Impact of Routine Angiographic Follow-Up on the Clinical Benefits of Paclitaxel-Eluting Stents

Journal of the American College of Cardiology © 2006 by the American College of Cardiology Foundation Published by Elsevier Inc.

Vol. 48, No. 1, 2006 ISSN 0735-1097/06/$32.00 doi:10.1016/j.jacc.2006.02.060

Impact of Routine Angiographic Follow-Up on the Clinical Benefits of Paclitaxel-Eluting Stents Results From the TAXUS-IV Trial Duane S. Pinto, MD, FACC,* Gregg W. Stone, MD, FACC,† Stephen G. Ellis, MD, FACC,‡ David A. Cox, MD, FACC,§ James Hermiller, MD, FACC,储 Charles O’Shaughnessy, MD, FACC,¶ J. Tift Mann, MD, FACC,# Roxana Mehran, MD, FACC,† Yingbo Na, MSC,† Mark Turco, MD, FACC,** Ronald Caputo, MD, FACC,†† Jeffrey J. Popma, MD, FACC,‡‡ Donald E. Cutlip, MD, FACC,* Mary E. Russell, MD, FACC,§§ David J. Cohen, MD, MSC,* for the TAXUS-IV Investigators Boston and Natick, Massachusetts; New York and Syracuse, New York; Cleveland and Elyria, Ohio; Charlotte and Raleigh, North Carolina; Indianapolis, Indiana; and Tacoma Park, Maryland The objectives of the study were to evaluate the effect of angiographic follow-up on revascularization rates in the TAXUS-IV trial and to determine whether the relative benefit of paclitaxel-eluting stent implantation compared with bare metal stent implantation was modified by angiographic follow-up. BACKGROUND Although several clinical trials have demonstrated that drug-eluting stents (DES) reduce restenosis compared with bare-metal stents (BMS), virtually all of these studies have incorporated angiographic follow-up. METHODS In the TAXUS-IV trial, 1,314 percutaneous coronary intervention patients were randomized to receive paclitaxel-eluting stents (PES) (n ⫽ 662) or identical-appearing BMS (n ⫽ 652). Clinical outcomes were compared, stratified by assignment to angiographic follow-up or clinical follow-up alone. RESULTS Compared with clinical follow-up alone, angiographic follow-up patients had a significantly higher rate of target vessel revascularization (TVR) at 1 year (adjusted hazard ratio [HR] 1.46; p ⫽ 0.04), with similar relative increases in PES and BMS patients. Because PES reduced TVR by ⬃60% regardless of type of follow-up, assignment to angiographic follow-up tended to overestimate the absolute benefit of PES relative to clinical follow-up alone. In contrast, assessment of end points immediately before the time of follow-up angiography led to substantial underestimation of the absolute benefit of PES implantation. CONCLUSIONS Performance of mandatory angiographic follow-up increases rates of TVR among patients receiving both BMS and PES and overestimates the absolute clinical benefits of PES relative to clinical follow-up alone. Nonetheless, PES substantially reduces TVR regardless of assignment to mandatory angiographic follow-up or not. Future studies designed to determine the true clinical benefits of DES should either forgo routine angiographic follow-up or separate the time of repeat angiography from the primary clinical end point by as long as possible. (J Am Coll Cardiol 2006;48:32– 6) © 2006 by the American College of Cardiology Foundation OBJECTIVES

