Direct Stenting With the TAXUS Liberté Drug-Eluting Stent

Transcription

1 JACC: CARDIOVASCULAR INTERVENTIONS VOL. 1, NO. 2, BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION ISSN /08/$34.00 PUBLISHED BY ELSEVIER INC. DOI: /j.jcin ing With the TAXUS Liberté Drug-Eluting Stent Results From the TAXUS ATLAS DIRECT STENT Study John A. Ormiston, MB, CHB,* Ehtisham Mahmud, MD, Mark A. Turco, MD, FACC, FSCAI, Jeffrey J. Popma, MD, Neil Weissman, MD, Louis A. Cannon, MD, Tift Mann, MD,# Michael J. Lucca, MD, FACC, FSCAI,** Soo-Teik Lim, MBBS, Jack J. Hall, MD, Dougal McClean, MD, David Dobies, MD, FACC, FSCAI, Lazar Mandinov, MD, Donald S. Baim, MD, FACC Auckland and Christchurch, New Zealand; San Diego, California; Takoma Park, Maryland; Boston and Natick, Massachusetts; Washington, DC; Petoskey and Grand Blanc, Michigan; Raleigh, North Carolina; Duluth, Minnesota; Singapore; and Indianapolis, Indiana Objectives This study was conducted to determine whether direct stenting with TAXUS Liberté is noninferior to stenting after pre-dilation. Background Direct stenting is performed in approximately 30% of patients, but data on clinical and angiographic outcomes with drug-eluting stents are limited. Methods The TAXUS ATLAS DIRECT STENT is a single-arm, multicenter study that enrolled patients with de novo coronary lesions visually estimated to be 10 to 28 mm in length in vessels 2.5 to 4.0 mm in diameter. The control group is the quantitative coronary angiography (QCA) subset of the TAXUS ATLAS trial, which used identical inclusion and exclusion criteria but mandated pre-dilation. The primary end point is 9-month analysis-segment percent diameter stenosis (%DS). Results Baseline patient characteristics were similar between the groups. On QCA analysis, significantly shorter lesions with larger lumen diameter and less calcification were observed in the direct stent group. Direct stenting was successful in 97.6% of patients and was associated with a shorter procedure time and fewer complications. Follow-up %DS was noninferior for direct stent (26.41%) versus pre-dilation (29.14%) with a 1-sided 95% confidence interval of the difference between the groups ( 0.34%) well below the pre-specified noninferiority margin (6.75%). Additionally, significantly lower restenosis (5.9% vs. 11.4%, p ) and target lesion revascularization (TLR) 2.9% vs. 7.8%, p ) rates were seen for direct stent versus pre-dilation. Conclusions Direct stenting of TAXUS Liberté is feasible and highly successful in carefully selected lesions. Direct stenting is noninferior to stenting after pre-dilation on the basis of %DS and can significantly reduce procedural time, procedural complications, and possibly angiographic restenosis and TLR. (TAXUS Liberté-SR Paclitaxel-Eluting Coronary Stent System; NCT ). (J Am Coll Cardiol Intv 2008;1:150 60) 2008 by the American College of Cardiology Foundation From the *Mercy Angiography Unit, Mercy Hospital, Auckland, New Zealand; University of California, San Diego, San Diego, California; Center for Cardiac & Vascular Research, Washington Adventist Hospital, Takoma Park, Maryland; Angiographic Core Laboratory, Brigham and Women s Hospital, Boston, Massachusetts; Cardiovascular Research Institute, Washington Hospital Center, Washington, DC; Cardiac & Vascular Research Center, Northern Michigan Hospital, Petoskey, Michigan; #Wake Heart Associates, Wake Medical Center, Raleigh, North Carolina; **St. Mary s Duluth Clinic Regional Heart Center, Duluth, Minnesota; National Heart Centre, Singapore; The Heart Center, St. Vincent s Hospital, Indianapolis, Indiana; Christchurch Hospital, Christchurch, New Zealand; Genesys Regional Medical Center, Grand Blanc, Michigan; and Boston Scientific Corporation, Natick, Massachusetts. This work was supported by Boston Scientific Corporation. Drs. Ormiston and Turco have received consulting fees/honoraria from Boston Scientific Corporation. Drs. Turco and Cannon are on the Speakers Bureau of the Boston Scientific Corporation. Drs. Mahmud, Turco, Popma, and Weissman have received research

