Author + information
- Received December 18, 2007
- Revision received March 3, 2008
- Accepted March 15, 2008
- Published online June 1, 2008.
- Dean J. Kereiakes, MD, FACC⁎,⁎ (, )
- Mark A. Turco, MD, FACC†,
- Jeffrey Breall, MD, PhD‡,
- Naim Z. Farhat, MD§,
- Robert L. Feldman, MD¶,
- Brent McLaurin, MD∥,
- Jeffrey J. Popma, MD, FACC⁎⁎,
- Laura Mauri, MD, MSc⁎⁎,††,
- Peter Zimetbaum, MD††,‡‡,
- Joseph Massaro, PhD††,
- Donald E. Cutlip, MD††,‡‡,
- AMEthyst Study Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Dean J. Kereiakes, The Lindner Center, 2123 Auburn Avenue, Suite 424, Cincinnati, Ohio 45219.
Objectives We sought to evaluate the relative safety and efficacy of the novel Interceptor PLUS Coronary Filter System (Medtronic Vascular, Santa Rosa, California) compared with approved embolic-protection devices (e.g., GuardWire, Medtronic Vascular/FilterWire EZ, Boston Scientific, Natick, Massachusetts) during percutaneous coronary intervention (PCI) of degenerative saphenous vein grafts (SVG).
Background Percutaneous coronary intervention of degenerative SVG is associated with embolization of atherothrombotic debris and subsequent myocardial infarction in a significant portion of patients. The use of distal embolic-protection devices has previously been demonstrated to reduce major adverse cardiovascular events associated with PCI in these patients.
Methods In this multicenter, randomized noninferiority trial, 797 patients undergoing PCI with stenting of SVG stenoses (de novo or restenotic) with reference vessel diameter 2.5 mm to 5.25 mm were randomly assigned 2:1 to either the Interceptor PLUS (n = 533) or control distal-protection devices (GuardWire [n = 191], FilterWire EZ [n = 73]) at the physician's discretion.
Results The trial primary clinical end point (composite occurrence of death, myocardial infarction, or urgent repeat revascularization through 30 days) was observed in 8% and 7.3% of Interceptor and control-treated patients, respectively (p = 0.025 for noninferiority; p = 0.77 for difference). Key secondary end points for device and procedural success were similar between randomly assigned treatment strategies.
Conclusions The Interceptor PLUS Coronary Filter System is noninferior in safety and efficacy to 30 days when compared with the GuardWire and FilterWire EZ distal embolic protection devices.
Autologous saphenous veins remain a common source of conduit for use in surgical coronary bypass graft revascularization (CABG) procedures. Within 10 years of CABG, half of venous grafts may become severely diseased or occluded (1–3). Repeat CABG is associated with greater procedural risk, less durable graft patency, and less symptomatic improvement when compared with the initial surgical procedure (4,5). Saphenous vein graft (SVG) atherosclerotic plaque is frequently complex, friable and may be associated with thrombus. Percutaneous coronary intervention (PCI) with either balloon angioplasty or stent deployment may be associated with a high (47% to 54%) incidence of enzymatic myocardial infarction (MI), including large (>5 times upper limit of normal for creatine kinase [CK]-myocardial band [MB]) infarctions in 16% to 18% of procedures (6,7). Both periprocedural MI and reduced antegrade graft flow (“no reflow”) have been ascribed to embolization of atherothrombotic debris and distal microvascular occlusion (1,4,8–12). By using a distal balloon occlusion-aspiration device (GuardWire, Medtronic Vascular, Santa Rosa, California) to trap and retrieve particulate debris, Webb et al. (10) identified particulate matter in 21 of 23 SVG angioplasty procedures. In the pivotal multicenter SAFER (Saphenous vein graft Angioplasty Free of Emboli Randomized) trial of 801 patients undergoing percutaneous SVG interventions, deployment of the GuardWire device was associated with a 42% relative reduction in major adverse cardiac events (MACEs) at 30 days after stenting when compared with conventional (no embolic protection) PCI (9.6% GuardWire vs. 16.5% control; p = 0.004) (13). The potential advantage of catheter-based filter devices as an alternative approach to capture and retrieve atherothrombotic debris is the maintenance of coronary blood flow with enhanced clinical stability and patient tolerance. In a multicenter randomized trial involving 651 patients undergoing percutaneous intervention of 682 SVG lesions, a catheter-based filter device (FilterWire EZ, Boston Scientific, Natick, Massachusetts) demonstrated similar efficacy to the GuardWire device with respect to the occurrence of MACE to 30 days (9.9% vs 11.6%, respectively; p = 0.0008 for noninferiority) (14). However, filter devices may differ in their crossing profile, ability to circumferentially appose the target conduit, efficiency of particulate capture, and ability to maintain conduit flow which is, in large part, determined by the open pore area of the deployed device (15). The purpose of the AMEthyst (Assessment of the Medtronic AVE Interceptor Saphenous Vein Graft Filter System) trial was to determine the relative safety and efficacy of a novel catheter-based filter device (the Interceptor PLUS Coronary Filter System; Medtronic Vascular) when compared with Food and Drug Administration (FDA)-approved, commercially available distal protection devices (the GuardWire and the FilterWire EZ devices) when used as an adjunct to PCI with or without stenting, of degenerative saphenous vein bypass grafts.