Since its first description by Gruentzig et al. (1), restenosis has remained the “Achilles heel” of percutaneous coronary revascularization, necessitating repeat revascularization in 15% to 30% of patients during the first year of follow-up, depending on a variety of patient- and lesion-specific characteristics (2). In several pivotal trials (3–5), polymerbased sirolimus- and paclitaxel-eluting stents reduced the incidence of clinical and angiographic restenosis by 60% to 80% compared with bare-metal stents (BMS), leading to From the *Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, Massachusetts; †Columbia University Medical Center and Cardiovascular Research Foundation, New York, New York; ‡Cleveland Clinic Foundation, Cleveland, Ohio; §Mid Carolina Cardiology, Charlotte, North Carolina; 储St. Vincent’s Hospital, Indianapolis, Indiana; ¶Elyria Memorial Hospital, Elyria, Ohio; #WakeMed, Raleigh, North Carolina; **Washington Adventist Hospital, Tacoma Park, Maryland; ††St. Joseph’s Hospital, Syracuse, New York; ‡‡Brigham and Women’s Hospital, Boston, Massachusetts; and §§Boston Scientific Corporation, Natick, Massachusetts. Supported by a grant from Boston Scientific Corporation. Manuscript received November 18, 2005; revised manuscript received February 22, 2006, accepted February 28, 2006.

U.S. Food and Drug Administration (FDA) approval of both drug-eluting stent systems and a revolution in the practice of interventional cardiology. Unresolved questions persist, however, despite widespread adoption of drugeluting stents (DES). In particular, because all of the major DES trials to date have required angiographic follow-up (3–5), which is known to accentuate the rates of repeat revascularization compared with clinical follow-up alone (6), the true benefit of DES implantation in routine clinical practice is unknown. To address these issues, we used data from the TAXUS-IV trial to examine the relative and absolute clinical benefits of paclitaxel-eluting stent (PES) implantation in the absence of mandatory angiographic follow-up. In addition, we sought to determine whether angiographic follow-up has a differential effect on rates of repeat revascularization among patients undergoing BMS or PES implantation. Finally, we used these data to examine how altering the timing of ascertainment of clinical end points

Pinto et al. Angiographic Follow-Up and DES

JACC Vol. 48, No. 1, 2006 July 4, 2006:32–6

Abbreviations and Acronyms BMS ⫽ bare-metal stent DES ⫽ drug-eluting stent PCI ⫽ percutaneous coronary intervention PES ⫽ paclitaxel-eluting stent TLR ⫽ target lesion revascularization TVR ⫽ target vessel revascularization

33

clinical, and angiographic factors. The independent effects of treatment assignment and planned angiographic follow-up were tested by adding appropriate dummy variables to the baseline model. Finally, an interaction term was added to the model to test whether the effect of angiographic follow-up on TVR differed according to stent type.

RESULTS relative to angiographic follow-up might affect the apparent rates of clinical restenosis in the clinical trial setting.

METHODS Study design. The TAXUS-IV trial design has been described previously (5). Briefly, 1,314 patients undergoing percutaneous coronary intervention (PCI) for a de novo lesion were randomly assigned to either the Taxus stent (Boston Scientific Corp., Natick, Massachusetts) or the indistinguishable bare-metal Express stent (Boston Scientific Corp.). Originally, the first 536 patients enrolled were assigned to planned angiographic follow-up at 9 months (⫾2 weeks). To allow adequate evaluation of long stents, the angiographic follow-up population was subsequently expanded by 196 patients, all receiving 32-mm stents. Because the goal of the present study was to directly compare outcomes between patients assigned to undergo angiographic follow-up versus clinical follow-up alone, we excluded these additional 196 patients from our primary analytic cohort to enhance comparability between the angiographic and clinical follow-up cohorts. Therefore, the initial 536 patients comprised the angiographic follow-up group, and the subsequent 582 patients who were managed without mandatory angiographic follow-up, comprised the clinical follow-up group. Qualitative and quantitative angiographic analyses were performed using previously described techniques and standardized definitions (5). Clinical end points (death [cardiac or noncardiac], myocardial infarction, and target lesion revascularization [TLR]) and the primary end point (target vessel revascularization [TVR]) were determined by contacting patients at 30 days and 6, 9, and 12 months after randomization. An independent committee, blinded to treatment assignment, adjudicated all events, including whether repeat revascularization was clinically indicated. Statistical analysis. We compared clinical outcomes stratified by treatment assignment and by whether patients were assigned to angiographic follow-up or clinical follow-up alone. Continuous variables are described as mean ⫾ standard deviation and were compared by t tests. Categoric variables are presented as frequencies and were compared using chi-squared tests. One-year clinical outcomes are reported as Kaplan-Meier estimates and were compared by the log rank statistic. The Cox proportional hazards model was used to identify independent predictors of TVR at 1 year, based on a prespecified set of 24 sociodemographic,