2 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, Conventional stenting generally employs balloon predilation of the target lesion to simplify stent delivery and facilitate complete stent expansion. The introduction of more flexible, lower-profile, balloon-expandable stents has made direct stenting (i.e., without pre-dilation) more common. Studies using bare metal stents have proven that direct stenting is safe and effective in patients selected for favorable anatomy, with outcomes similar to stenting after predilation but with reduced procedural time and cost (1 5). Drug-eluting stents have demonstrated a clear benefit over bare metal stents in suppressing neointimal growth and reducing the rate of repeat revascularization, resulting in better long-term outcomes even in complex lesions (6 8). Similar to the bare-metal stent experience, drug-eluting stents are implanted without pre-dilation in up to 34% of patients (9 11). There are few data from controlled trials on direct stenting with drug-eluting stents, beyond early evidence suggesting it to be as safe as stenting after pre-dilation (12 14). The TAXUS Liberté stent (Boston Scientific Corporation, Natick, Massachusetts) combines the polymer-based, paclitaxel-eluting TAXUS technology with a more advanced stent geometry and was proven to be effective in the pivotal TAXUS ATLAS study (15). The TAXUS Liberté features thinner 316L stainless steel struts ( vs inches); a continuous cell architecture for improved deliverability, conformability, and homogeneity of drug distribution; and a smaller tip profile to allow easier access to the target lesion (16). The TAXUS ATLAS DIRECT STENT was a prospective, single-arm, historically controlled study evaluating the hypothesis that direct stenting with TAXUS Liberté is noninferior to stenting with balloon pre-dilation. Methods Device description. The TAXUS Liberté-SR Stent consists of a balloon-expandable Liberté stent with a polymer coating containing 1 g/mm 2 of paclitaxel in a slow-release formulation. Patient selection, procedure, and follow-up. A total of 247 direct stent patients were enrolled at 24 centers in the U.S., New Zealand, and Singapore from March to September The study protocol was approved by local ethics review committees, and all patients provided written informed consent. Trial results are reported on the National Institutes of Health website (see registry information after the abstract). Eligible patients were 18 years or older undergoing percutaneous coronary intervention of a single de novo lesion 10 and 28 mm in length (visual estimate) in a native coronary artery with a reference vessel diameter (RVD) of 2.5 and 4.0 mm (visual estimate) and a left ventricular ejection fraction of 25%. Patients with myocardial infarction within 72 h were prohibited. Angiographic exclusion criteria prohibited left main, ostial, or bifurcation lesions; total occlusion or thrombus; severe or moderate lesion calcification; and tortuosity or excessive bending. Enrollment occurred after successful treatment of a non-target lesion (if applicable) but before advancing a guidewire across the target lesion. Loading doses of clopidogrel (300 mg) or ticlopidine (500 mg), and aspirin (300 mg) were administered before the procedure. After stent implantation, all patients were prescribed clopidogrel (75 mg daily) or ticlopidine (250 mg twice daily) for a minimum of 6 months. Aspirin ( 100 mg daily) was mandated for at least 9 months and recommended indefinitely. Clinical follow-up was scheduled at 1, 4, and 9 months and yearly thereafter for 5 years. Quantitative coronary angiography (QCA) and intravascular ultrasound (IVUS) follow-up for all patients were scheduled at 9 months. Control group. The 543 patients from the QCA subset of the pivotal TAXUS ATLAS study served as the pre-dilation control for TAXUS ATLAS DIRECT STENT (15). Inclusion and exclusion criteria were identical for TAXUS ATLAS and TAXUS ATLAS DIRECT STENT. Enrollment, however, occurred after successful pre-dilation of Abbreviations and Acronyms %DS percent diameter stenosis ARC Academic Research Consortium CI confidence interval HR hazard ratio ITT intention-to-treat IVUS intravascular ultrasound MACE major adverse coronary events MLD minimum lumen diameter PP per-protocol QCA quantitative coronary angiography RVD reference vessel diameter TLR target lesion revascularization TVR target vessel revascularization the target lesion in TAXUS ATLAS but before wiring of the target lesion in TAXUS ATLAS DIRECT STENT. Additionally, whereas all direct stent patients were required to have both IVUS and QCA follow-up at 9 months, only a subset of pre-dilation control subjects underwent IVUS follow-up (Fig. 1). Definitions. Major adverse cardiac events (MACE) and stent thromboses were adjudicated by an independent Clinical Events committee. The clinical and angiographic end grants from Boston Scientific Corporation. Dr. Ormiston is on the advisory board of Boston Scientific Corporation. Dr. Baim is a stockholder of Boston Scientific Corporation. Drs. Mandiov and Baim are full-time employees of Boston Scientific Corporation. Manuscript received November 1, 2007; revised manuscript received January 18, 2008, accepted January 20, 2008.

