Author + information
- Received July 13, 2005
- Revision received February 28, 2006
- Accepted March 2, 2006
- Published online July 18, 2006.
- Arshad Ali, MD⁎,⁎ (, )
- David Cox, MD†,
- Nabil Dib, MD‡,
- Bruce Brodie, MD§,
- Daniel Berman, MD∥,
- Navin Gupta, MD§,
- Kevin Browne, MD¶,
- Robert Iwaoka, MD†,
- Michael Azrin, MD#,
- Dwight Stapleton, MD⁎,
- Cindy Setum, PhD⁎⁎,
- Jeffrey Popma, MD††,
- AIMI Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Arshad Ali, Williamsport Hospital, Williamsport, Pennsylvania 17701.
Objectives The goal of this work was to determine whether rheolytic thrombectomy (RT) as an adjunct to primary percutaneous coronary intervention (PCI) reduces infarction size and improves myocardial perfusion during treatment of ST-segment elevation myocardial infarction (STEMI).
Background Primary PCI for STEMI achieves brisk epicardial flow in most patients, but myocardial perfusion often remains suboptimal. Distal embolization of thrombus during treatment may be a contributing factor.
Methods This prospective, multicenter trial enrolled 480 patients presenting within 12 h of symptom onset and randomized to treatment with RT as an adjunct to PCI (n = 240) or to PCI alone (n = 240). Visible thrombus was not required. The primary end point was infarct size measured by sestamibi imaging at 14 to 28 days. Secondary end points included final Thrombolysis In Myocardial Infarction (TIMI) flow grade, tissue myocardial perfusion (TMP) blush, ST-segment resolution, and major adverse cardiac events (MACE), defined as the occurrence of death, new Q-wave myocardial infarction, emergent coronary artery bypass grafting, target lesion revascularization, stroke, or stent thrombosis at 30 days.
Results Final infarct size was higher in the adjunct RT group compared with PCI alone (9.8 ± 10.9% vs. 12.5 ± 12.13%; p = 0.03). Final TIMI flow grade 3 was lower in the adjunct RT group (91.8% vs. 97.0% in the PCI alone group; p < 0.02), although fewer patients had baseline TIMI flow grade 3 in the adjunct RT group (44% vs. 63% in the PCI alone group; p < 0.05). There were no significant differences in TMP blush scores or ST-segment resolution. Thirty-day MACE was higher in the adjunct RT group (6.7% vs. 1.7% in the PCI alone group; p = 0.01), a difference primarily driven by very low mortality rate in patients treated with PCI alone (0.8% vs. 4.6% in patients treated with adjunct RT; p = 0.02).
Conclusions Despite effective thrombus removal, RT with primary PCI did not reduce infarct size or improve TIMI flow grade, TMP blush, ST-segment resolution, or 30-day MACE.
Primary percutaneous coronary intervention (PCI) in patients with ST-segment elevation myocardial infarction (STEMI) results in higher coronary reperfusion rates and improved clinical outcomes than fibrinolytic therapy (1). Despite its value, conventional PCI using balloons and stents alone may dislodge thrombotic material adherent to the ruptured plaque, causing distal embolization and microcirculatory impairment and limiting myocardial salvage. Reduced myocardial perfusion despite normal coronary flow is associated with a reduced long-term survival, and can be detected using a variety of methods, including angiographic perfusion grades, persistence in the ST-segment elevation, and noninvasive imaging modalities, such as sestamibi scintigraphy (2,3).
Rheolytic thrombectomy (RT) with the AngioJet catheter (Possis Medical, Inc., Minneapolis, Minnesota) is a proven technique for the removal of thrombus. As shown in the VEGAS-2 (Vein Graft AngioJet) trial, adjunct RT was safer and more effective than an extended local infusion of urokinase for the treatment of angiographically evident thrombus in saphenous vein grafts (SVGs) and native coronary arteries (4). As a result of this study, the AngioJet RT catheter (Possis Medical, Inc.) was approved for use in the U.S. on March 12, 1999, for treatment in thrombus-containing native coronary arteries and SVGs. Plaque rupture and angiographic thrombus formation are central to the pathogenesis of STEMI, but the extent of the thrombus burden is often not apparent at the time of initial angiography. Dislodgement of the thrombotic material during PCI and the ensuing complications of “no reflow” and distal embolization may increase the myocardial infarction (MI) size. Based on a favorable initial experience using adjunct RT in patients with acute MI (5,6), we sought to determine whether RT as an adjunct to PCI in treatment of STEMI would lessen the degree of myocardial necrosis, as assessed by sestamibi infarct size and myocardial perfusion measured by tissue myocardial perfusion (TMP) blush and ST-segment resolution.