Patient population. Baseline clinical and angiographic characteristics of the study population, stratified by treatment group (Taxus [PES] vs. control [BMS]) and intention to perform angiographic follow-up are summarized in Table 1. There were no significant differences in clinical or angiographic characteristics comparing patients assigned to angiographic or clinical follow-up alone. Of the patients assigned to angiographic follow-up, 76.3% completed the follow-up angiogram, with a similar proportion for the PES and BMS groups (74.8% vs. 77.9%). Baseline clinical and angiographic characteristics were generally well matched between the PES and BMS groups, although diabetes was more prevalent in patients assigned to BMS in the clinical follow-up cohort. Clinical outcomes. At 1 year, there was a 38% increase in the primary end point of adjudicated TVR among patients assigned to angiographic versus clinical follow-up (13.7% vs. 9.9%; p ⫽ 0.06). In particular, TVR-PCI was significantly increased among those assigned to angiographic follow-up (11.9% vs. 6.7%; p ⫽ 0.005), whereas there was no difference in TVR-coronary artery bypass graft (2.1% vs. 3.8%; p ⫽ 0.12). The relative increase in TVR associated assigned to angiographic follow-up was similar between patients who received BMS or DES (36% vs. 46%, respectively; p ⫽ NS for interaction). A higher rate of myocardial infarction occurred among the clinical follow-up group (2.3% vs. 4.7%; p ⫽ 0.023), with no difference in mortality (1.1% vs. 1.0%; p ⫽ 0.89). With clinical follow-up alone, randomization to PES was associated with a 65% reduction in adjudicated TVR compared with BMS (5.5% vs. 14.3%; p ⬍ 0.001) (Fig. 1). With mandatory angiographic follow-up, randomization to PES was associated with a 59% reduction in TVR at 1 year compared with BMS (8.5% vs. 18.9%; p ⬍ 0.001). Similar findings were seen for TLR. Randomization to PES was associated with a 66% reduction in TLR in the angiographic follow-up group (5.9% vs. 15.9%; p ⬍ 0.001) and a 79% reduction in the clinical follow-up group (2.8% vs. 13.1%; p ⬍ 0.001). For both the clinical and the angiographic follow-up groups, the Kaplan-Meier event curves began to diverge at approximately 3 months from the index procedure. With angiographic follow-up, however, there was a striking increase in TVR at 9 months, reflecting the impact of angiographically driven repeat revascularization. Multivariable analysis. Multivariable analysis identified diabetes mellitus (adjusted hazard ratio [HR] 1.80; p ⫽ 0.002), maximum device diameter (adjusted HR 0.56; p ⫽

34

Pinto et al. Angiographic Follow-Up and DES

JACC Vol. 48, No. 1, 2006 July 4, 2006:32–6

Table 1. Baseline Characteristics Stratified by Angiographic or Clinical Follow-Up and Treatment Assignment Angiographic Follow-Up

Clinical characteristics Age Male Previous smoker Diabetes mellitus Previous CABG Previous MI Hypertension Hyperlipidemia LVEF Unstable angina Angiographic characteristics Lesion location: RCA Lesion location: LCx Lesion location: LAD ACC/AHA: B2 ACC/AHA: C Total occlusion Baseline RVD, mm Baseline lesion length, mm Baseline DS, % Final RVD, mm

Clinical Follow-Up

Taxus (n ⴝ 269)

Control (n ⴝ 267)

Taxus (n ⴝ 287)

