0.001). Symptomatic brain hemorrhage occurred in 10 % of the r-proUK group and 2 % of the control group. In PROACT II, as in the NINDS trial, despite the higher early symptomatic brain hemorrhage rate, patients overall benefited from the therapy and there was no excess mortality (r-proUK 24 %, control 27 %).

Aggregate data indicates an 8.3 % risk of symptomatic brain hemorrhage with IA thrombolysis in the carotid territory and a 6.5 % risk in the vertebrobasilar territory [41]. There is no evidence that the rate of symptomatic brain hemorrhage is lower with IA thrombolysis than with IV thrombolysis, but direct comparisons are difficult. In an uncontrolled series, Gönner et al. [33] reported a 4.7 % rate of symptomatic brain hemorrhage in 42 patients treated with IA thrombolysis. This series differed from PROACT II in that only 26 out of the 42 patients received heparin; the remainder received aspirin. The higher rate of intracranial hemorrhage with neurological deterioration with IA rpro-UK in PROACT II (10.2 %) compared to IV rt-PA in NINDS (6.4 %) [2], ATLANTIS (7.2 %) [52] and ECASS II (8.8 %) [53] must be understood within the context of the greater baseline stroke severity, longer time to treatment, and 66 % MCA recanalization rate in PROACT II. The median baseline NIHSS score in ATLANTIS and ECASS II was 11, in NINDS 14 and in PROACT II 17. Greater baseline stroke severity was first associated with increased intracranial hemorrhage risk in NINDS and ECASS I. All symptomatic intracranial hemorrhages in PROACT II occurred in patients with a baseline NIHSS score ≥11, and in NINDS the rate of symptomatic brain hemorrhage in patients with a NIHSS >20 was 18 %.

Although brain hemorrhage complicating thrombolysis for acute stroke likely reflects reperfusion of necrotic tissue, several IV thrombolysis series have found no direct relationship between recanalization and hemorrhage risk [54, 55]. Kidwell et al. [56] retrospectively reviewed 89 patients treated with IA thrombolysis at their center and found that the leading predictors of any hemorrhage in this population were the baseline NIHSS score, platelet count, glucose level and a longer time to recanalization. Also, this uncontrolled series from Kidwell et al. showed a strong trend toward a higher rate of hemorrhagic transformation in patients without recanalization vs. with recanalization (54 % vs. 33 %, respectively; p = 0.1) The amount of ischemic damage is a key factor in the development of brain hemorrhage after thrombolysis induced recanalization. Major early CT changes and severity of the initial neurological deficit, both indicators of the extent of ischemic damage, are some of the best predictors for the risk of hemorrhagic transformation [55, 57].

Several other factors have been associated with hemorrhage after thrombolysis for both stroke and myocardial infarction, including thrombolytic dose [58], blood pressure [55, 59, 60], advanced age, prior head injury [61] and blood glucose >200 mg/dl. Levels of blood matrix metalloproteinase-9 (MMP-9) in patients prior to thrombolytic therapy has been shown to be a risk factor for hemorrhagic transformation [62], although a specific genotypic relationship could not be elucidated [63]. Other studies to identify patients at risk have used advanced T2-weighted MRI sequences such as gradient echo planar imaging to identify patients with a history of microbleeds which may predispose to a higher rate of hemorrhagic transformation during thrombolysis [64]. Adjunctive antithrombotic therapy may also play a role during IA thrombolysis. Age was the most important risk factor in one of the largest series of thrombolysis-related intracranial hemorrhage [61]. A relationship between advanced age and hemorrhage was demonstrated in the NINDS [3] and ECASS trials [53, 57]. Although there is no strict age cutoff for administering thrombolytics for stroke, physicians need to take age into account, especially in patients over age 75 when determining the risk of angiography and IA thrombolysis.