Between July 2003 and April 2007, 797 patients were enrolled at 73 U.S. sites. Eligible patients were at least 18 years of age with symptomatic ischemic heart disease due to a target lesion with >50% (but <100%) diameter stenosis of at least 1 SVG with a reference vessel diameter of ≥2.5 and ≤5.25 mm, and a Thrombolysis In Myocardial Infarction (TIMI) flow grade ≥1. Percutaneous coronary intervention could be performed on as many as 2 SVGs per patient, with no limit to the number of target lesions (de novo or first-time restenotic with no previous stent deployed) attempted. Target lesion treatment was limited to balloon angioplasty and stent deployment. Major exclusion criteria included allergy or contraindication to study medications, stent materials, or contrast; acute MI, or recent (>24 h) MI with total CK >2 times normal within the past 24 h and cardiac enzymes (CK-MB fraction) greater than normal limits at the time of treatment; major surgery within 30 days; left ventricular ejection fraction <25%; serum creatinine >2.5 mg/dl (if not on chronic hemodialysis); stroke or transient ischemic attack within 2 months; active peptic ulcer or upper gastrointestinal bleeding within 3 months; or history of bleeding diathesis or coagulopathy and concurrent medical condition with life expectancy <12 months.
In addition to the target SVG-PCI, PCI of as many as 2 nontarget native coronary lesions was permitted during the index procedure. Nontarget lesions were required to have stenosis severity >50, but <100% with lesion length <24 mm and reference vessel diameter >2.5 mm. Percutaneous coronary intervention of nontarget lesions with the following characteristics was not permitted: location distal to either the target SVG or other bypass graft, location distal to a >45° bend in the vessel, moderate-to-severe calcification; ostial or bifurcation lesions; an unprotected left main lesion; or evidence of thrombus.
Eligible patients agreed to all postprocedure follow-up and signed a written informed consent. The study protocol was approved by the institutional review boards of all participating centers. An independent data and safety monitoring board performed a prespecified interim review of the first 300 randomized patients and again after completion of the 30-day follow-up.
The Interceptor PLUS Coronary Filter System consists of a steerable vascular filter, an actuator handle, a torque handle, a peel-away introducer, and a guidewire introducer (Fig. 1). The vascular filter component comprises the self-expanding and free-floating filter, constructed of nitinol and platinum-filled (for enhancement of radiopacity) nitinol wire and constrained between 2 mechanical stops on a steerable 0.014-inch diameter guidewire of 180 cm in length. The filter is deployed and collapsed by means of the actuator handle. After the filter is expanded, blood and embolic material pass through 4 large in-flow openings (>1,400 μm) located at the proximal end of the device. The distal filter has numerous 100-μm pores that provide perfusion while trapping embolic debris.
The FilterWire EZ is also a catheter-based distal filtration device that incorporates a nonocclusive filter (pore size 110 μm) in the shape of a windsock, mounted on a self-expanding nitinol loop, and fixed on its own guidewire. The FilterWire EZ is deployed through a 3.2-F delivery sheath and recaptured with a 4-F retrieval sheath (14,16). The difference in open pore area between the Interceptor PLUS and FilterWire EZ is illustrated in Figure 2 and has been reported to be associated with a 60% difference in flow (in vitro) (15).