Study design and patients
The AIMI (AngioJet Rheolytic Thrombectomy In Patients Undergoing Primary Angioplasty for Acute Myocardial Infarction) study was a prospective, multicenter trial that randomly assigned patients to treatment with RT as an adjunct to conventional PCI or to conventional PCI alone. Patients were considered eligible for enrollment if they were >18 years of age; presented with anterior MI (new ST-segment elevation of >1 mm in at least 2 contiguous leads within V1to V6) or large inferior MI (new ST-segment elevation of >1 mm in 2 contiguous leads of II, III, and aVF); presented within 12 h of symptom onset; and had a reference coronary artery >2.0 mm in diameter. Patients were excluded if they had known prior ejection fraction <35%, cardiogenic shock (systolic blood pressure <80 mm Hg and requiring inotropic support), contraindication to treatment with glycoprotein IIb/IIIa inhibitors, major surgery within the preceding 6 weeks, or history of stroke within 30 days or any history of hemorrhagic stroke. Study enrollment did not require angiographically visible thrombus. Randomization occurred after angiographic confirmation of the reference coronary artery diameter.
All participating study sites received approval from their local hospital investigational review boards before beginning enrollment. Informed consent was obtained from all patients before their inclusion in the study.
Description of the RT and PCI procedure
The RT system consists of a drive unit control console, a disposable pump set, and a disposable single-use RT catheter, as described previously (7). The 5-F LF140 RT catheter (Possis Medical, Inc.) was initially used in 20 randomized patients and was replaced with the 4-F XMI RT catheter (Possis Medical, Inc.) for all remaining patients. Both catheter models are indicated for intracoronary use and were to be used in the study in accordance with the approved instructions for use (8).
After diagnostic angiography confirmed vessel size >2.0 mm, patients randomized to RT received treatment. Placement of a temporary pacing wire was recommended before initiating treatment. The RT catheter was advanced over a 0.014-inch guidewire that had been positioned across the target lesion and into the distal coronary artery. The catheter was activated and 1 or more passes through the target lesion were made, as assessed by the operator. In lesions without clear angiographically evident thrombus, a minimum of 3 slow passes through and at least 10 mm distal to the lesion were to be made.
Definitive PCI with balloon angioplasty and/or stenting was performed immediately after RT. Control patients were treated with standard PCI without RT. Choice of coronary stents was at the discretion of individual operators. Use of embolic protection devices or other thrombectomy devices was not allowed. Standard 12-lead electrocardiograms (ECG) were obtained at the time of initial evaluation and at 90 min after the procedure (ECGs obtained at up to 180 min were allowed).
Concomitant medical therapy and follow-up
Eptifibatide (Millennium Therapeutics, Inc., Cambridge, Massachusetts) was started either in the emergency room or administered in the catheterization laboratory for all patients (9). All patients received oral aspirin 325 mg, clopidogrel 300 mg loading dose was given before the procedure, and clopidogrel 75 mg daily was continued for at least 4 weeks thereafter. Ticlopidine, 500 mg loading dose and 250 mg twice a day dose was given in the event of intolerance to clopidogrel. Unfractionated heparin was administered during the procedure to attain an activated clotting time of >250 s. Use of beta-adrenergic blockers, nitroglycerin, and angiotensin-converting enzyme inhibitors before and after the procedure was at the discretion of individual physicians.