3 152 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, 2008 Figure 1. TAXUS ATLAS DIRECT STENT Study Flow *Only 60% of pre-dilation control patients versus 100% of direct stent patients were included in the intravascular ultrasound (IVUS) subset; starting at 12 months, only per-protocol population (PP) patients will continue to be followed. Angio angiography; F/U follow-up; ITT intent-to-treat. point definitions were identical to those in TAXUS ATLAS (15). Stent thrombosis was determined angiographically or, in the absence of angiographic confirmation, as sudden cardiac death within 30 days of implantation or myocardial infarction in the territory of the target vessel, with a protocol definition that closely parallels the Academic Research Consortium (ARC) definite or probable criteria. The QCA and IVUS analyses were performed by the same core laboratories (Brigham and Women s Hospital, Boston, Massachusetts, and Washington Hospital Center, Washington, DC, respectively) for both TAXUS ATLAS DIRECT STENT and TAXUS ATLAS (15). Both QCA and IVUS measures within the stent and the analysis segment (including the stented region and the 5-mm edge regions) were reported separately. Paired analyses were conducted on patients for whom both procedure and 9-month data were available. End points. The primary end point was percent diameter stenosis (%DS) of the analysis segment at 9 months. Secondary end points included MACE, stent thrombosis, all-cause death, and QCA and IVUS outcomes. Statistical methodology. All end points were analyzed with both the intention-to-treat (ITT) and the per-protocol (PP) population, which excluded patients who did not receive either the study stent or the assigned implantation procedure. The PP population was the main population of interest for noninferiority testing of the primary end point to eliminate possible dilution of the true differences between the assigned stent and implantation procedures, thus preventing erroneous conclusions. The pre-specified noninferiority margin of 6.75% was based on one-half of the effect exhibited by the TAXUS Express stent versus the bare metal control in the TAXUS IV trial (8). Assuming a noninferiority margin of 6.75%, a 9-month %DS analysis segment rate of 26.3% (on the basis of TAXUS IV), and 30% attrition, the required enrollment for TAXUS ATLAS DIRECT STENT was 250 subjects for 89% power. Baseline, post-procedure, and follow-up information were summarized with descriptive statistics. The p value for comparing 2 independent continuous variables was from Student t test and for comparing 2 proportions was from the chi-square test; Fisher exact test was used if the total number of samples was 40 or at least 1 cell count in the 2 2 table was expected to be 5. The Kaplan-Meier product-limit method and log-rank test were used to assess time-to-event end points across study groups. Propensity score analysis was performed with a hierarchical logistic regression model with a stepwise selection process. The significance level for entry and exit of independent variables was set at Multivariate analysis was used to determine predictors for target lesion revascularization (TLR) at 9 months. All covariates were modeled univariately for each outcome and multivariately with a stepwise procedure in an appropriate regression model. Statistical significance was set at p Results Baseline characteristics. The groups were well-matched for demographic data and clinical characteristics, but direct stent patients had shorter lesions, larger minimum lumen diameter (MLD), less %DS, and less calcification (Table 1). There was a high prevalence of type B2/C lesions in both

4 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, Table 1. Baseline Patient and Lesion Characteristics (n 247) p Value* Male 68.1% (370/543) 70.4% (174/247) 0.52 Age (yrs) (543) (247) 0.61 Cardiac history Stable angina 58.7% (319/543) 57.9% (143/247) 0.82 Unstable angina 33.3% (181/543) 35.6% (88/247) 0.53 Silent ischemia 7.9% (43/543) 6.5% (16/247) 0.47 Previous MI 27.4% (149/543) 30.8% (76/247) 0.34 Previous PCI 28.4% (154/543) 29.6% (73/247) 0.73 Cardiac risk factors Current smoking 25.2% (137/543) 25.9% (64/247) 0.84 Diabetes, medically treated 26.0% (141/543) 22.7% (56/247) 0.32 Hypertension 72.9% (396/543) 66.0% (163/247) Hyperlipidemia 75.9% (412/543) 75.3% (186/247) 0.86 Lesion characteristics (by QCA) Target lesion vessel 0.17 LAD 41.6% (226/543) 46.6% (115/247) 0.19 LCX 25.4% (138/543) 27.1% (67/247) 0.61 RCA 33.0% (179/543) 26.3% (65/247) 0.06 Lesion location 0.90 Ostial 2.8% (15/543) 3.2% (8/247) 0.71 Proximal 39.2% (213/543) 41.3% (102/247) 0.58 Mid 51.0% (277/543) 49.0% (121/247) 0.60 Distal 7.0% (38/543) 6.5% (16/247) 0.79 RVD (mm) (543) (247) 0.25 MLD (mm) (543) (247) %DS (543) (247) Lesion length (mm) (543) (247) Bend % (151/543) 