The GuardWire temporary occlusion-aspiration system consists of a guidewire incorporating a central inflation lumen, to which an elastomeric balloon is attached. The device has a 2.8-F crossing profile, and contrast inflation of the balloon arrests anterograde graft flow (13). The intervention is then performed over the wire, liberated debris is trapped proximal to the balloon and is aspirated through a 5-F monorail export catheter before balloon deflation, which restores graft flow (13).
Randomization and interventional protocol
Successful treatment of eligible nontarget lesions was required before study enrollment. Use of a platelet glycoprotein (GP) IIb/IIIa receptor inhibitor was at the discretion of the treating physician but was determined before randomization, which was stratified by the intent for GP IIb/IIIa receptor inhibitor use in an attempt to ensure equal distribution of adjunctive therapies in the 2 arms of the trial. Eligible patients meeting these conditions were enrolled and randomized (in a 2:1 ratio) to receive either the Interceptor PLUS or 1 of the 2 control devices (GuardWire or FilterWire EZ, at the investigator's discretion). During the trial, GuardWire was approved for PCI of SVG with reference vessel diameter 2.5 to 6.0 mm and FilterWire EZ was approved for SVG with reference vessel diameter 3.5 to 5.5 mm. In patients with target SVG reference diameter <3.5 mm randomized to the “control” device, the investigator was instructed by protocol to select GuardWire to ensure proper device sizing.
All patients received a loading dose of 325 mg of nonenteric-coated water-soluble aspirin, and either 300 mg of clopidogrel or 500 mg of ticlopidine, unless the patient was on a chronic thienopyridine therapy. A complete blood cell count, blood urea nitrogen, and creatinine, in addition to a pregnancy test (for women of childbearing age) were obtained within 7 days before the procedure, and CK enzyme, CK-MB isoenzyme levels, and electrocardiogram (ECG) were obtained within 24 h before the procedure. After the initial intravenous bolus of unfractionated heparin (dose per participating hospital standard), a baseline activated clotting time was obtained and monitored at 30-min intervals throughout the procedure with additional heparin administered as needed to maintain the activated clotting time ≥250 s for patients receiving concomitant GP IIb/IIIa blockade and ≥300 s for those who did not.
Baseline and post-PCI angiography of the target vessel was performed after intracoronary graft nitroglycerine administration in at least 2 near-orthogonal views, which showed the target lesion free of foreshortening or vessel overlap. The assigned embolic protection device was deployed according to the manufacturer's instructions for use, and the target lesion was then treated with balloon angioplasty and/or stenting. The choice of drug-eluting (DES) or bare-metal stent (BMS) was at the discretion of the treating physician.
After the procedure, all patients were continued on a daily maintenance dose of aspirin (81 to 325 mg) and patients who received stents were prescribed clopidogrel (75 mg every day) or ticlopidine (250 mg twice a day) according to the stent type (30 days for BMS and 3 to 6 months' minimum for DES). An ECG, complete blood cell count, blood urea nitrogen, and creatinine were performed 16 to 24 h after the procedure or before discharge. In addition, CK and CK-MB were measured at 6 to 8 h, 12 to 16 h, and 20 to 24 h after procedure. Clinical follow-up with ECG was obtained at 30 ± 5 days, with repeat blood testing and angiography repeated if clinically indicated.
Procedural angiograms were forwarded to an angiographic core laboratory for independent review. Lesion length was defined as the axial extent of the lesion that contained a shoulder-to-shoulder lumen reduction by >20%. Thrombus was defined as none, haziness, small size comprising one-half of the lumen diameter, moderate sized comprising more than one-half but <2 lumen diameters, and large comprising more than 2 lumen diameters. Coronary aneurysms were defined as a maximum lumen diameter within the treatment zone that was 1.2 times larger than the average reference diameter of the vessel.
The degree of SVG degeneration was assessed using an ordinal metric of the extent of lumen irregularities and ectasia (>20% of the reference normal segment) within the SVG that comprises <25% (SVG degeneration score 0), 26% to 50% (SVG degeneration score 1), 51% to 75% (SVG degeneration score 2), or >75% (SVG degeneration score 3) of the total SVG length. Reference vessel diameter and minimal lumen diameter were determined from 2 projections using an automated edge-detection algorithm (CMS Medis, Leiden, the Netherlands) (17). Electrocardiograms were interpreted by an independent core laboratory at the Harvard Clinical Research Institute. The clinical events committee reviewed and adjudicated all clinical events, and the data coordination also was performed by an independent core laboratory at the Harvard Clinical Research Institute.
