Author + information
- Received May 8, 2007
- Revision received June 15, 2007
- Accepted June 25, 2007
- Published online October 9, 2007.
- Laura Mauri, MD, MSc, FACC⁎,†,⁎ (, )
- David Cox, MD, FACC‡,
- James Hermiller, MD, FACC§,
- Joseph Massaro, PhD†,
- Joyce Wahr, MD∥,
- Sew Wah Tay, PhD∥,
- Michael Jonas, MD⁎,
- Jeffrey J. Popma, MD, FACC⁎,
- Jim Pavliska, BS∥,
- Dennis Wahr, MD, FACC∥ and
- Campbell Rogers, MD, FACC⁎
- ↵⁎Reprint requests and correspondence:
Dr. Laura Mauri, Brigham and Women’s Hospital, 75 Francis Street, Boston, Massachusetts 02115.
Objectives To determine if outcomes could be further improved, we investigated an embolic protection device placed proximal to the target lesion that could provide protection before lesion instrumentation, allow the use of conventional guidewires, and permit embolic protection in anatomy unfavorable for distal devices.
Background Embolic complications during stenting of degenerated saphenous vein coronary bypass grafts are reduced, but not eliminated, by distal protection.
Methods A total of 594 patients undergoing stenting of 639 saphenous vein grafts were prospectively randomized, using a noninferiority design, to compare 2 treatment strategies: control (distal protection whenever possible) or test (proximal protection when possible, distal when not).
Results The primary composite end point of death, myocardial infarction, or target vessel revascularization at 30 days by intention to treat analysis occurred in 10.0% of control and 9.2% of test patients; difference = −0.8% (95% confidence interval [CI] −5.5% to 4.0%); p for noninferiority = 0.0061. In device specific analysis, this composite end point occurred in 11.7% of distal protection patients and 7.1% of proximal protection patients (difference = −4.6% [95% CI −9.6% to 0.3%]; p for superiority = 0.10, p for noninferiority = 0.001). Finally, in the subset of patients with lesions amenable to treatment with either proximal or distal protection devices (n = 410), the primary composite end point occurred in 12.2% of distal protection patients and 7.4% of proximal protection patients; difference = −4.7% (95% CI −10.4% to 1.0%), p for superiority = 0.14, p for noninferiority = 0.001.
Conclusions Using proximal embolic protection whenever possible during treatment of diseased saphenous vein grafts produced outcomes similar to those with distal embolic protection.
Distal embolic protection has become standard therapy during percutaneous stenting of diseased saphenous vein (SVG) aortocoronary bypass grafts, reducing rates of myocardial infarction and major adverse cardiac events (MACE) by roughly 50% (1). Both distal occlusive and distal filtering approaches have proven similar in efficacy, reducing events related to debris dislodged during stenting (2). However, despite distal embolic protection, MACE accompanies SVG stenting far more frequently (approximately 10% of patients) than native coronary artery (non-SVG) stenting. Inherent limitations of distal protection include initial passage of a device across the stenosis without protection in place, distal occlusion devices requiring several minutes of ischemia, and distal filters capable of trapping small particles (3) but tending occasionally to become occluded with debris (4) and lacking ability to prevent transit of soluble mediators into the myocardium. Furthermore, many SVG lesions lie distally, precluding distal protection device use (5), or have “Y” or sequential limbs that preclude the simultaneous protection of multiple downstream myocardial beds.
The Proxis Embolic Protection System (St. Jude Medical, Maple Grove, Minnesota) is a unique single-operator catheter that is deployed proximal to the target lesion before crossing (Fig. 1).Inflation of the sealing balloon interrupts antegrade flow during the period of lesion intervention. Stagnated blood and emboli liberated during intervention is then retrieved by gentle aspiration or via ancillary flushing of the vessel (6).
To evaluate whether use of proximal protection could circumvent the limitations of distal embolic protection, we performed a pivotal, multicenter prospective randomized trial. Inherent differences between proximal and distal protection systems obviated use of a traditional study design. Therefore, the PROXIMAL (Proximal Protection During Saphenous Vein Graft Intervention) trial compared 2 treatment strategies in a noninferiority format. Patients with SVG stenoses were randomized to 1 of 2 treatment strategies: a current-care control arm (distal embolic protection device whenever possible, and no embolic protection when not) or a test arm (proximal protection device whenever possible and distal embolic protection when anatomy precluded proximal protection).
