We retrospectively analyzed clinical and imaging data of patients with acute cerebral infarction who were admitted to Chungnam National University Hospital, Daejeon, South Korea between January 2007 and September 2011 and who underwent IAT to recanalize acute occlusions of intracranial arteries. IAT was applied to patients who had a National Institutes of Health Stroke Scale (NIHSS) score of 4 or more, were older than 18 years, arrived at the emergency room within 6 hours of symptom onset, and had angiographically confirmed occlusion of the unilateral MCA and/or terminal internal carotid artery or the basilar artery. Conventional IAT using AMCD was performed from January 2007 to October 2010, while SAT using a retrievable stent (Solitaire AB®, ev3, Irvine, CA, USA) was performed from November 2010 to September 2011. Patients who had acute cerebral infarctions caused by occlusion of the distal common carotid artery, proximal ICA, or extracranial vertebral artery were excluded, as were those who met the exclusion criteria of the National Institute of Neurological Disorders and Stroke protocol for thrombolysis.15 The study protocol was reviewed and approved by the Institutional Review Board of Chungnam National University Hospital.

Brain computed tomography (CT) was initially performed within 3 hours after symptom onset. After confirmation that there was neither hemorrhage nor low density involving greater than one-third of the MCA territory on the CT scan, 0.9 mg/kg recombinant tissue plasminogen activator was infused for 1 hour after a bolus injection of 10% of the calculated amount of recombinant tissue plasminogen activator. We continued AMCD or SAT if the neurological status of the patient did not improve after intravenous therapy and mismatches between diffusion-weighted images and perfusion-weighted images re-mained on magnetic resonance imaging (MRI) scans. When a patient arrived between 3 and 6 hours after symptom onset, AMCD or SAT was performed without intravenous thrombolysis after confirming the presence of DWI-PWI mismatch on MRI and no hemorrhage on CT.

AMCD and SAT have both been used for IAT in our hospital. During the first part of the study period, from January 2007 to October 2010, AMCD was performed in acute-intracranial-occlusion patients. We performed aggressive mechanical catheterization using a microcatheter and/or a J-shaped guide-wire tip within the occluded arterial lumen.4 The efficacy of thrombolysis was increased in some patients by also injecting urokinase. During the second part of the study period, from November 2010 to September 2011, we performed SAT for acute-cerebral-infarction patients. A retrievable stent (Solitaire AB®, ev3) was first deployed within the impacted thrombus at the occluded vessels. After 10-15 minutes the stent was retrieved within a guiding catheter. We then withdrew the entire guiding catheter and simultaneously applied negative pressure with an aspiration device to reduce the likelihood of distal migration of the captured thrombi. Finally, we performed angiography to confirm recanalization of the occluded vessel. If reocclusion occurred within the recanalized artery after SAT, intra-arterial urokinase injection, balloon angioplasty, and/or stenting were performed after SAT.

We compared clinical parameters between patients who received SAT and those who received AMCD. First, to evaluate whether SAT recanalized the occluded vessels more rapidly than AMCD, we compared the door-to-puncture times and puncture-to-recanalization times between the two thrombolysis groups. We also compared the rates of recanalization, which was evaluated using the Thrombolysis in Cerebral Infarction (TICI) Scale,16 with successful recanalization defined as grade 2 or 3 on the TICI Scale.4

Next, to compare differences in early and late functional recoveries between the AMCD and SAT groups, we evaluated the NIHSS and modified Rankin Scale (mRS) scores before and immediately after interventions, at discharge, and at 3 and 6 months after discharge. To evaluate clinical outcomes after IAT, the proportion of good outcomes (mRS score of 0 to 317) was compared between the two groups.