Study end points and statistical methods
The intention-to-treat (ITT) sample consisted of all randomized patients except for 1 patient, who consented after receiving protocol-specified medications (excluded from ITT per directive of site institutional review board) and 2 patients erroneously randomized (not included per directive of the FDA [ITT n = 797]). A per-protocol (PP) analysis included all randomized patients treated with the assigned device (n = 719). The study primary end point was the composite end point of MACE at 30 days, (defined as the occurrence of death, MI, or repeat revascularization [CABG or PCI] of the target vessel). Myocardial infarction was defined as CK-MB elevation >3 times normal. Secondary end points included in-hospital MACE; device success, defined as successful delivery deployment and retrieval of the study device to the target site; clinical success, defined as device success with no in-hospital MACE; and final TIMI flow grade.
The null hypothesis for this study was that the Interceptor PLUS Coronary Filter System would have a primary end point event rate greater than or equal to that of the control device (GuardWire or FilterWire EZ) plus a noninferiority margin delta. The alternative hypothesis was that the Interceptor PLUS would have a primary end point event rate less than that of the control device plus a noninferiority margin delta.
The initial study sample size calculation was made using the following assumptions: the 30-day MACE rate for Interceptor and control groups would be 10% based on observations made in the SAFER and FIRE trials (13,14); the power of the study would be ≥80%; the 1-sided alpha error would be 5% and the noninferiority margin delta would be 5.5% (as in the precedent FIRE trial) (14). On the basis of these assumptions, enrollment of 600 patients in a 2:1 ratio of the Interceptor PLUS (400 patients) to control (200 patients) was planned with enrichment of the control population by borrowing data from the GUARDWIRE arm of the SAFER trial with the use of Bayesian methodology.
At the prescribed review by the Data and Safety Monitoring Board of the first 300 enrolled patients, the observed 30-day MACE rate was substantially <10%, raising concerns about both the adequacy of the proposed sample size as well as the validity of borrowing control data from a study (SAFER) with a potentially different (i.e., greater-risk) population. In conjunction with the FDA, the following revised assumptions were made: 30-day MACE rate for the control and Interceptor device groups would be 6%; the power of the study would ≥80%; and the noninferiority margin delta was set at 4.5%. In this context, by the use of a 2:1 randomization and the Farmington-Manning approach (18) to assess noninferiority, the planned study enrollment was increased to total 800 patients.
Categorical variables were tested by the use of appropriate contingency table analyses (exact or chi-square approximations), and continuous variables were tested with the Student t test or Wilcoxon rank sum test, depending on variable distributions. All statistical analyses were performed with the use of SAS for Windows (versions 8 and 9; SAS Institute, Cary, North Carolina) and NCSS 2004–PASS 2005 software (NCSS, Kaysville, Utah). The 95% confidence intervals (CIs) for the treatment difference on the primary end point were 1-sided and used a significance level of 0.05. Confidence intervals for the treatment difference on all other end points were 2-sided and used a significance level of 0.05.
An analysis was conducted to compare the consistency of treatment effect in patients with and without GP IIb/IIIa inhibitor treatment. Logistic regression was used to assess the significance of embolic protection treatment group by GP IIb/IIIa inhibitor treatment interaction on 30-day MACE rate with the use of a 0.10 level of significance.
After enrollment and randomization of the first 54 patients using the Interceptor P2 filter system, the next-generation (Interceptor PLUS) device became available and was deployed for the remaining 479 patients randomized to the treatment group. Iterative improvements provided by the Interceptor PLUS included lower crossing profile (2.7 vs. 2.9 F), shorter distal landing zone requirement (7.5 vs. 4.5 cm), no delivery or retrieval sheath requirement, and rapid exchange monorail catheter compatibility. The comparability of the Interceptor P2 and PLUS subgroups was confirmed by analysis of baseline characteristics and 30-day MACE rates in the ITT populations, and these groups were pooled for the final analysis, which compared all Interceptor device(s) to the control device.