27.1% (67/247) 0.84 Tortuosity 13.1% (71/543) 10.1% (25/247) 0.24 Calcification 31.7% (172/543) 23.5% (58/247) Moderate 26.3% (143/543) 21.5% (53/247) 0.15 Severe 5.3% (29/543) 2.0% (5/247) Branch vessel disease 11.6% (63/541) 15.2% (37/244) 0.17 Modified ACC/AHA lesion type 0.14 A 5.3% (29/543) 4.9% (12/247) 0.78 B1 16.6% (90/543) 22.3% (55/247) B2 44.4% (241/543) 45.7% (113/247) 0.72 C 33.7% (183/543) 27.1% (67/247) 0.06 B2 or C 78.1% (424/543) 72.9% (180/247) 0.11 Numbers are % (count/sample size) or mean SD (n). *p values were calculated by Student t test or chi-square test unless otherwise noted. p values calculated by Fisher exact test. The B2 or C category was not included in the calculation of the overall p value for the American College of Cardiology (ACC)/American Heart Association (AHA) lesion type. CABG coronary artery bypass graft; CAD coronary artery disease; DS diameter stenosis; LAD left anterior descending artery; LCX left circumflex (artery); MI myocardial infarction; MLD minimum lumen diameter; PCI percutaneous coronary intervention; QCA quantitative coronary angiography; RCA right coronary artery; RVD reference vessel diameter. groups. For both groups, 50% of target lesions were 13 mm in length, but 95% of lesions were 23.5 mm (maximum 31.1 mm) for the direct stent group, versus 95% of lesions being 28 mm for the control (maximum 52.7 mm), indicating a skew toward shorter lesions for the direct stent group (Fig. 2). Procedural outcomes. The TAXUS Liberté stent was successfully implanted by direct stenting in 97.6% of direct stent patients, with crossover to balloon pre-dilation required in only 6 patients (2.4%; Table 2). One control patient (0.2%) underwent direct stenting. All 247 direct stent patients (100%) and 539 (99.3%) control patients successfully received the TAXUS Liberté stent. A mean of 1.1 stents/patient was implanted in both groups (591 stents in 543 control patients; 269 stents in 247 direct stent patients). The majority of implanted stents were 3.0 mm in

5 154 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, 2008 Figure 2. Cumulative Frequency Distribution of Baseline Lesion Length as Determined by QCA Median and 95th percentile lengths are shown. QCA quantitative coronary angiography. diameter (32.0% control; 45.4% direct stent) and 16 mm in length (37.4% control; 50.9% direct stent). There were no episodes of stent detachment or embolization in the direct stent group. Procedural time was similar between the groups (Table 2). However, 100% of direct stent patients but only 60.2% of control patients were included in the IVUS follow-up. For those patients in the IVUS subset, both procedure and fluoroscopy time were significantly shorter for direct stent versus control by nearly 5 and 2 min, respectively. Maximum stent-deployment inflation pressure was similar between the groups. Post-dilation was performed 13.0% more frequently for the direct stent group (Table 2). The maximum balloon/artery size ratio and the overall maximum device size were both significantly smaller for direct stent than control. There were significantly fewer complications in the direct stent group than the control (Table 2). Table 2. Procedural Characteristics and Success (n 247) p Value* Procedure time (min) (542) (246) 0.35 Procedure time IVUS subset (min) (326) (246) Fluoroscopy time IVUS subset (min) (327) (246) Pre-dilation 99.8% (542/543) 2.4% (6/247) Maximum pressure overall (atm) (543) (247) 0.28 Maximum balloon/artery ratio (543) (247) Maximum device size (mm) (543) (247) Post-dilation 49.7% (270/543) 62.8% (155/247) Multiple stents implanted 9.0% (49/543) 8.1% (20/247) 0.67 Technical and procedural performance Successful direct stenting n/a 97.6% (241/247) n/a Clinical procedural success 96.7% (525/543) 96.8% (239/247) 0.96 Procedural complications 3.1% (17/541) 0.4% (1/247) Abrupt closure 0.9% (5/540) 0.4% (1/246) 0.67 No reflow 0.6% (3/540) 0.0% (0/247) 0.56 Dissection 2.4% (13/540) 0.0% (0/247) Others 1.1% (6/541) 0.0% (0/247) 0.18 Numbers shown are mean SD (n) or % (count/sample size). *p values were calculated by Student t test or chi-square test unless otherwise noted. All direct stent patients were to receive IVUS, whereas only a subset of pre-dilation control patients were assigned to the IVUS subset. Maximum balloon size investigator-reported; RVD from core lab QCA. Defined as mean lesion diameter stenosis of 30% with TIMI flow grade 3, as visually assessed by the Investigator, without the occurrence of MACE as of the time of hospital discharge. Defined as the occurrence of no reflow, abrupt closure, dissection, embolus, perforation, or post-procedure thrombus. p values calculated by Fisher exact test. IVUS intravascular ultrasound; MACE major adverse cardiac events; QCA quantitative coronary angiography; RVD reference vessel diameter.