Patients >18 years old undergoing percutaneous coronary intervention (PCI) with planned stenting of lesion(s) in up to 2 diseased SVGs were eligible for enrollment. Lesions were ≥50% and <100% stenosed with at least Thrombolysis In Myocardial Infarction flow grade 1 and reference vessel size 3.0 to 5.0 mm at the site of deployment of the protection device. Exclusion criteria included acute or recent myocardial infarction or current elevation of creatine phosphokinase-MB isoenzyme; uncontrollable allergy or contraindication to contrast or any of the study medications; previous stroke or transient ischemic attack within 2 months; serum creatinine >2.5 mg/dl; previous cardiac surgery within 30 days; SVG age <2 months; lesions within 5 mm of the proximal anastomotic site; planned use of devices other than balloon angioplasty and stenting; and left ventricular ejection fraction <20%. The protocol was approved by the investigational review board at each site, and all patients provided informed written consent.
Randomization and interventional protocol
Aspirin 100 to 325 mg was administered within 24 h before the procedure. A loading dose of 300 mg of clopidogrel or 500 mg of ticlopidine was recommended before the procedure and required within 2 h after the procedure. Glycoprotein IIb/IIIa inhibitor use was left to the discretion of the investigator. Intravenous heparin was administered to maintain the procedural activated clotting time ≥250 s (≥200 s if a glycoprotein IIb/IIIa inhibitor was used).
Eligible patients were randomized without regard to lesion location in a 1:1 ratio, stratified by study center and intended glycoprotein IIb/IIIa use. Patients were assigned to one of two treatment strategies. In the control arm, current care strategy was used (treatment with distal protection whenever possible as defined by the current instructions for use for the chosen device) with no protection when lesions were not amenable to distal protection. Either the FilterWire EX, the FilterWire EZ system (Boston Scientific, Natick, Massachusetts), or the GuardWire Plus (Medtronic, Santa Rosa, California) could be used. In the test arm, a proximal protection strategy was used (treatment with proximal protection [Proxis, St. Jude Medical] for all lesions >15 mm distal to the proximal anastamosis, and distal protection when lesions were within 15 mm of the proximal anastamosis).
Once embolic protection was established, stenting was performed using standard techniques. The decision whether to perform balloon dilation before and/or after stent implantation was left to the discretion of the operator. The levels of creatine phosphokinase-MB were measured at 3 times within the first 24 h after the procedure, and a 12-lead electrocardiogram was obtained immediately after procedure and before discharge. Aspirin and a thienopyridine were continued at least through the 30-day follow-up period.
Quantitative coronary angiography
Standard image acquisition was performed at the clinical sites using 2 or more angiographic projections of the stenosis, and the angiograms were forwarded to the Brigham and Women’s Hospital Angiographic Core Laboratory for independent review. All procedural angiograms were reviewed using standard morphologic criteria (7). Lesion length was defined as the axial extent of the lesion that contained a shoulder-to-shoulder lumen reduction of >20% or more. The SVG degeneration score was a continuous measure of the extent of lumen irregularities and ectasia within the SVG as a percentage of the total SVG length. Using the contrast filled injection catheter as the calibration source, we performed quantitative angiographic analysis with a validated automated edge detection algorithm (Medis CMS, Leiden, the Netherlands) (8). Selected images for analysis were identified with 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 implantation. Minimal lumen diameters were measured at baseline, after stent implantation within the stent (in-stent analysis), within the 5 mm proximal and distal edges of the stent, and within the segment between the proximal and distal reference vessel (in-lesion analysis). Plaque volume was estimated from reference vessel diameter, lesion length and minimal lumen diameter, and SVG degeneration score is an ordinal metric of the extent of lumen irregularities and ectasia within the SVG that comprises <25%, 25% to 50%, 51% to 75%, or >75% of the total SVG length, as previously described (9).