A total of 800 patients were enrolled and randomized to either the Interceptor arm (n = 533) or control (n = 264) treatment arm. In the control group, 191 patients received GuardWire and 73 received FilterWire EZ device treatment. No differences in baseline clinical demographic parameters were observed between the Interceptor and control groups. A high prevalence of diabetes, hypertension, and dyslipidemia was observed in both treatment groups, and ∼10% of both groups had previous balloon angioplasty to the target lesion (Table 1). Baseline angiographic characteristics were similar between randomized treatment groups with the exceptions that SVG age was slightly older for Interceptor-treated patients and baseline TIMI flow grades were slightly less in control-treated patients (Table 2). The extent of saphenous vein graft degeneration was also similar between groups. Approximately 20% of patients in both treatment groups had extensive (>50% graft length) degeneration.
Of the 797 enrolled patients, 4 patients did not have a procedure performed and 8 did not have the assigned treatment strategy attempted. No differences in any procedural parameters were observed between randomly assigned treatment groups (Table 3). Approximately 40% of patients in both groups received periprocedural GP IIb/IIIa inhibitor therapy, and a similar portion of each group received treatment with DES (vs. BMS). Device success in the Interceptor group (90.5%; 478 of 528) was similar to that in controls (93.8%; 241 of 257; p = 0.134). Clinical success was similar in the Interceptor (85.4%, 451 of 528) and control groups (86.8%; 230 of 257; p = 0.118). Postprocedure in-stent minimal lumen diameter and percent diameter stenosis were slightly larger in the Interceptor treatment group. The final TIMI flow grade was 3 in 98.1% (528 of 538) of Interceptor and 97.4% (267 of 271) of control patients. Similarly, no differences in the incidence of no-reflow, abrupt vessel closure, or distal embolization were observed between treatment groups.
In-hospital and 30-day clinical outcomes
No differences between treatment groups were observed with respect to in-hospital outcomes (Table 4). Similarly, to 30 days' follow-up, no differences were observed in any adverse clinical events with the exception that target vessel revascularization (TVR) of a nontarget lesion was slightly more prevalent (p = 0.04) in control device-treated patients compared with the Interceptor group. These revascularization events involved the distal SVG site of device deployment and occurred within 12 h and 26 days after GuardWire (n = 2) and at 17 days after FilterWire EZ (n = 1) treatment, respectively. Target vessel revascularization of a nontarget lesion was not observed after Interceptor treatment. Angiographic core lab determined target vessel reference diameters were for patients who incurred this adverse event were 2.10 and 1.61 mm in the GuardWire and 3.45 mm in the FilterWire-treated patient, respectively. Clinical events associated with TVR of a nontarget lesion included non–Q-wave MI (n = 1) and unstable angina (n = 2).
The primary end point of 30-day MACE (Table 4) analyzed on patients with available follow-up (≥25 days after procedure or incidence of MACE by 30 days) occurred at a rate of 8.0% (40 of 501) in the Interceptor group and 7.3% (18 of 247) in the control group, yielding a risk difference of 0.7% (upper 1-sided 95% CI of 4.0% by the Farrington-Manning approach). This upper CI was less than the noninferiority margin of 4.5% and indicates that the Interceptor device met noninferiority criteria when compared with the control (p = 0.025). This analysis was repeated on the entire ITT population (n = 533 Interceptor; n = 264 control) with multiple imputation used to estimate 30-day MACE incidence for patients who were prematurely withdrawn and yielded similar results (upper 1-sided 95% CI of 4.3%). Similarly, analysis of the PP population which included 719 patients (735 lesions), of whom 478 were successfully treated with the Interceptor and 241 were successfully treated with control (172 GuardWire, 69 FilterWire EZ), demonstrated MACE to 30 days in 6.2% of Interceptor and 6.3% of control patients, yielding an upper-bounded 95% CI of 3.0% (well less than the noninferiority margin of 4.5%; noninferiority p = 0.023. There was no significant interaction between treatment assignment and prerandomization intended use of GP IIb/IIIa inhibitors on 30-day MACE (p = 0.569).