6 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, Antiplatelet use. At hospital discharge, 99.6% of predilation control subjects and 100% of direct stent patients were receiving dual antiplatelet therapy (p 1.00). A total of 91.6% of pre-dilation control patients and 93.3% of direct stent patients (p 0.40) received dual antiplatelet therapy through the protocol-mandated 6-month duration. By 9 months, the use of dual antiplatelet therapy had fallen to 60.5% for pre-dilation and 57.9% for direct stent (p 0.49); however, 94.7% of control patients and 97.9% of direct stent patients (p ) were taking aspirin. 9-month angiographic diameter stenosis. The primary end point of noninferiority of analysis segment %DS was analyzed for both the ITT and PP populations (Fig. 3). The PP population exhibited unadjusted %DS values of 29.14% for the control and 26.41% for the direct stent group. The difference in %DS values between the groups was 2.73%, and the upper 1-sided 95% confidence interval (CI) of this difference was 0.34%, which is significantly less than the pre-specified noninferiority margin of 6.75% (p ); thus, noninferiority of direct stenting versus pre-dilation was demonstrated. Similar results were obtained with the ITT population and after propensity score adjustment for differences in baseline characteristics. Angiographic and IVUS paired lesion analysis. At 9 months, paired lesion QCA analysis demonstrated similar MLD, %DS, and late loss in both groups (Table 3). The IVUS analysis revealed no differences between the groups for changes in vessel, lumen, or neointima area from post-procedure to 9 months. The percent in-stent net volume obstruction and the percent neointima-free length of the stent were also similar (Table 3). Additionally, both groups showed comparable rates for early (direct stent: 8.7% [22 of 254]; control: 6.2% [12 of 193]; p 0.33) and late incomplete apposition (direct stent: 4.3% [9 of 209]; control: 3.9% [7 of 180]; p 0.84). Despite similar amounts of neointimal hyperplasia, late loss, and %DS between the groups, binary angiographic restenosis rates were reduced significantly for direct stent versus the control, even after propensity score adjustment (Fig. 4). The restenosis rates remained similar between the groups at the proximal and distal edges, but direct stenting was associated with a significant reduction of focal in-stent restenosis compared with control. The prevalence of diffuse, proliferative, and totally occluded in-stent restenosis was low and comparable for the groups (Fig. 4). Clinical outcomes through 9 months. Early MACE rates were comparable between groups (in-hospital control 2.6%, direct stent 3.2%, p 0.60; 30-day control 3.1%, direct stent 3.3%, p 0.93). However, 9-month MACE rates were significantly lower for direct stent than control (Table 4). This difference was mainly attributable to a 70% reduction in TLR for direct stent versus control. The groups were similar for myocardial infarction, cardiac death, and noncardiac and total death. Of note, adjusted target vessel revascularization (TVR) and TLR rates were concordant with the unadjusted rates (Table 5). 12-month clinical outcomes. At 12 months, the direct stent group exhibited a sustained 70% reduction for TLR versus the control, which was reflected in a 63% reduction of both TVR and MACE (Fig. 5). Propensity score adjusted rates were concordant with the unadjusted rates (Table 5). Additionally, the reduction in TLR was not affected by Figure 3. Primary End Point: 9-Month Analysis Segment %DS (A) Unadjusted (Unadj) and propensity score-adjusted (Adj) rates are shown for both the intention-to-treat (ITT) and per-protocol (PP) populations. The primary end point is based on the adjusted PP analysis. (B) Schematic representation of noninferiority testing results of analysis segment % diameter stenosis (%DS). The pre-specified noninferiority margin, difference (Diff) between the control and direct stent values, and upper 1-sided 95% confidence interval (CI) are presented.

7 156 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, 2008 Table 3. Angiographic and IVUS Outcomes: Paired Lesion Analysis (n 247) p Value* Pre-procedure QCA RVD (mm) (449) (221) 0.32 MLD (mm) (449) (221) %DS (449) (221) Post-procedure QCA RVD (mm) (447) (221) 0.29 MLD, in-stent (mm) (446) (221) MLD, analysis segment (mm) (447) (221) %DS, in-stent (446) (221) 0.14 %DS, analysis segment (447) (221) Acute gain, in-stent (mm) (446) (221) 0.84 Acute gain, analysis segment (mm) (447) (221) month QCA RVD (mm) (449) (221) 0.58 MLD, in-stent (mm) (448) (221) 0.14 MLD, analysis segment (mm) (449) (221) 0.11 %DS, in-stent (448) (221) 0.09 %DS, analysis segment (449) (221) Late loss, in-stent (mm) (446) (221) 0.76 Late loss, analysis segment (mm) (447) (221) 0.91 Loss index, in-stent (446) (221) 0.49 Loss index, analysis segment (447) (221) month IVUS Change in mean vessel area (mm 2 ) (154) (126) 0.57 Change in mean lumen area (mm 2 ) (177) (144) 0.14 Change in mean neointimal area (mm 2 ) (177) (144) 0.68 In-stent net volume obstruction (%) (177) (144) 