Arterial reocclusion in stroke patients treated with intravenous tissue plasminogen activator

Subjects and methods.

We studied consecutive patients treated with standard IV TPA therapy (0.9-mg/kg dose, maximum 90 mg, 10% bolus, 90% continuous infusion) that was initiated within the first 3 hours after symptom onset.1 Patients were included in this study if they had an M1 or M2 middle cerebral artery (MCA) occlusion on pretreatment TCD according to previously validated criteria.7 Our TCD criteria for the proximal MCA occlusion have 91% sensitivity and 98% specificity compared with angiography7 and a prospective multicenter validation study is underway (A.M. Demchuk et al., unpublished data).

Before TPA bolus, an experienced sonographer identified residual flow signals at the presumed thrombus location using the Thrombolysis in Brain Ischemia (TIBI) flow-grading system.8 A 2-MHz transducer was positioned at a constant angle of insonation with a standard head frame (Marc series; Spencer Technologies, Seattle, WA). The depth that displayed the worst residual TIBI flow signal was selected, and TCD monitoring was performed according to an institutional review board-approved protocol (University of Texas Committee for Protection of Human Subjects). Patients were continuously monitored with TCD starting before bolus until 2 hours after infusion.

Arterial recanalization on TCD was determined using previously validated criteria.9 Recanalization on TCD was graded as complete, partial, or none similar to the Thrombolysis in Myocardial Infarction (TIMI) criteria.10 In brief, complete recanalization was diagnosed when a normal waveform or a low-resistance stenotic signal appeared at the selected depth of insonation, suggesting low resistance of the distal circulatory bed. These flow findings correlate with unobstructed passage of contrast agent on angiography.9 Partial recanalization was diagnosed if the abnormal signals (high resistance dampened signals or flattening of the systolic upstroke with “blunted waveform) were still seen at the distal portion. No change in the abnormal flow signals indicated that no recanalization has occurred, with minimal flow signal or absent flow corresponding to complete arterial occlusion on angiography. Our TCD criteria for thrombolysis in the proximal MCA to predict complete recanalization on angiography had 91% sensitivity and 93% specificity.9

All patients were monitored under direct supervision of an experienced sonographer who interpreted all waveforms from the real-time display at bedside. Standard monitoring of vital signs (i.e., blood pressure, pulse oximetry, and heart rhythm) was performed during TPA therapy in all patients to exclude systemic reasons for TCD flow changes.

Reocclusion was first suspected by a sonographer when a decrease in the flow signal by ≥1 TIBI grade was seen on TCD display and vital signs were stable (figure 1, top, white arrow). Worsening of flow signals by 1 TIBI grade indicates an increase in resistance to flow and therefore progression in the degree of arterial obstruction. Treating physicians were immediately informed about these findings. Reocclusion was then diagnosed by a treating physician if repeat CT showed no hemorrhage after completion of TPA infusion (see figure 1, bottom). Indications for repeat CT scanning included stroke symptom progression, deterioration following initial improvement by ≥4 NIHSS points, or evidence of flow worsening on TCD with or without clinical changes.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Early arterial reocclusion on transcranial Doppler (TCD). (Top) Images show TCD monitoring spectra of the distal right M1 middle cerebral artery (MCA). Frame 1 shows a blunted (Thrombolysis in Brain Ischemia [TIBI] grade II) residual flow signal suggesting MCA near occlusion (sketch) at the time of tissue plasminogen activator (TPA) bolus (NIH Stroke Scale [NIHSS] score of 15 at 10:12 am). Frame 2 shows a stenotic (TIBI grade IV) flow signal with clusters of microemboli, suggesting MCA thrombus dissolution. Frames 3 to 4 show MCA flow improvement to normal TIBI V flow grade, which indicates complete M1 MCA recanalization. In frame 4, white arrow points to a sudden drop in the MCA velocity that started reocclusion while the heart rate and blood pressure remained stable. This sudden M1 MCA obstruction was likely caused by a thrombus that lodged in the proximal MCA and then propagated to the site of insonation (next frame). Frame 5 shows a blunted (TIBI grade II) flow with clusters of high-intensity emboli and bruits. Frame 6 shows a minimal (TIBI grade I) flow signal that suggests a complete M1 MCA reocclusion owing to the absence of diastolic flow. Note continuing appearance of emboli during the last cardiac cycle. The NIHSS scores obtained at time-corresponding intervals showed deterioration following improvement (15 to 3 to 15). (Bottom) Images show noncontrast pre-TPA CT and post-TPA CT scans with no signs of intracerebral hemorrhage as a cause of deterioration following improvement. Digital subtraction angiography (DSA) was performed 38 minutes after the end of TPA infusion and showed a Thrombolysis in Myocardial Infarction grade I right M1 MCA occlusion.