Control (n ⴝ 295)

p Value*

63.7 ⫾ 10.8 70.6% 41.0% 28.3% 9.7% 30.1% 73.5% 67.8% 55.2 ⫾ 10.0 37.5%

61.3 ⫾ 10.4 73.0% 47.5% 22.8% 10.9% 26.6% 66.2% 65.0% 56.0 ⫾ 11.0 33.7%

62.4 ⫾ 11.2 72.8% 40.4% 17.8% 8.0% 30.0% 66.4% 64.3% 55.8 ⫾ 10.0 33.8%

62.2 ⫾ 11.4 72.9% 46.3% 26.4%† 8.5% 34.2% 71.9% 69.1% 55.2 ⫾ 9.8 30.8%

0.71 0.74 0.81 0.21 0.26 0.17 0.85 0.95 0.84 0.26

24.6% 34.0% 41.4% 33.2% 13.4% 1.9% 2.7 ⫾ 0.5 12.3 ⫾ 5.0 66.4 ⫾ 10.4 2.8 ⫾ 0.5

31.6% 26.3% 42.1 34.6% 14.7% 0.4% 2.8 ⫾ 0.5 11.9 ⫾ 5.2 65.1 ⫾ 10.0 2.8 ⫾ 0.5

31.1% 30.1% 38.8% 37.1% 12.6% 1.4% 2.6 ⫾ 0.5 12.0 ⫾ 5.5 66.8 ⫾ 10.8 2.8 ⫾ 0.5

29.6% 27.2% 43.2% 29.9% 15.3% 1.7% 2.7 ⫾ 0.5 12.1 ⫾ 4.6 65.4 ⫾ 10.3 2.7 ⫾ 0.5

0.43 0.60 0.81 0.90 0.99 0.61 0.50 0.85 0.60 0.67

*The p values are for the comparison between the entire angiographic follow-up and entire clinical follow-up groups. †p ⬍ 0.05 for comparison of the Taxus and control groups among patients assigned to clinical follow-up alone; p ⫽ NS for all other comparisons for Taxus and control groups. ACC/AHA ⫽ American College of Cardiology/American Heart Association; CABG ⫽ coronary artery bypass graft surgery; CHF ⫽ congestive heart failure; DS ⫽ diameter stenosis; LCx ⫽ left circumflex artery; LAD ⫽ left anterior descending artery; LVEF ⫽ left ventricular ejection fraction; MI ⫽ myocardial infarction; RCA ⫽ right coronary artery; RVD ⫽ reference vessel diameter.

0.008), and total stent length (adjusted HR 1.37 per 10 mm; p ⫽ 0.001) as independent predictors of TVR. Randomization to PES was associated with a 61% reduction in TVR (adjusted HR 0.39; p ⬍ 0.001), whereas assignment to angiographic follow-up was associated with a 46% increase in this end point (adjusted HR 1.46; p ⫽ 0.04). There was no evidence for a significant interaction between treatment assignment and the impact of angiographic follow-up on TVR (multivariate p value for interaction 0.77). For the angiographic cohort randomization to PES was associated with a 59% reduction in TVR (adjusted HR 0.41; p ⬍ 0.001), and for the clinical follow-up group randomization to PES was associated with a 65% reduction in this end point (adjusted HR 0.65; p ⬍ 0.001). To determine whether restriction of our analysis to the time frame before performance of angiographic follow-up would correct for the observed angiographic bias, we repeated all analyses after censoring events beyond 8.5 months, a time point immediately before performance of the angiographic follow-up procedures (Table 2). These analyses revealed no difference in the relative benefit conferred by PES implantation according to the type of follow-up. At 8.5 months, randomization to the PES was associated with a 54% reduction in the rate of TVR among the angiographic follow-up group compared with a 69% reduction in the clinical follow-up group (p value for interaction 0.77). On the other hand, the type of follow-up impacted the magnitude of absolute risk reduction provided by PES implantation when events after 8.5 months were excluded. Among the angiographic follow-up group,

randomization to PES was associated with an absolute reduction of 46 events per 1,000 patients treated compared with an absolute risk reduction of 79 events per 1,000 among the clinical follow-up group. Thus, censoring patients before mandatory angiographic follow-up at 8.5 months was associated with substantial underestimation of the absolute benefit of PES implantation compared with clinical follow-up alone.