Percutaneous coronary intervention of degenerative SVGs is complicated by adverse ischemic events, particularly myocardial infarction, because of distal embolization of atherothrombotic debris and microvascular occlusion (1,4,8–12). The risk of atheroembolism may be increased in proportion to the age and extent of SVG degeneration, plaque burden (determined by conduit diameter and lesion length), as well as the presence of thrombus (19,20). The use of an embolic protection device during SVG interventions is associated with a significant reduction in adverse clinical events (vs. no embolic protection) and, hence, has been accorded a Class 1 recommendation in the current American College of Cardiology/American Heart Association guidelines for the performance of PCI (11–15,21).
Several strategies for embolic protection exist and include distal conduit occlusion and aspiration as well as distal filter devices (12,22). Limited available data suggest that both the efficiency and size of particulate debris capture is similar between occlusion and filter devices (23). Distal occlusion and aspiration may offer the potential advantage of removing humoral factors (both lesion and platelet derived) that promote microvascular spasm (24). Alternatively, filter devices that maintain conduit flow may be symptomatically and hemodynamically better tolerated by the patient and provide angiographic visibility of the target lesion(s) during PCI. Filter devices may differ in crossing profile and ability to atraumatically circumferentially oppose the target conduit as well as in open pore area available to provide flow and capture debris (15). Indeed, open-pore area (that portion of the expanded filter which is composed of pores) appears to be a critical determinant of both flow and capacity to capture debris without becoming obstructed. Despite the availability of randomized controlled trial data demonstrating efficacy for embolic protection in reducing periprocedural MACEs and the commercial availability of multiple embolic-protection devices, an embolic-protection strategy is used in only a minority (≤30%) of patients undergoing SVG-PCI (25). This apparent lack of use remains unexplained but may, at least in part, be due to the relative complexity of operator use for available devices. The purpose of the present study was to evaluate the relative safety and efficacy of a novel guidewire based filter distal embolic protection device, when compared with devices which were commercially available at the time of study enrollment. From this study, several important observations can be made.
First, the Interceptor PLUS filter system was demonstrated to be not inferior to the control embolic protection device(s) with respect to the occurrence of MACE to 30 days in both the primary (ITT) analysis as well as the secondary (PP-treated) subgroup analyses. No difference in the occurrence of any adverse clinical event was observed between randomly assigned treatment groups (Interceptor vs. control) with the exception that TVR of a nontarget lesion was observed slightly more frequently among control device treated patients (1.2% vs. 0% Interceptor respectively; p = 0.04). In addition, no differences were observed in either device or clinical procedural success. Furthermore, there was consistency of the treatment effect of the Interceptor device compared with control devices across GP IIb/IIIa-treated and -untreated patients. This observation is consistent with both the lack of apparent GP IIb/IIIa receptor inhibitor benefit for either GuardWire or FilterWire EZ treated patients in both the SAFER and FIRE trials (13,14), as well as the overall lack of GP IIb/IIIa benefit during SVG PCI previously reported from multiple randomized, placebo controlled trials (19,26,27).
Although the ability of GP IIb/IIIa receptor blocking agents to disaggregate platelet-rich thrombus (28) has intuitive theoretic appeal to prevent filter pore plugging and enhance filter efficiency, no objective clinical or angiographic data support this premise. The lack of GP IIb/IIIa inhibitor benefit has been attributed to the possibility that platelet inhibition may be overwhelmed when large amounts of prothrombotic embolic material are present or conversely, that periprocedural ischemic events during SVG-PCI may not be thrombus related but instead due to the friable particulate debris characteristic of older SVG atherosclerotic lesions (19,22,27).
Given the friability of aging SVGs, there is a potential concern for placement of any distal protection device through and beyond the target lesion. In this regard, the trend toward less frequent TVR (of nontarget lesions) to 30 days among Interceptor compared with control device treated patients is noteworthy. Although the numbers of events are small (3 in control vs. 0 in Interceptor) and could be due to chance, both the nature and time-course of these events are clinically relevant. In each case, angiographic evidence of endoluminal disruption was evident in the region of distal graft where device deployment occurred. As device deployment occurs distal (25 to 35 mm) to target lesion, the small target vessel reference diameters of 2.10 and 1.61 mm in the GuardWire-treated patients are noteworthy and suggest caution for device deployment in smaller caliber conduits. Endoluminal trauma due either to device expansion or traction on the device during stent deployment or device retrieval may have occurred.