As treating physicians were advised when reocclusion occurred, patients who subsequently underwent experimental intra-arterial thrombolysis were excluded from the study to minimize selection bias. Also, our protocol requires independent scoring of the neurologic deficit and flow grade assessments on TCD to minimize bias in clinical patient evaluations; that is, TCD information was disclosed after stroke severity assessment.

Neurologic status was repeatedly assessed during the first 2 hours after TPA bolus by a neurologist who was not involved in TCD but who was informed about worsening of flow signals on TCD, if these occurred. All neurologists who performed serial neurologic examinations in the emergency room were certified in the NIHSS scoring. The NIHSS scores at 24 hours and modified Rankin scores (mRS) at 3 months were obtained by a neurologist who was not aware of TCD findings and the purposes of this study.

Clinical outcome measurements included the NIHSS scores at 2 hours after TPA bolus and at 24 hours. “Dramatic recovery” was defined as decrease of the total NIHSS score to ≤3 points,11 and “early neurologic improvement” was defined as reduction by ≥10 points in the total NIHSS score.12 Deterioration following improvement was determined as deterioration of the neurologic deficit by ≥4 NIHSS points after any initial improvement. Good long-term outcome was defined as mRS scores of 0 to 1.6 Statistical analysis included descriptive statistic and χ² and two-sample Student’s t-tests to compare patient groups with normal data distribution.

Results.

A total of 60 patients with a proximal MCA occlusion on prebolus TCD were treated with standard IV TPA therapy (35 men, 25 women; mean age 70 ± 14 years, median 71 years, range 42 to 93 years). These patients had median prebolus NIHSS scores of 16 points (range 6 to 28 points, 90% with ≥10 points). TPA bolus was given at 130 ± 32 minutes (median 120 minutes), and TPA infusion was initiated within the first 2 hours after symptom onset in 58% of all patients.

Recanalization on TCD during the first 2 hours after TPA bolus was complete in 18 patients (30%) and partial in 29 patients (48%). No recanalization was found in 13 patients (22%). Reocclusion occurred in 16 patients. In the reocclusion group, four patients (25%) were spontaneously recanalizing and developed reocclusion during diagnostic TCD performed within 5 to 25 minutes before TPA bolus, three patients (19%) had reocclusion by 30 minutes after bolus, three patients (19%) had reocclusion by the end of TPA infusion, and six patients (37%) had reocclusion within 60 to 120 minutes (figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Timing of early reocclusion in tissue plasminogen activator (TPA)-treated patients. Reocclusion was detected in four patients (25%) before TPA bolus, three patients (19%) by 30 minutes after bolus, three patients (19%) by the end of TPA infusion, and in six patients (37%) within 60 to 120 minutes.

Reocclusion occurred in 34% of patients (16/47) with initial complete or partial recanalization on TCD: Reocclusion occurred in 4 of 18 patients with complete (TIBI grade IV to V) recanalization (22%) and in 12 of 29 patients with partial (TIBI grade II to III) recanalization (41%). Prior to reocclusion, those patients had earlier median timing of recanalization compared with those who recanalized without reocclusion: 130 versus 180 minutes after stroke onset (p = 0.01). After reocclusion occurred, only 4 of these 16 patients (25%) had subsequent complete recanalization within 2 hours of TPA bolus, 3 of 16 patients (19%) had partial recanalization, and 9 of 16 patients (56%) had no recanalization.