Study end points and analysis
The primary end point was infarct size measured by Tc 99m sestamibi imaging at 14 to 28 days after procedure. Secondary angiographic end points included post-procedure Thrombolysis In Myocardial Infarction (TIMI) flow grade, TMP blush, corrected TIMI frame count, and procedural complications. Secondary clinical end points included the frequency of complete (>70%) ST-segment resolution 90 min after PCI, left ventricular ejection fraction (acquired at time of nuclear imaging), and 30-day major adverse cardiac events (MACE), defined as the occurrence of death, new Q-wave MI, emergent coronary artery bypass grafting (CABG), target lesion revascularization (TLR), stroke, or stent thrombosis. Q-wave MI was defined as MI with development of new Q waves in 2 or more leads, and elevated creatine kinase-MB. Emergent CABG was defined as surgery performed on an urgent basis within 24 h of the index PCI, for recurrent ischemia.
An independent clinical events committee adjudicated MACE, perforation, pericardial hemorrhage, tamponade, and device failures. A Data and Safety Monitoring Board (DSMB) assessed blinded randomized data at prespecified interim reviews using predetermined stopping rules for demonstrated superiority in 1 treatment group. After each interim review, the DSMB recommended that the study continue without modification.
Infarct size measurement
A resting Tc-99m sestamibi-gated single-photon emission computed tomography (SPECT) myocardial perfusion imaging protocol was performed, after validation of each participating site by the Nuclear Core Laboratory (Cedars-Sinai Medical Center, Los Angeles, California) before study enrollment. All raw acquisition data were uniformity- and center-of-rotation-corrected and forwarded to the Core Laboratory for processing by staff blinded to the treatment assignment. The total perfusion deficit was determined using a QPS module that compares the amount of radioactivity on a pixel-by-pixel basis within the left ventricular myocardium to the lower limit of normal distribution derived in patients with a low likelihood of coronary artery disease (10), and was used as the measure of the infarct size. Left ventricular ejection fraction, end-diastolic and end-systolic volumes were measured using the software program QGS (Cedars-Sinai Medical Center) (11).
Standard image acquisition was performed at the clinical sites using 2 or more angiographic projections of the coronary occlusion. Angiograms were to be obtained at baseline, after guidewire recanalization, after RT, and after stent deployment, and then forwarded to the Angiographic Core Laboratory (Brigham and Women’s Hospital, Inc., Boston, Massachusetts). Thrombolysis in Myocardial Infarction flow grades, TMP blush grades, and corrected TIMI frame counts were determined according to previously defined criteria (12–14). The TIMI frame count was corrected for the frame acquisition rate but not for the vessel location. For TIMI flow grade 0/1 at baseline, 150 frames were imputed. Thrombus was graded using an ordinal scale: grade 0: no thrombus present; grade 1: possible thrombus, characteristics include reduced contrast density, haziness, irregular lesion contour or a smooth convex meniscus at the site of total occlusion; grade 2: small thrombus, greater dimension of thrombus is <1/2 the vessel diameter; grade 3: moderate thrombus, greater linear dimension is >1/2 but <2 vessel diameters; grade 4: large thrombus, largest dimension is >2 vessel diameters; grade 5: total vessel occlusion.
Using the contrast-filled injection catheter as the calibration source, quantitative angiographic analysis was performed using a validated automated edge detection algorithm (Medis CMS, Leiden, the Netherlands). Selected images for analysis were identified using angiographic projections that demonstrated the stenosis in an unforeshortened view, minimized the degree of vessel overlap, and displayed the stenosis in its “sharpest and tightest” view. A 5-mm segment of reference diameter proximal and distal to the stenosis was used to calculate the average reference vessel diameter at baseline and after stent dilatation. Minimal lumen diameter (MLD) was measured at these same points. Total occlusions were assigned a MLD of 0 mm and a 100% diameter stenosis.
Electrocardiograms were forwarded to the ECG Core Laboratory (Guthrie Clinic/Robert Packer Hospital, Sayre, Pennsylvania) and analyzed by reviewers blinded to treatment assignment and angiographic and clinical end points, using a hand-held electronic caliper. The ST-segment was measured 40 ms after the J point, and the sum of ST-segment deviation was determined before and after intervention. The percent resolution of ST-segment deviation was calculated and categorized using the Schröder classification as complete (>70%), partial (30% to 70%), or absent (<30%) (15).