DISCUSSION Although randomized clinical trials incorporating high rates of angiographic follow-up have demonstrated that both sirolimus-eluting and PES substantially suppress neointimal proliferation after stent implantation (3–5), the precise magnitude of clinical benefit conferred by DES is less well defined. By incorporating a large cohort of patients managed without mandatory angiographic follow-up, the TAXUS-IV trial provided the opportunity to examine the true clinical benefit of DES compared with BMS. The principal finding of our study was that, although angiographic follow-up increased the rate of adjudicated TVR by ⬃40%, PES still provided substantial benefit regardless of assignment to clinical or angiographic follow-up. The observation that angiographic follow-up increases TLR and TVR rates compared with clinical follow-up alone (the “oculostenotic reflex”) has been made previously with both balloon angioplasty (7) and BMS (6). The present results differ somewhat from these previous trials, however. In the BENESTENT-II study (BElgium-NEtherlands

JACC Vol. 48, No. 1, 2006 July 4, 2006:32–6

Pinto et al. Angiographic Follow-Up and DES

35

Figure 1. Curves showing freedom from target vessel revascularization at 1 year among patients assigned to clinical follow-up alone (A) or to mandatory 9-month angiographic follow-up (B). Events are summarized as unadjusted Kaplan-Meier estimates and comparisons between the groups were made with a log-rank test.

STENT), the only previous study to compare the impact of protocol-specified angiographic follow-up on clinical outcomes for two treatments with different rates of angiographic restenosis, the effect of angiographic follow-up nearly obscured the true clinical benefit of reduced restenosis with stenting. Among BENESTENT-II patients assigned to clinical follow-up alone, stenting led to a 50% relative reduction in TVR compared with angioplasty, whereas the reduction in TVR was only 12% among those patients assigned to angiographic follow-up (7). In contrast, in TAXUS-IV trial angiographic follow-up increased the rate of TVR by ⬃40% for both DES and BMS, but the relative clinical benefit of DES versus BMS remained similar regardless of assignment to angiographic follow-up or not. Because the overall rate of TVR was higher in the

angiographic follow-up group, however, the absolute clinical benefit of DES placement at 1 year was greater in the group assigned to angiographic follow-up compared with clinical follow-up alone (104 vs. 88 events prevented per 1,000 patients treated) (Table 2). The precise explanation for this important difference between our study and previous studies is unknown. One possibility is that in BENESTENT-II the visible outline of the vessel provided by the stent at angiographic follow-up may have led the operator to overestimate the degree of restenosis and the need for reintervention preferentially among stent patients. Because TAXUS-IV trial was a double-blind study, however, revascularization decisions at angiographic follow-up were less likely to have been affected by treatment assignment.

36

Pinto et al. Angiographic Follow-Up and DES

JACC Vol. 48, No. 1, 2006 July 4, 2006:32–6

Table 2. Absolute and Relative Reductions in Target Vessel Revascularization at 8.5 Months and 1 Year Angiographic Follow-Up

1 year 8.5 months

Clinical Follow-Up

PES

BMS

p Value*

ARR

RRR

PES

BMS

p Value*

ARR

RRR

8.5% 4.1%

18.9% 8.7%

⬍0.001 0.03

104/1000 46/1000

59% 54%

5.5% 3.5%

14.3% 11.4%

⬍0.001 ⬍0.001

88/1000 79/1000

65% 69%

*The p values are for the comparison between PES and BMS patients within the angiographic or clinical follow-up groups. ARR ⫽ absolute risk reduction for PES compared with BMS; BMS ⫽ bare-metal stent; PES ⫽ paclitaxel-eluting stent; RRR ⫽ relative risk reduction for PES compared with BMS.