Definitions, end points, and statistical analysis
The primary end point of the study was the 30-day composite occurrence of MACE, including death, Q-wave or non–Q-wave myocardial infarction (MI), emergent coronary artery bypass graft surgery, or target vessel revascularization. An MI was defined as any postprocedure creatine phosphokinase-MB isoenzyme ≥3 times normal. The development of new, pathological Q waves in ≥2 contiguous leads (assessed by a blinded electrocardiography core laboratory technician) was required to diagnose a Q-wave MI. Neither patients nor investigators were blinded to treatment group, but primary end point events were adjudicated by an independent committee blinded to treatment allocation. Device success was defined as successful delivery to the target site, deployment, and use and subsequent retrieval of the intact device when all device criteria were met. Clinical success was defined as device success without in-hospital MACE.
The study was designed to test whether a treatment strategy in which proximal protection was used whenever possible would be noninferior to a strategy of distal protection only. Rejection of the null hypothesis would indicate that the proximal protection strategy has a primary end point rate (30-day MACE) not exceeding that of the distal protection strategy plus a prespecified noninferiority margin (δ). Assuming 30-day MACE rates for both strategies of 12.7%, randomizing 600 patients would provide 80% power to demonstrate noninferiority with 1-sided alpha of 5% and a noninferiority absolute margin of 7%. Categorical variables were tested using appropriate contingency table analyses (exact or chi-square approximations), and continuous variables were tested using unpaired Student ttest or a nonparametric test. Categorical variables are presented as percentages, whereas continuous variables are presented as mean ± SD.
The primary analysis sample was defined as all randomized patients based on the principle of intention to treat. Noninferiority was assessed using a 1-sided 0.05 level of significance. Unless specified otherwise, statistical significance was declared for all remaining analyses at the 2-sided 0.05 level. The influence of baseline demographic features, angiographic variables, and treatment assignment on 30-day MACE was evaluated using multiple stepwise logistic regression analyses, with an entry criterion of 0.20 and a stay criterion of 0.10.
The randomization scheme used in this study and the diverse nature of SVG lesion locations meant that some patients randomized to test received embolic protection with a distal device; and some patients randomized to control were treated with no embolic protection. Therefore, in addition to the intent-to-treat analysis, a prespecified analysis was also performed according to treatment received (i.e., by device: protection with proximal vs. distal devices vs. no embolic protection). Finally, an analysis was performed on the subset of patients whose lesions were suitable for treatment with either proximal or distal protection (i.e., lesions >15 mm from the proximal anastamosis and>20 mm from the distal anastamosis.)
Between July 2003 and April 2005, 600 patients at 68 sites in the U.S., Canada, and Europe undergoing stent placement in 655 SVG lesions were randomized (300 control [i.e. distal embolic protection whenever possible] and 300 test [i.e. proximal embolic protection whenever possible, distal when not]). Figure 2demonstrates how patients were randomized in this trial and what treatment they actually received. Six patients randomized to the test arm were treated with a 7-F rather than an 8-F guide-compatible Proxis catheter. Because of its late introduction and its limited use, the interchangeability of the 7-F system for pooled analysis could not be tested with sufficient power, and these 6 patients (contributing 7 lesions) were excluded yielding a final sample size of 294 patients. Inclusion of the 7-F patients, however, was not found to alter the primary end point results.
As reported in other SVG embolic protection and stenting trials, the population had a high incidence of multiple cardiovascular risk factors and previous MI and was elderly (Table 1).Baseline angiographic characteristics (Table 2)were well matched between the 2 arms, except that slightly more patients randomized to test had a target lesion in an SVG to the left anterior descending artery, whereas those randomized to control were slightly more likely to have a target lesion in the SVG supplying the left circumflex artery. Plaque volume was similar between arms.
Procedural results of intention to treat
Device and clinical success rates were similar with both devices (Table 3);glycoprotein IIb/IIIa receptor antagonists were used in approximately 40% of patients in both groups. Quantitative and qualitative angiographic results and epicardial coronary flow rates were similar in both arms, except for a slightly lower residual in-lesion diameter stenosis in the control group. In the control group, 74% of lesions were treated with distal protection (81% with FilterWire, 19% with GuardWire) and 26% were treated without embolic protection. In the test group, 82% were treated with proximal protection, 16% with distal protection, and 2% without embolic protection. The median occlusion time for each Proxis inflation was 3.23 ± 2.71 min, and the mean number of inflations per lesion was 1.38.