0.55 Neointimal free length of stent (%) (177) (144) 0.81 Data shown are for paired analysis subset, including patients for whom both procedure and 9-month data were available. Numbers shown are mean SD (n) or % (n). *p values were calculated by Student t test. Abbreviations as in Tables 1 and 2. greater use of post-dilation in the direct stent group (data not shown). Throughout the 12-month follow-up, there were no episodes of stent thrombosis in the direct stent group versus an incidence of 1.1% in the control. After propensity score adjustment, this difference was statistically significant. The ARC definite or probable stent thrombosis was 0.0% for direct stent versus 1.5% (8 of 523) for pre-dilation control (p ). Predictors of TLR. Multivariate modeling identified RVD, stent expansion index, and implantation technique as predictors of 12-month TLR. Smaller vessels increased the risk for TLR ( 1.24; hazard ratio [HR] 0.29, 95% CI 0.13 to 0.62; p ), such that for each 1-mm decrease in baseline RVD, there was a 29% increased risk for TLR. The stent expansion index was also a negative predictor for TLR ( 0.04; HR 0.96, 95% CI 0.93 to 0.99; p 0.018), with 1% underexpansion resulting in a 4% increased risk of TLR. Finally, direct stenting was an independent negative predictor for TLR ( 0.97; HR 0.38, 95% CI 0.17 to 0.86; p ) and corresponded to a 38% risk reduction for TLR. Discussion The TAXUS ATLAS DIRECT STENT demonstrates that direct deployment of TAXUS Liberté in selected patients is feasible and can shorten procedural times versus pre-dilation while decreasing the rate of procedural complications in simpler lesions. Neointimal hyperplasia was similar after both direct stenting with TAXUS Liberté and stenting with pre-dilation, but direct stenting was associated with reduced rates of binary angiographic restenosis and ischemia-driven TLR. There was also a trend for a reduced risk of stent thrombosis in the first year after direct stent implantation. However, it is unclear whether these benefits are due to direct stenting itself or to the favorable vessel and lesion characteristics that led to selection for the direct stenting group.

8 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, Figure 4. Binary Restenosis Edge Analysis Nine-month results from the intention-to-treat population are shown. (A) Unadjusted binary restenosis rates for the proximal edge, in-stent segment, and distal edge. In-stent restenosis patterns are also indicated for the in-stent segment. (B) Unadjusted and propensity score-adjusted binary restenosis rates of the analysis segment. D diffuse; F focal; P proliferative; TO total occlusion. With the advent of more deliverable stents, direct stenting has become increasingly common. As a thin-strut drug-eluting stent with a lower profile and increased flexibility for enhanced deliverability and conformability, TAXUS Liberté is suitable for direct stenting in selected patients. Indeed, direct stenting was successful in 97.6% of patients in this study, which is similar to the best results reported with bare metal stents in somewhat simpler lesions (1 5). A major benefit of direct stenting is a significant reduction in procedure time and radiation exposure (1-5). However, direct stenting might lead to errors in stent placement because of limited vessel opacification, incorrect stent sizing, underexpansion, or incomplete strut apposition. Stent dislodgement and embolization are theoretical possibilities either at the lesion site or upon withdrawal of an unsuccessfully attempted stent placement. Thus, direct stenting might increase procedural risk or suboptimal late clinical outcomes. In the present study, the benefits of reduced procedural time and radiation exposure were confirmed for TAXUS Liberté without any potential adverse consequences. In particular, the occurrence of incomplete apposition was low and comparable between the groups. The stent expansion index was similar in both groups, and procedural complications such as no reflow, distal embolization, dissection, or thrombus formation were significantly reduced after direct stenting versus pre-dilation. Thus, Table 4. 9-Month Clinical Outcomes in the Intention-to-Treat Population Unadjusted Adjusted (n 247) p Value* (n 247) p Value MACE 14.7% (79/537) 9.1% (22/242) % 10.1% 0.09 Cardiac death 1.3% (7/537) 0.8% (2/242) % 1.0% 0.69 MI 4.3% (23/537) 4.5% (11/242) % 5.7% 0.41 Q-wave 0.9% (5/537) 0.4% (1/242) % 0.3% 0.18 Non Q-wave 3.4% (18/537) 4.1% (10/242) % 5.5% 0.22 TVR, overall 11.2% (60/537) 5.0% (12/242) % 5.4% TLR 7.8% (42/537) 2.9% (7/242) % 3.2% TVR, remote 4.5% (24/537) 2.9% (7/242) % 2.9% 0.32 TVF 14.3% (77/537) 8.7% (21/242) % 9.9% 0.10 Stent thrombosis 0.9% (5/533) 0.0% (0/240) % 0.0% Numbers shown are % (count/total). *p values were calculated by chi-square test unless otherwise noted. p values were calculated by Student t test. p values calculated by Fisher exact test. MACE major adverse cardiac events; MI myocardial infarction; TLR target lesion revascularization; TVF target vessel failure; TVR target vessel revascularization.