All continuous measurements were summarized descriptively by treatment group; ttests were used to test for differences in treatment means. Dichotomous outcomes were presented as proportions of subjects per treatment group and tested for differences using Fisher exact tests. Ordinal results (TIMI flow and blush grades) were presented as proportions of subjects per treatment group and tested for group differences using Cochran-Mantel-Haenszel mean score tests. Cochran-Mantel-Haenszel tests were also used when testing for group difference in final TIMI flow and blush grades, when results were stratified by baseline level. The significance level of all statistical tests was set at p = 0.05. Analyses were by intention-to-treat and were performed with SAS version 8.2 (SAS Institute, Inc., Cary, North Carolina). The sample size of 480 patients assumed 90% nuclear image ascertainment (<10% lost to follow-up) and detection of a 30% treatment effect with 80% power.
The study randomized 480 patients from 31 contributing sites in the U.S. and Canada from July 26, 2001 to January 21, 2004 (Appendix). Patients were treated with either RT as an adjunct to PCI (n = 240) or PCI alone (n = 240). Baseline demographic, clinical, and lesion characteristics were similar for both groups (Tables 1 and 2).⇓⇓As expected, use of temporary pacemakers was greater in the adjunct RT group, because prior experience has shown that bradycardia may be observed during activation of the RT catheter (4,7). Although crossover treatment was not permitted by protocol, 6 patients received crossover RT treatment. Moderate-to-large thrombus was present in a minority of patients at baseline: 21.3% in the adjunct RT group and 19.6% in the PCI alone group. Of note, baseline TIMI flow grade 3 was present in significantly more patients in the PCI alone group (Tables 3 to 5).⇓⇓
Procedure and angiographic results
The RT catheter was successfully delivered in 95% of cases. The unsuccessful deliveries were attributed to: delivery not attempted due to patient deterioration (n = 11), unable to cross the lesion with a guidewire (n = 2), unable to set up the RT system (n = 2), and attempted but unsuccessful delivery of the RT catheter (n = 4). Complete (88%) or partial (7%) thrombus removal was achieved, as assessed by the operator. The mean number of passes made with the RT catheter was 3 (range = 1 to 9), using a distal-to-proximal approach in 52% and a proximal-to-distal approach in 48%. Total procedure time was greater in the adjunct RT group (Table 1).
Quantitative angiographic analysis showed a higher incidence of final TIMI flow grade 3 in the PCI alone group, although there were significantly more patients with baseline TIMI flow grade 3 in this group, and the difference in final TIMI flow grade 3 was less pronounced when it was adjusted for baseline TIMI flow grade (Tables 3 to 5). There were no significant differences between groups in final TMP blush or corrected TIMI frame count. Angiographic complication rates were low and were similar in both groups (Tables 3 and 4).
Baseline and 90-min 12-lead ECGs that were deemed technically adequate for ST-segment analysis were available in 176 patients in the adjunct RT group and 164 patients in the PCI alone group. An ST-segment resolution >70% occurred in 105 patients (60%) in the adjunct RT group and in 111 patients (68%) treated with PCI alone (p = 0.14). An ST-segment resolution of 30% to 70% occurred in 50 patients (28%) treated with adjunct RT and 35 patients (21%) treated with PCI alone (p = 0.17), and ST-segment resolution <30% occurred in 21 patients (12%) treated with adjunct RT and 18 patients (11%) treated with PCI alone (p = 0.87).
The primary study end point, Tc 99m sestamibi infarct size, was significantly greater in the adjunct RT group (Table 6).By subgroup analysis, this difference was primarily driven by differences in inferior infarct size, and there were no significant differences in anterior infarct size between groups. Ejection fraction was similar in both groups.
Adverse events to 30 days
There were no intraprocedure deaths, strokes, or emergency CABG surgeries. Adverse events to 30 days are shown in Table 7.Major adverse cardiac events were significantly greater in the adjunct RT group, attributable to a higher mortality rate in this group. Among the patients who died, final TIMI flow grade 3 was achieved in all but 1 patient, who was assigned to adjunct RT treatment, but had unsuccessful reperfusion due to inability to cross the lesion with a guidewire. All TLR events in both groups were due to subacute stent thrombosis. Site-reported complications, including bleeding and arrhythmia events, occurred with similar frequency between groups.