30-day clinical outcomes, intention-to-treat analysis
The primary end point of 30-day MACE occurred in 10.0% of control patients compared with 9.2% of test patients (difference = −0.8% [95% confidence interval (CI) −5.5% to 4.0%], p = 0.0061 for noninferiority) (Figs. 2 and 3).⇓Most MACE events were non–Q-wave MI (Table 4).The extent of periprocedural myonecrosis was essentially identical in both the control and test groups (Fig. 4),and adverse event rates stratified by patient characteristics were similar in the 2 arms.
Outcomes according to device used
Of the 594 enrolled patients, 282 patients were treated with distal protection (66 patients received GuardWire, 216 received FilterWire), 242 received proximal protection, and 70 were treated without protection. The primary end point of 30-day MACE occurred in 11.7% of distal protection device-treated patients and 7.1% of Proxis catheter-treated patients; difference = −4.6% (95% CI −9.6% to 0.3%); p for superiority = 0.10, p for noninferiority = 0.001 (Fig. 3). Of patients whose lesions were treated without embolic protection (7 test and 63 control patients), 10% (7 of 70) experienced 30-day MACE. As noted in Figure 5,the extent of periprocedural myonecrosis appeared somewhat greater in patients treated with distal protection, although the difference did not reach statistical significance (p = 0.12).
Outcomes in patients amenable to treatment with either device
In the 410 patients with at least 1 lesion amenable to treatment with either device (i.e., patients whose lesions were not located within 15 mm of the proximal or within 20 mm of the distal anastamosis), 30-day MACE occurred in 12.2% (27 of 222) of distal protection patients and 7.4% (14 of 188) of the proximal protection patients; difference = −4.7% (95% CI −10.4% to 1.0%), p for superiority = 0.14, p for noninferiority = 0.001 (Fig. 3). In this subgroup of patients the extent of periprocedural myonecrosis was numerically but not statistically greater in patients treated with distal protection (p = 0.09) (Fig. 6).
Univariate and multivariate predictors of adverse events
Multivariate predictors of MACE are presented for the intention-to-treat analysis in Table 5and for the by device analysis in Table 6.In the by device analysis, but not the intention-to-treat analysis, treatment with Proxis was an independent predictor of decreased MACE with an odds ratio of 0.4 for MACE (p = 0.0064).
At the initiation of the trial, to provide operators with early training using the investigational proximal embolic protection, 117 patients who met the same inclusion and exclusion criteria as the randomized cohort were enrolled in a nonrandomized roll-in phase and were assigned to treatment with the Proxis System. These patients were analyzed separately from the primary noninferiority analysis. The 30-day MACE rate in these patients was 7.8%.
Stenting of SVGs carries a high risk of periprocedural complications, primarily myonecrosis (10). Furthermore, a link between these complications and late events is strongly suspected (10,11). Distal embolic protection devices have enhanced the safety of SVG stenting, reducing MACE rates in a series of trials by approximately 50% (1,2,12,13). For this reason, distal embolic protection devices are standard of care for SVG stenting procedures (14). However, a finite MACE rate of roughly 10% has persisted despite distal embolic protection use. Putative culprits for such adverse events include emboli released during initial lesion crossing (before embolic protection is in place), emboli arising from injury at the site of protection device deployment, incomplete capture of debris or soluble mediators, occlusive overloading of filter-based devices with debris and platelet/fibrin, or emboli shed after the completion of the procedure (when embolic protection is no longer in place). Furthermore, as many as 40% of patients requiring SVG stenting are not suited to distal protection devices by virtue of lesion location (5). For these reasons, novel approaches to enhance the safety of SVG stenting and extend embolic protection to a greater proportion of patients have been devised. The PROXIMAL study is the first pivotal, prospective, randomized trial to test whether a strategy of preferential use of proximal, rather than distal, embolic protection can afford protection to a greater proportion of patients and provide as good, or better, outcomes.
The PROXIMAL trial compared 2 treatment strategies in a noninferiority format. Patients with SVG stenoses were randomized either to a current care control arm (distal embolic protection device whenever possible, and no embolic protection when anatomy precluded distal protection device use) or to a test arm (proximal protection device whenever possible and distal embolic protection when anatomy precluded proximal protection). The primary intention-to-treat analysis showed that these 2 strategies resulted in similar outcomes after 30 days (10.0% vs. 9.2% for control and test arms, respectively), with MACE rates in both arms close to those reported in other recent embolic protection trials (1,2,12,13). Consistent with this observation, a previous comparison of the particulate burden retrieved during SVG PCI showed no significant difference between proximal occlusion, distal occlusion, and distal filtering (15).