It is of some concern that the overall incidence of MACE to 30 days in distal embolic protection device-treated patients in this trial appears to be less than in previous trials in which the control devices (GuardWire, FilterWire EZ) were evaluated (13,14). Indeed, GuardWire-treated patients enrolled into the SAFER trial experienced a 9.6% incidence of MACE to 30 days using a similar definition as was employed in the current study (13). Similarly, MACE to 30 days was observed in 11.6% of GuardWire and 9.9% of FilterWire-treated patients in the FIRE trial (14). These precedent observations prompted the 10% MACE rate assumption used in the original AMEthyst study sample size calculation. After enrollment and prespecified interim analysis of the first 300 patients in the AMEthyst trial, a significantly lower primary end point event rate (∼6%) was observed and prompted expansion of trial enrollment from 600 to 800 patients. This observation has important implications for the design of future SVG intervention trials. Potential explanations for this finding include differences in the SVG disease being treated or in the type and prevalence of adjunctive pharmacotherapy as well as in the type of stent devices deployed. Preliminary comparisons suggest no significant differences exist between GuardWire-treated patients enrolled into the SAFER trial and the AMEthyst trial with respect to reference target vessel diameter, lesion length, or plaque burden. Although the prevalence of platelet GP IIb/IIIa inhibitor use was more frequent in the SAFER (∼60%) compared with the AMEthyst (∼40%) trial, fewer patients in the SAFER trial had adequate pretreatment with either statins or clopidogrel, which have been demonstrated to reduce periprocedural ischemic events, particularly enzymatic infarction (29,30).
Finally, iterative evolution in available BMS and DES platforms has occurred, which could influence the occurrence of early (≤30 day) events. The AMEthyst is the first large-scale randomized trial of embolic protection in which patients could receive any FDA-approved stent whether DES or BMS. Of note, DES were deployed in ∼60% of patients enrolled into AMEthyst. Stent thrombosis through 30 days was observed in 0.4% of DES- and 0.7% of BMS-treated patients, respectively. The variety and distribution of stent type deployed in AMEthyst differs from those used in both the SAFER and FIRE trials and may, in part, contribute to the apparent differences in clinical end points observed.
Several limitations of the current study deserve mention. First, iterative device evolution resulted in design improvements being incorporated into the study device early in the trial. The design improvements made in the Interceptor PLUS (compared with the Interceptor P2 device) were in large part related to operator ease of use (no delivery or retrieval sheath), simplified device deployment with the actuator handle and monorail catheter compatibility. There were no differences between these devices in filter pore size or open-pore area. Furthermore, the validity of pooling Interceptor P2 (n = 54) and Interceptor PLUS (n = 479) treated patients is supported by the absence of significant difference in any clinical outcomes between these treatment groups.
Second, the original study sample size calculation used an assumption of primary end point event occurrence rate (∼10%) based on available precedent randomized controlled trials. The remarkable lower incidence of event rate occurrence observed after the enrollment of the first 300 patients required trial enrollment expansion (from 600 to 800 patients) to provide adequate statistical power and a more stringent noninferiority delta than used in the FIRE trial (4.5% vs. 5.5% respectively). Several potential explanations for the lower-than-expected event rate exist, of which the prevalence of background adjunctive pharmacologic pretreatment (particularly with clopidogrel and statin therapy) appears most plausible. Importantly, the appropriate adjustment in sample size was made without study unblinding to provide adequate statistical power to the primary end point observations.
Third, assignment to the control device (GuardWire vs. FilterWire EZ) was not randomized, and pooling of these devices was performed to provide an aggregate control device event rate. Pooling of control devices is supported by the lack of difference in any clinical endpoint through 30 days between these nonrandomly assigned groups. In this respect, the introduction, relative ease of use, and rapid clinical adoption of the FilterWire EZ embolic protection device after conception of the AMEthyst trial severely limited trial enrollment and mandated inclusion of this device to maintain clinical relevance of the primary end point comparisons.
Finally, as more devices have become FDA approved and commercially available, the selection of a single device to serve as an active control in a randomized trial comparison has become progressively difficult. Indeed, it could be argued that the comparison of a novel device with the available devices allows a more clinically relevant comparison to current standards of care.