Patients with reocclusion tended to have lower median prebolus NIHSS scores of 13.5 points compared with 17 points in the rest (p = 0.1, NS). However, at 2 and 24 hours, their NIHSS scores were higher than in those who recanalized without reocclusion: 14 versus 9 points (p ≤ 0.04) and 16 versus 6 points (p ≤ 0.03). Also, deterioration following improvement within the first 24 hours was observed in 12 (20%) of all study patients, and reocclusion was seen in 8 of 12 or 67% of these patients (table).

As many as 50% of patients in the reocclusion group had deterioration on NIHSS by ≥4 points (8/16) compared with 10% in the rest of the study patients (4/44) (p < 0.05; see the table). Symptomatic intracerebral hemorrhage within the first 72 hours occurred in 3 of 60 patients (5%): 2 in the reocclusion group (12.5%) and 1 in the non-reocclusion group (2%). In-hospital mortality was 25% in the reocclusion group and 3% in non-reocclusion group (p < 0.0001).

At discharge, stroke mechanisms in reocclusion patients were as follows: six atherothrombotic, five cardioembolic, three of other etiologies (dissections and hypercoagulable state), and two undetermined. In the rest (n = 44), 11 were atherothrombotic, 19 cardioembolic, 1 dissection, and 13 undetermined (p = 0.1, NS).

Long-term follow-up was completed in 80% of all studied patients: 12 with no recanalization, 12 with reocclusion, and 24 with stable recanalization. The main reason for patients being lost for follow-up was unavailability of the next-of-kin phone numbers or new addresses if patients moved. At 3 months, good outcome (mRS score of 0 to 1) was achieved by 8% of patients with no recanalization, by 33% of patients with reocclusion, and by 50% of patients with stable recanalization (p ≤ 0.05; figure 3). Mortality at follow-up was 42% in patients with no recanalization, 33% in patients with reocclusion, and 8% in patients with stable recanalization (p ≤ 0.05).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Pretreatment stroke severity, clinical improvement, and long-term outcome in patients with no recanalization, reocclusion, and stable recanalization. Based on transcranial Doppler (TCD) monitoring within 2 hours after tissue plasminogen activator (TPA) bolus, patients were divided into the following groups: No Recanalization (n = 13), Early Reocclusion (n = 16), Stable Recanalization (n = 31). Bar graphs indicate stroke severity at pretreatment (b), 2 hours, and 24 hours after stroke onset expressed as NIH Stroke Scale (NIHSS) scores 0 to 3, 4 to 9, and ≥10 points. Long-term outcome at 3 months was obtained in 80% of study patients (no recanalization [n = 12], early reocclusion [n = 12], stable recanalization [n = 24]): modified Rankin score (mRS) 0 to 1, 2 to 3, 4 to 5 points, and death (black squares).

Among patients lost for long-term follow-up (n = 12), discharge information was as follows: One patient with no recanalization was transferred to a skilled nursing facility with an NIHSS score of 20 points. In the reocclusion group, one patient was discharged with no deficit, one patient went home after in-hospital rehabilitation (NIHSS score of 8 points), and two were transferred to a skilled nursing facility (NIHSS scores of 19 and 25 points). In the group of patients with stable recanalization, two went home with no or minimal deficits (NIHSS scores of 0 and 2 points), one patient went home after in-hospital rehabilitation (NIHSS score of 8 points), and one was transferred to a skilled nursing facility (NIHSS score of 20 points).

Among those four patients who developed reocclusion before TPA bolus, pretreatment NIHSS scores were 15 to 28 points. Only one of these patients had a favorable outcome at 3 months, whereas the rest had persistent severe neurologic deficits and poor outcomes (three patients had mRS scores of 4 to 5 and one patient died during the follow-up period).

Discussion.