9 158 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, 2008 Table Month Clinical Outcomes in the Per-Protocol Population Unadjusted Adjusted (n 247) p Value* (n 247) p Value MACE 16.3% (86/528) 8.9% (21/237) % 9.9% Cardiac death 1.3% (7/528) 0.8% (2/237) % 1.2% 0.87 MI 4.5% (24/528) 4.2% (10/237) % 5.3% 0.63 Q-wave 0.9% (5/528) 0.4% (1/237) % 0.3% 0.17 Non Q-wave 3.6% (19/528) 3.8% (9/237) % 5.0% 0.38 TVR, overall 12.5% (66/528) 4.6% (11/237) % 4.8% TLR 8.3% (44/528) 2.5% (6/237) % 2.6% TVR, remote 5.5% (29/528) 3.0% (7/237) % 3.0% 0.11 TVF 15.9% (84/528) 8.4% (20/237) % 9.6% Stent thrombosis 1.1% (6/523) 0.0% (0/235) % 0.0% Total death 1.5% (8/528) 0.8% (2/237) % 1.2% 0.72 Noncardiac death 0.2% (1/521) 0.0% (0/235) % 0.0% 0.31 Numbers shown are % (count/total). *p values were calculated by chi-square test unless otherwise noted. p values were calculated by Student t test. p values calculated by Fisher exact test. MACE major adverse cardiac events; MI myocardial infarction; TLR target lesion revascularization; TVF target vessel failure; TVR target vessel revascularization. TAXUS ATLAS DIRECT STENT demonstrates that direct stenting can be a safe implantation technique for drug-eluting stents and might decrease the occurrence of procedural complications in selected lesions. It is important to underscore that patients were carefully selected for direct stenting. The QCA analysis revealed that 95% of lesions were 23.5 mm in the direct stent group but 28 mm in the pre-dilation group, indicating that longer lesions were clearly avoided for direct stenting. Lesion calcification was significantly less in the direct stent group, with only 2% of direct stent lesions severely calcified. Additionally, stenosis severity in the direct stent group was less than in the pre-dilation group. However, with more than 70% of lesions classified as American College of Cardiology/American Heart Association type B2 or C, the lesions treated in this study are the most complex enrolled in any direct stenting study to date (1 3,5). Preclinical studies have suggested that direct stenting might be less traumatic than stenting after pre-dilation in that it limits endothelial denudation, edge injury, and the occurrence of dissections or thrombus formation (17,18). Rogers et al. (17) demonstrated that the vascular response to stent implantation in arteries denuded by balloon pre-dilation, including monocyte recruitment, reendothelialization, and neointima development, can be dramatically reduced by direct stenting, which leaves some remnant endothelium. Presumably, this can increase the speed of regeneration of an intact endothelial monolayer and reduce the requirement for endothelial proliferation and migration. However, randomized clinical trials evaluating Figure 5. Freedom From TLR Through 1 Year TLR target lesion revascularization.

10 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, direct stenting of bare metal stents have failed to show a consistent benefit in reducing angiographic restenosis or TLR versus stenting after pre-dilation (1 3,5). Therefore, TAXUS ATLAS DIRECT STENT is one of the first studies to report a beneficial effect of direct stenting on binary restenosis and the need for TLR (48% and 70% relative reduction, respectively, at 9 months). The conflicting impact of direct stenting on TLR and restenosis between the current study and previous bare metal stent studies might be due to differences in vascular healing after implantation of drug-eluting stents versus bare metal stents. Because drug-eluting stents are associated with delayed endothelialization versus bare metal stents, leaving areas of intact endothelium might be more important for drug-eluting stents than bare metal stents. In support of this concept, there was a significant reduction of TLR after direct stenting with drug-eluting stents in a recent small randomized trial (19) and a trend for reduced revascularization in an ad hoc analysis of the sirolimus-eluting stent (13). This outcome might also result from reduced local dissection and post-procedural thrombus after direct stenting, as suggested in this study and by others (17,18). Alternatively, lesions treated in the present study were of higher complexity than those previously studied, so a less traumatic approach might have had a greater effect on these more complex lesions. Conversely, although direct stenting remained a predictor of lower TLR rates after multivariate modeling, we cannot exclude the confounding effect of the slightly shorter lesion length and larger MLD in the direct stent cohort or associated lesion factors that might also have contributed to reduced restenosis. Further studies are necessary to clarify these hypotheses. Study limitations. The TAXUS ATLAS DIRECT STENT is a nonrandomized study using historical controls from TAXUS ATLAS. Although the inclusion/exclusion criteria of both studies were identical, the lack of randomization allows for imbalance in baseline