This study demonstrates that safety and efficacy of the novel Interceptor PLUS Coronary Filter System for distal embolic protection during PCI of degenerative SVG is similar to that of commercially available devices (GuardWire; FilterWire EZ) to 30 days after procedure. A trend toward less-frequent TVR of nontarget lesions to 30 days was observed in Interceptor PLUS-treated patients.
The following individuals were involved in the management of the study and clinical data: Emily Gmitter, Principal Clinical Research Specialist at Medtronic Vascular; Christopher Weeks, Clinical Research Associate Medtronic Vascular; and Manuela Negoita, Clinical Director—Data Management and Biostatistics.
For a complete list of investigators, please see the online version of this article.
This study was supported by a research grant from Medtronic Vascular, Santa Rosa, California. A complete list of investigators is listed in the Appendix. Dr. Kereiakes received modest research grants from Conor Medsystems, Pfizer, Cordis/Johnson & Johnson, Boston Scientific, Medtronic, and Daiichi Sanyko, and modest consulting fees from ConorMedsystems, Cordis/Johnson & Johnson, Core Valve, Eli Lilly & Co., Boston Scientific, and Abbott/ Bioadsorbable Vascular Solutions. Dr. Turco received research support from and is a consultant for Medtronic Vascular. Dr. Popma received research grants from Cordis, Boston Scientific, Medtronic, Abbott, Ev3, and Biosensors, and is on the Speakers' Bureau of Pfizer, Medicines Company, Bristol-Meyer Squibb, and Sanofi. Dr. Mauri received a hononarium (U.S. $<10,000) from Medtronic. Dr. Massaro has served as a consultant for the Harvard Clinical Research Institute.
- Abbreviations and Acronyms
- bare-metal stent(s)
- coronary bypass graft surgery
- creatine kinase
- drug-eluting stent(s)
- Food and Drug Administration
- intention to treat
- major adverse cardiovascular events
- myocardial infarction
- percutaneous coronary intervention
- per protocol
- saphenous vein graft
- Thrombolysis In Myocardial Infarction
- target vessel revascularization
- Received December 18, 2007.
- Revision received March 3, 2008.
- Accepted March 15, 2008.
- American College of Cardiology Foundation
- de Feyter P.J.
- Fitzgibbon G.M.,
- Kafka H.P.,
- Leach A.J.,
- Keon W.J.,
- Hooper G.D.,
- Burton J.R.
- Motwani J.G.,
- Topol E.J.
- Keeley E.C.,
- Velez C.A.,
- O'Neill W.W.,
- Safian R.D.
- Weintraub W.S.,
- Jones E.L.,
- Morris D.C.,
- King S.B. III.,
- Guyton R.A.,
- Craver J.M.
- Hong M.K.,
- Mehran R.,
- Dangas G.,
- et al.
- Hong M.K.,
- Mehran R.,
- Dangas G.,
- et al.
- Topol E.J.,
- Yadav J.S.
- Piana R.N.,
- Paik G.Y.,
- Moscucci M.,
- et al.
- Webb J.G.,
- Carere R.G.,
- Virmani R.,
- et al.
- Grube E.,
- Gerckens U.,
- Yeung A.C.,
- et al.
- Gorog D.A.,
- Foale R.A.,
- Malik I.
- Baim D.S.,
- Wahr D.,
- George B.,
- et al.
- Stone G.W.,
- Rogers C.,
- Hermiller J.,
- et al.
- Stone G.W.,
- Rogers C.,
- Ramee S.,
- et al.
- van der Zwet P.M.,
- Reiber J.H.
- Coolong A.,
- Baim D.S.,
- Kuntz R.E.,
- et al.
- Smith S.C. Jr..,
- Feldman T.E.,
- Hirshfeld J.W.,
- et al.
- Kereiakes D.J.
- Rogers C.,
- Huynh R.,
- Seifert P.A.,
- et al.
- Salloum J.,
- Reddy B.,
- Vaughan D.E.,
- Zaho D.X.
- Ellis S.G.,
- Lincoff A.M.,
- Miller D.,
- et al.
- Roffi M.,
- Mukherjee D.,
- Chew D.P.,
- et al.
- Goto S.,
- Tamura N.,
- Ishida H.
- Pasceri V.,
- Patti G.,
- Nusca A.,
- et al.
- Patti G.,
- Pasceri V.,
- Colonna G.,
- et al.