Some patients were unable to be treated with any embolic protection device. In this group, the 30-day MACE rate was not greater than either device-treated group. However, caution should be used in this comparison because the number of events is small (7) and subject to sampling error and should not be taken as evidence of equivalent outcomes with no protection in this small and selected group.
Secondary analyses provided some evidence that proximal protection may provide additional benefit to distal protection in selected populations. Support for this hypothesis is found in the fact that, for the subset of patients whose lesions fell within an SVG segment amenable to treatment with either type of protection device, proximal protection was associated with numerically although not statistically lower 30-day MACE and less periprocedural myonecrosis. In addition, the by device analysis found that proximal protection was an independent predictor of lower 30-day MACE (Table 6). The lower risk of MACE with proximal protection in these secondary analyses might be explained by avoidance of embolization of debris liberated during initial lesion crossing or filter occlusion with debris.
There are inherent differences in lesions eligible for proximal versus distal protection devices wherein some lesions are amenable only to proximal protection, some only to distal protection, and others (the majority) amenable to both. Therefore, although the lesion types treated in this study with distal versus proximal protection are overlapping, they are not identical. The PROXIMAL study therefore was designed to compare treatment strategies in a broad patient population rather than a direct comparison of devices in only those lesions amenable to treatment with either device, a smaller cohort of all lesions.
This study provides evidence that a treatment strategy of proximal protection results in outcomes equivalent to those seen with a strategy of distal protection, and that, in lesions amenable to treatment with either device, proximal protection results in outcomes at least as good as distal protection. Demonstration of noninferiority of a treatment strategy using proximal embolic protection allows the application of effective embolic protection to a broader population, including those in whom distal embolic protection is not possible because of disease in the distal vein or not desirable because of difficulty or embolic risk of crossing the obstructive stenosis with the protection device prior to dilation.
Use of the Proxis proximal protection system whenever possible during SVG PCI results in similar MACE rates 30 days after SVG PCI as the use of distal embolic protection whenever possible. At the very least, this supports use of proximal protection for lesions not amenable to treatment with a distal protection device. Secondary analyses suggest that incremental benefit may be afforded by proximal protection perhaps via protection from debris liberated by lesion crossing. These observations provide an additional approach to improving outcomes in patients undergoing SVG PCI.
For a list of study sites, investigators, and organizations, please see the online version of this article.
The PROXIMAL Trial: Proximal Protection During Saphenous Vein Graft Intervention Using the Proxis Embolic Protection System: A Randomized, Prospective, Multicenter Clinical Trial
Proximal Protection during Saphenous Vein Graft Intervention using the Proxis™ Embolic Protection System The PROXIMAL Trial
A complete list of the PROXIMAL investigators and study organization appears in the Appendix. This trial was funded by St. Jude Medical, formerly known as Velocimed, Inc., Maple Grove, Minnesota. Drs. Dennis Wahr and Tay and Jim Pavliska were employees of Velocimed/St. Jude Medical at the time of the study. Dr. Dennis Wahr is the co-founder of Velocimed, Inc. See accompanying online Cardiosource Slide Set.
- Abbreviations and Acronyms
- major adverse cardiac event
- myocardial infarction
- percutaneous coronary intervention
- saphenous vein graft
- Received May 8, 2007.
- Revision received June 15, 2007.
- Accepted June 25, 2007.
- American College of Cardiology Foundation
- Baim D.S.,
- Wahr D.,
- George B.,
- et al.
- Stone G.W.,
- Rogers C.,
- Hermiller J.,
- et al.
- Rogers C.,
- Huynh R.,
- Seifert P.,
- et al.
- Stone G.W.,
- Rogers C.,
- Ramee S.,
- et al.
- Lansky A.,
- Popma J.
- van der Zwet P.,
- Reiber J.
- Hong M.K.,
- Mehran R.,
- Dangas G.,
- et al.
- Carrozza J.J.,
- Mumma M.,
- Breall J.,
- Fernandez A.,
- Heyman E.,
- Metzger C.
- Mauri L.,
- Rogers C.,
- Baim D.S.