Our study showed a high rate of early arterial reocclusion on TCD with standard IV TPA therapy for MCA occlusion. Fully one-third of patients with early recanalization experienced reocclusion within 2 hours of TPA bolus. Also, two-thirds of patients with deterioration following improvement experienced early reocclusion. Our results suggest that early reocclusion may be the most common mechanism of early clinical fluctuation and worsening after thrombolytic therapy for stroke.

Although deterioration following improvement was not reported for the ECASS-I study patients, the overall incidence of early stroke progression was 38%.2 Davalos et al. showed that focal hypodensity on CT, hyperdense MCA sign, and longer treatment delays were independent prognostic factors for early stroke progression.2 Our study population had a 20% rate of deterioration by ≥4 NIHSS points, and it is different from the ECASS-I cohort because all patients in our study were treated within the first 3 hours after stroke onset. Our results, however, are in agreement with ECASS-I observations that hemorrhagic transformation is an unlikely cause of early stroke progression.2

Our study showed that early and stable recanalization predicts clinical improvement and good long-term outcome in patients with MCA occlusion, similar to a previous report.13 However, unlike our results, a previous study did not show that occlusion propagation on serial TCD is related to clinical worsening in stroke patients.4 Our patient population and methods are different because 58% of our patients were studied under 2 hours of stroke onset and we used more detailed definitions of acute arterial obstruction and residual flow signals. Also, continuous TCD monitoring as opposed to serial diagnostic TCD offers high-time resolution to assess recanalization and reocclusion that may take only a few seconds or minutes to occur.6,14⇓

One intriguing observation in our study is that patients with early reocclusion have better long-term outcomes than patients with no early recanalization on TCD. One explanation may be that prior to reocclusion, these patients may have had some degree of early recanalization and reperfusion of the penumbral areas, resulting in additional time to tolerate further ischemia.15

Our study showed that 25% of early reocclusions occur even before the TPA bolus. This finding probably reflects the dynamic stepwise progression of MCA thrombosis to a complete occlusion within the first 3 hours after stroke onset. This finding also suggests that there is no point in delaying thrombolysis in patients with fluctuating clinical course and TCD demonstration of occlusion because further stroke progression can be expected in many of these patients. Only 25% of patients who experienced reocclusion developed subsequent complete recanalization within 2 hours of TPA bolus. However, the number of patients in our study was too small to draw a definitive conclusion about the prognostic significance of recanalization after reocclusion.

Our study has other limitations as no immediate angiographic tests were implemented to verify TCD flow changes. Also, a relatively small number of patients was studied, and the long-term follow-up was successfully completed in only 80% of our study patients. Our TCD findings may not be generalized to centers with less experienced sonographers and implementation of contrast-enhanced transcranial duplex imaging may yield different results.

If stable MCA recanalization was achieved within 2 hours of TPA bolus, 50% of these patients had mRS scores of 0 to 1 and 66% had mRS scores of 0 to 2 at 3 months. These tentative long-term data indicate that TPA-associated thrombolysis of a proximal MCA occlusion can lead to recovery sustained at 3 months. Our data showed that if patients fail to recanalize early after TPA bolus, they are at a greater risk for poor outcome. This population of patients might be good candidates for additional clot-disrupting interventions such as intra-arterial lysis.16

A final implication of our data is that outcome of IV TPA therapy might be improved if reocclusion could be prevented. As reocclusion was observed up to 120 minutes after TPA bolus, perhaps more sustained dosing of a longer-acting lytic drug or coupling lytic therapy with antiplatelet, antithrombin, or other antithrombotic therapy would prevent reocclusion and improve outcome. Only a few series of patients have been reported when TPA was combined with heparin, but no data are available on reocclusion rates.17,18⇓ Heparin is also routinely used with intra-arterial therapy,19 and hemorrhage rates were closely linked to the dose of heparin. Further studies of efforts to prevent reocclusion are needed. However, any studies combining a lytic drug with antithrombotic therapy will need careful assessment of added risk versus added benefit.