Finally, the safety of SAT and AMCD were compared based on the frequencies and doses of urokinase injections and incidences of hemorrhagic transformation in the two groups. Hemorrhagic transformation was classified into four categories: 1) small petechiae within the infracted lesion, 2) confluent petechiae without space-occupying effects, 3) small hematoma (≤30% of infarcted area) with a slight space-occupying effect, and 4) large hematoma (>30% of the infarcted area) or outside hemorrhage with a substantial space-occupying effect.18,19 Symptomatic intracerebral hemorrhage (ICH) was defined as clinical aggravation (increase in the NIHSS score of ≥4) and simultaneous large hematoma or outside hemorrhage on CT or MRI performed within 24 hours of thrombolysis.20

Changes in the NIHSS score between admission and discharge, and the time intervals from symptom onset or start of intra-arterial interventions to recanalization of the occluded vessels were compared between the SAT and AMCD groups using t-tests. The recanalization rates and hemorrhagic complications of the two groups were compared using Fisher's exact test.

Differences in the clinical outcomes between the two groups were analyzed by comparing the proportion of patients with a good outcome at discharge (mRS score of 0-3) with the proportion of patients discharged, and the proportion of patients with good outcomes at 3 and 6 months after discharge using Fisher's exact test. All statistical analyses were performed using SPSS 19.0 (SPSS Inc., Chicago, IL, USA).

Recanalization was achieved more rapidly from the start of intra-arterial interventions by SAT (69.2±39.6 minutes) than by AMCD (94.4±48.0 minutes, p<0.05). The rate of successful recanalization (TICI-Scale score of 2 or 3) was higher in SAT patients (91%) than in AMCD patients (80%), but it was not statistically significant (p>0.05) (Table 2).

With regard to clinical outcomes at discharge, the improvement in the NIHSS score (SAT, 5.8±6.1; AMCD, 4.7±4.9; p>0.05) and the proportion of good early neurological outcomes, as defined by an NIHSS score of 0 or 1 or an improvement of 8 or more in the NIHSS score (SAT, 27%; AMCD, 31%; p>0.05), were similar in the SAT and IAT groups. The proportion of good outcomes (mRS score of 0 to 3) was similar in the two groups at discharge (SAT, 46%; AMCD, 39%; p>0.05), but was significantly higher in patients who underwent SAT than those that underwent AMCD at both 3 months (SAT, 64%; AMCD, 40%; p<0.05) and 6 months (SAT, 67%; AMCD, 42%; p<0.05) after discharge (Table 2). Interestingly, the improvements in the mRS score were more favorable in the SAT group than in the AMCD group; the proportion of patients with a good outcome (mRS score of 0 to 3) increased from 46% to 67% after 6 months, while the proportion of patients with a not-good outcome (mRS score of 4 to 6) decreased from 54% to 33% during the same time period, despite the AMCD group having similar proportions of patients with good and not-good outcomes (defined based on mRS scores) during the follow-up period (Fig. 1).

Urokinase for recanalizing the occluded vessel was used more frequently in the AMCD group (92%) than in the SAT group (21%). Furthermore, more urokinase was injected in the AMCD group than in the SAT group (median urokinase doses of 430000 IU vs. 270000 IU) (Table 3). The incidence rates of confluent petechiae and smaller hematomas were higher in the SAT group than in the AMCD group (Table 3). However, the incidence of symptomatic ICH with a large hematoma was significantly higher in the AMCD group than in the SAT group (p<0.05). One patient in the SAT group showed symptomatic ICH combined with subarachnoid hemorrhage.

Stent-related complications occurred in three patients in the SAT group (Table 3). Two of these patients experienced accidental fracture of the stent during retrieval occurred at the occluded MCA sites: one underwent balloon dilatation within the remaining stent, while the other underwent intracranial stent deployment within the fractured stent using a metal stent. The intraluminal patency of the ballooned-fractured stent or metal stent was well maintained on follow-up angiography performed 2 days after SAT. The third patient with a stent-related complication experienced distal migration of an initially captured thrombus within the stent occurred. In this patient the thrombus captured at the proximal MCA migrated and impacted the ipsilateral ACA during stent retrieval. The thrombus located in the ACA was immediately and completely removed by repeated SAT, and no further infarction was detected after this intervention.