characteristics. Although propensity score adjustments can partially compensate for the lack of randomization, they cannot substitute for randomization. Furthermore, because lesions chosen for direct stenting in this trial were carefully selected, the results might not apply to higher-risk lesions and patients not studied. Finally, although the reduction in TLR is interesting, the study was not powered to detect differences in TLR. Conclusions Direct stent implantation with TAXUS Liberté is a feasible and highly successful technique for carefully selected lesions of limited length and freedom from severe calcification. Direct stenting with TAXUS Liberté is noninferior to stenting with pre-dilation in terms of 9-month %DS and can significantly shorten procedure time, reduce procedural complications in simpler lesions, and might decrease restenosis and TLR rates. Acknowledgments The authors thank Leslie E. Stolz, PHD (Boston Scientific Corporation), for assistance in drafting this manuscript. The authors also thank the investigators and investigational sites for their contributions to the TAXUS ATLAS Direct Stent trial. Reprint requests and correspondence: Dr. John A. Ormiston, Mercy Angiography Unit, Ltd., 1st Floor, 98 Mountain Road, Epsom, Auckland 1003, New Zealand. johno@ mercyangiography.co.nz. REFERENCES 1. Baim DS, Flatley M, Caputo R, et al. Comparison of PRE-dilatation vs direct stenting in coronary treatment using the Medtronic AVE S670 Coronary Stent System (the PREDICT trial). Am J Cardiol 2001;88: Barbato E, Marco J, Wijns W. Direct stenting. Eur Heart J 2003;24: Burzotta F, Trani C, Prati F, et al. Comparison of outcomes (early and six-month) of direct stenting with conventional stenting (a metaanalysis of ten randomized trials). Am J Cardiol 2003;91: Carrie D, Khalife K, Citron B, et al. Comparison of direct coronary stenting with and without balloon predilatation in patients with stable angina pectoris. BET (Benefit Evaluation of Direct Coronary Stenting) Study Group. Am J Cardiol 2001;87: Dawkins KD, Chevalier B, Suttorp M-J, et al. Effectiveness of direct stenting without balloon predilatation (from the Multilink Tetra Randomised European Study [TRENDS]). Am J Cardiol 2006;97: Dawkins KD, Grube E, Guagliumi G, et al. Clinical efficacy of polymer-based paclitaxel-eluting stents in the treatment of complex, long coronary artery lesions from a multicenter, randomized trial: support for the use of drug-eluting stents in contemporary clinical practice. Circulation 2005;112: Stone GW, Ellis SG, Cannon L, et al. Comparison of a polymer-based paclitaxel-eluting stent with a bare metal stent in patients with complex coronary artery disease: a randomized controlled trial. JAMA 2005; 294: Stone GW, Ellis SG, Cox DA, et al. A polymer-based, paclitaxeleluting stent in patients with coronary artery disease. N Engl J Med 2004;350: Agema WRP, Monraats PS, Zwinderman AH, et al. Current PTCA practice and clinical outcomes in The Netherlands: the real world in the pre-drug-eluting stent era. Eur Heart J 2004;25: Abizaid A, Chan C, Lim Y-T, et al. Twelve-month outcomes with a paclitaxel-eluting stent transitioning from controlled trials to clinical practice (the WISDOM Registry). Am J Cardiol 2006;98: Urban P, Gershlick AH, Guagliumi G, et al. Safety of coronary sirolimus-eluting stents in daily clinical practice: one-year follow-up of the e-cypher registry. Circulation 2006;113: Silber S, Hamburger J, Grube E, et al. Direct stenting with TAXUS stents seems to be as safe and effective as with predilatation. A post hoc analysis of TAXUS II. Herz 2004;29: Schluter M, Schofer J, Gershlick AH, Schampaert E, Wijns W, Breithardt G. Direct stenting of native de novo coronary artery lesions with the sirolimus-eluting stent: a post hoc subanalysis of the pooled E- and C-SIRIUS trials. J Am Coll Cardiol 2005;45: Hirohata A, Morino Y, Ako J, et al. Comparison of the efficacy of direct coronary stenting with sirolimus-eluting stents versus stenting with pre-dilation by intravascular ultrasound imaging (from the DIRECT Trial). Am J Cardiol 2006;98:

11 160 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 1, NO. 2, Turco MA, Ormiston JA, Popma JJ, et al. Polymer-based, paclitaxeleluting TAXUS Liberté stent in de novo lesions: the pivotal TAXUS ATLAS trial. J Am Coll Cardiol 2007;49: Ahmed WH. Review of the TAXUS Liberté SR paclitaxel-eluting coronary stent. Expert Rev Med Devices 2007;4: Rogers C, Parikh S, Seifert P, Edelman ER. Endogenous cell seeding: remnant endothelium after stenting enhances vascular repair. Circulation 1996;94: Yun SH, Tearney GJ, Vakoc BJ, et al. Comprehensive volumetric optical microscopy in vivo. Nat Med 2006;12: Katritsis DG, Korovesis S, Karvouni E, et al. Direct versus predilatation drug-eluting stenting: a randomized clinical trial. J Invasive Cardiol 2006;18: APPENDIX For a list of investigators and investigational sites for the TAXUS ATLAS trial, please see the online version of this article.