The present study is also the first to formally examine the impact of angiographic follow-up on clinical end points when ascertained immediately before the time of follow-up angiography. Although previous studies have used a strategy of analyzing clinical end points before the time frame of planned angiographic follow-up to minimize the impact of the “oculostenotic reflex” on outcomes (8), we found that this approach resulted in overcorrection for angiographic bias, with underestimation of both the true rates of clinical restenosis as well as the benefits of DES implantation. These findings were particularly striking when examining differences in absolute risk reduction. At 8.5 months, the Taxus stent was associated with a reduction of 46 TVR events per 1,000 with angiographic follow-up compared with a reduction of 79 events per 1,000 patients undergoing clinical follow-up alone. Therefore, assessment of the clinical benefit of the Taxus stent at 8.5 months in the angiographic follow-up group would have led to a substantial underestimate in the absolute risk reduction for TVR at 8.5 months and an even greater underestimate relative to the true clinical benefit of PES at 12 months (88 events per 1,000). Inspection of the Kaplan-Meier curves in patients treated with BMS (Fig. 1) suggests that this type of bias might be avoided by performing routine angiographic follow-up at 12 months, a point where the vast majority of clinical restenosis has become apparent (6). These findings have important implications for future studies of DES (or other antirestenotic technologies). Because there was no evidence of differential bias in revascularization decisions between the angiographic and clinical follow-up arms, studies with high rates of angiographic follow-up can provide valid estimates of the relative clinical benefit of DES in double-blinded studies. However, conclusions about the absolute benefit of DES (as required for risk/benefit or cost-effectiveness calculations) can only be derived from studies not requiring angiographic follow-up or from studies delaying such follow-up until well after the clinical end points have accrued. Conclusions. Incorporation of protocol-specified mandatory angiographic follow-up in studies of angioplasty devices

and techniques is a well accepted method for studying the in vitro response to vascular injury. As demonstrated in the present report, however, routine follow-up angiography after coronary stenting results in a significant number of additional revascularization procedures, despite careful attempts to adjudicate and include only those events that are truly driven by ischemia. Therefore, trials designed to determine the true relative and absolute clinical benefits of new antirestenosis technologies (including DES) should either forgo routine angiographic follow-up or separate the time of repeat angiography from the primary clinical end point by as long as possible. Reprint requests and correspondence: Dr. Duane S. Pinto, Division of Cardiology, Beth Israel Deaconess Medical Center, 1 Deaconess Road, Boston, Massachusetts 02115. E-mail: dpinto@ bidmc.harvard.edu.

REFERENCES 1. Gruentzig A, Senning A, Siegenthaler W. Nonoperative dilatation of coronary-artery stenosis: percutaneous transluminal coronary angioplasty. N Engl J Med 1979;301:61– 8. 2. Holmes DR, Kip KE, Yeh W, Kelsey SF, Detre KM, Williams DO. Long-term analysis of conventional coronary balloon angioplasty and an initial “stent-like” result: the NHLBI PTCA registry. J Am Coll Cardiol 1998;32:590 –5. 3. Morice M-C, Serruys PW, Sousa JE, et al. A randomized comparison of a sirolimus-eluting stent with a standard stent for coronary revascularization. N Engl J Med 2002;346:1773– 80. 4. Moses JW, Leon MB, Popma JJ, et al. Sirolimus-eluting stents versus standard stents in patients with stenosis in a native coronary artery. N Engl J Med 2003;349:1315–23. 5. Stone GW, Ellis SG, Cox DA, et al. A polymer-based, paclitaxeleluting stent in patients with coronary artery disease. N Engl J Med 2004;350:221–31. 6. Cutlip DE, Chauhan MS, Baim DS, et al. Clinical restenosis after coronary stenting: perspectives from multicenter clinical trials. J Am Coll Cardiol 2002;40:2082–9. 7. Serruys PW, van Hout B, Bonnier H, et al. Randomised comparison of implantation of heparin-coated stents with balloon angioplasty in selected patients with coronary artery disease (Benestent II). Lancet 1998;352:673– 81. 8. Stone GW, Grines CL, Cox DA, et al. Comparison of angioplasty with stenting, with or without abciximab, in acute myocardial infarction. N Engl J Med 2002;346:957– 66.