Subset of patients with angiographically visible thrombus
While RT provided no benefit to reducing final infarct size in the general STEMI patient sample enrolled, there was interest in selective patient subsets, especially those with angiographically visible thrombus. A total of 96 patients (50 treated with adjunct RT, and 46 treated with PCI alone) had large or moderate baseline thrombus by angiographic analysis (excluding total occlusions). In this subset, the mean final infarct size was 10.8% in the adjunct RT group, compared with 8.1% in the PCI alone group (p = 0.23), but this subset also had a strong trend toward higher baseline TIMI flow grade in the PCI alone group (TIMI flow grade 0/1 = 36.0% in the adjunct RT group vs. 19.6% in the PCI alone group, and TIMI flow grade 2/3 = 64.0% in the adjunct RT group vs. 80.4% in the PCI alone group, p = 0.11). In this subset, there were 3 MACE events in the adjunct RT group (a death, stent thrombosis, and stroke) and no MACE in the PCI alone group (p = 0.24). While these observations may suggest safety of RT in patients with significant visible thrombus, this subset is too small for definitive analysis, and our study was not powered to detect small subgroup differences.
The results of this study demonstrate that, in patients with STEMI, the routine adjunctive use of RT did not reduce infarction size compared with primary PCI alone. Final TIMI flow grade 3 was achieved less often in patients undergoing adjunct RT. The 30-day MACE rate was higher in patients treated with adjunct RT, a difference primarily attributable to a very low mortality rate in patients treated with PCI alone. Other angiographic and electrocardiographic end points were similar in the 2 groups.
Sestamibi infarct size is associated with subsequent patient mortality, and myocardial infarct size has been used as a surrogate end point to evaluate reperfusion strategies in patients with STEMI (16,17). In our study, final infarction size was larger in patients treated with adjunct RT, a difference that was more pronounced for inferior MIs. Recent studies using distal protection devices in patients with STEMI have also failed to show any benefit (18,19). Whether the delivery of these relatively bulky devices predisposes to distal embolization requires further investigation. Time to reperfusion is also a powerful predictor of success in STEMI (20). The longer procedure time in the RT group may have negated any benefit of thrombus removal in these patients. There were 94 randomized patients (42 in RT group, 52 in PCI alone group) that had no measurable final infarction. Though the possibility of small subendocardial infarction in these patients cannot be excluded using the SPECT imaging alone, clinical significance of any difference between such small final infarction is yet to be established.
Final TIMI flow grade 3 is a predictor of outcomes in STEMI (21). Further, patients with established TIMI flow grade 2/3 coronary flow before PCI in acute MI have also been shown to have superior outcomes (22). In our study, baseline TIMI flow grade 3 was present in significantly more patients in the PCI alone group, and the extent to which this may have contributed to the observed differences in final outcomes is unknown. Yet, despite very high rates of final TIMI epicardial flow grade 3, TMP blush grade 3 was achieved in a minority of patients in both groups, reflecting the failure of current modalities in establishing the normal myocardial perfusion.
Benefits of RT in visibly thrombotic lesions in native coronary arteries and SVGs have been documented in prior studies (4,23,24). In these studies, glycoprotein IIb/IIIa inhibitors were typically used in only a small percentage of patients. Use of glycoprotein IIb/IIIa inhibitors has been shown to improve outcomes of primary PCI in STEMI (25,26), and their use in the majority of patients in the present study may have negated any benefit of direct thrombectomy in these patients.
Our study showed no significant difference in the percentage of patients who achieved complete ST-segment resolution at 90 min after reperfusion. In another study, adjunct treatment with RT resulted in early ST-segment elevation resolution in 90% versus 72% in the stenting only group, p = 0.02 (27). Another study using a different thrombectomy device showed that a higher degree of early ST-segment resolution was achieved with thrombectomy, but this difference did not persist after 18 h (28).
The overall 30-day MACE rate was significantly higher in the RT group, a difference driven primarily by increased mortality in the RT group (4.6% vs. 0.8%, p = 0.02). Our study was not powered to detect differences in clinical events, including mortality. Although the mortality in RT group is similar to mortality in other clinical trials with primary PCI, the mortality rate in the PCI alone group (0.8%) was lower than that seen in other contemporary acute MI reperfusion trials in which 30-day mortality ranged from 1.8% to 6.6% (29–31). A detailed adjudication of deaths did not yield any specific device-related concerns. The DSMB requested to review mortality again at 6 months, which showed that 3 additional deaths had occurred in each group—6-month mortality was 5.8% in the adjunct RT group and 2.1% n the PCI only group, p = 0.06. Although any mortality conclusions from a study not designed to detect mortality differences should be drawn with caution and cannot be considered definitive, increased MACE in the thrombectomy arm raises concerns regarding routine use of RT in STEMI.
Randomization after angiography may have introduced selection bias against enrolling high-risk patients with a large amount of angiographically apparent thrombus, though no data is available to confirm such bias in enrollment. Another limitation of the study is that MI size rather than myocardial salvage was measured. As a result, the effect of a lower baseline TIMI flow grade in patients treated with adjunct RT cannot be determined. It is also difficult to ascertain whether the higher baseline TIMI flow grades contributed to the very favorable outcomes in patients treated with PCI alone.
Despite high technical success rates, adjunct use of RT in patients with STEMI did not reduce infarct size, improve TIMI flow or blush grades, improve ST-segment resolution, and resulted in higher MACE at 30 days. The increased infarct size and increased mortality in the RT group raise safety concerns about the routine use of AngioJet in STEMI, and further studies are needed to clarify safety issues and the possible role of AJ in subgroups of patients with STEMI, particularly those with large thrombus burden. Reperfusion injury with interstitial edema, capillary leakage, and generalized coronary endothelial dysfunction may play a much larger role in poor myocardial reperfusion.
Eptifibatide for use in the study was provided by its manufacturer, Millennium Pharmaceuticals, Inc., Cambridge, Massachusetts. The authors also gratefully acknowledge the efforts of Maureen Lyden, MS, biostatistician (in collaboration with Clinical Development Associates, Inc., Richmond, Virginia) and Maria Pyle, AIMI Clinical Research Project Manager.
For a list of the AIMI investigative centers and principal investigators, please see the online version of this article.
Source of support: Possis Medical, Inc. is the sponsor of the study reported in this article. Dr. Ali received research funding and has acted as a speaker/consultant for Possis Medical Inc. David Holmes, Jr., MD, FACC, served as guest editor for this report.
- Abbreviations and Acronyms
- AngioJet Rheolytic Thrombectomy In Patients Undergoing Primary Angioplasty for Acute Myocardial Infarction trial
- coronary artery bypass graft
- Data and Safety Monitoring Board
- major adverse cardiac events
- myocardial infarction
- minimal lumen diameter
- percutaneous coronary intervention
- rheolytic thrombectomy
- single-photon emission computed tomography
- ST-segment elevation myocardial infarction
- saphenous vein graft
- Thrombolysis In Myocardial Infarction
- target lesion revascularization
- tissue myocardial perfusion
- Received July 13, 2005.
- Revision received February 28, 2006.
- Accepted March 2, 2006.
- American College of Cardiology Foundation
- Angeja B.G.,
- Gunda M.,
- Murphy S.A.,
- et al.
- ↵(2004) Instructions for Use: AngioJet XMI-OTW Rheolytic Thrombectomy Catheter (Possis Medical, Inc, Minneapolis, MN).
- Germano G.,
- Kiat H.,
- Kavanagh P.,
- et al.
- The TIMI Study Group
- Gibson C.M.,
- Cannon C.,
- Daley W.,
- et al.
- van’t Hof A.,
- Liem A.,
- Suryapranata H.,
- et al.
- Schroder R.,
- Dissmann R.,
- Bruggemann T.,
- et al.
- Gibbons R.J.,
- Miller T.D.,
- Christian T.F.
- Burns R.J.,
- Gibbons R.J.,
- Yi Q.,
- et al.
- Gick M.,
- Jander N.,
- Bestehorn H.P.,
- et al.
- Cannon C.
- Neumann F.J.,
- Blasini R.,
- Schmitt C.,
- et al.
- Kastrati A.,
- Mehilli J.,
- Dirschinger J.,
- et al.,
- Stent versus Thrombolysis for Patients with Occluded Coronary Arteries in Patients with Acute Myocardial Infarction (STOPAMI-2) study investigators
- Beran G.,
- Lang I.,
- Schreiber W.,
- et al.