Author + information
- Received October 15, 2005
- Revision received February 27, 2006
- Accepted March 21, 2006
- Published online October 17, 2006.
- Jacqueline Saw, MD, FRCPC⁎,
- Deepak L. Bhatt, MD, FACC†,⁎ (, )
- David J. Moliterno, MD, FACC‡,
- Sorin J. Brener, MD, FACC†,
- Steven R. Steinhubl, MD, FACC‡,
- A. Michael Lincoff, MD, FACC†,
- James E. Tcheng, MD, FACC§,
- Robert A. Harrington, MD, FACC§,
- Maarten Simoons, MD, FACC∥,
- TingFei Hu, MS†,
- Mobeen A. Sheikh, MD†,
- Dean J. Kereiakes, MD, FACC¶ and
- Eric J. Topol, MD, FACC†
- ↵⁎Reprint requests and correspondence:
Dr. Deepak L. Bhatt, Department of Cardiovascular Medicine, Cleveland Clinic Foundation, 9500 Euclid Avenue, Desk F25, Cleveland, Ohio 44195
Objectives We aimed to evaluate clinical outcomes among peripheral arterial disease (PAD) patients following percutaneous coronary intervention (PCI).
Background A significant proportion of patients with coronary artery disease undergoing PCI have concomitant PAD, which may be associated with worse outcomes.
Methods We performed a pooled analysis of 8 randomized PCI trials. We included multicenter PCI trials that compared antiplatelet therapies (EPIC, EPILOG, EPISTENT, RAPPORT, CAPTURE, IMPACT-II, TARGET, and CREDO) and had baseline PAD status recorded. Multivariable analyses were performed with stepwise logistic regression for 7- and 30-day outcomes and Cox regression for 6-month and 1-year events.
Results In our pooled analysis of 19,867 patients undergoing PCI, 1,602 (8.1%) were previously diagnosed with PAD. Patients with PAD had higher incidences of 7-day death (1.0% vs. 0.4%; p < 0.001) or myocardial infarction (MI) (6.8% vs. 5.6%; p = 0.047), 30-day death (1.7% vs. 0.7%; p < 0.001) or MI (7.4% vs. 6.1%; p = 0.05), 6-month death (4.2% vs. 1.5%; p < 0.001) or MI (9.1%, vs. 7.7%; p = 0.048), and 1-year death (5.0% vs. 2.1%; p < 0.001). There was a trend toward higher major bleeding risk with PAD (4.8% vs. 3.9%; p = 0.06). With multivariable analyses, PAD remains a significant predictor of mortality at 30 days (hazard ratio [HR] 1.67, 95% confidence interval [CI] 1.03 to 2.70; p = 0.039), 6 months (HR 1.76, 95% CI 1.31 to 2.37; p < 0.001), and 1 year (HR 1.46, 95% CI 1.08 to 1.96; p = 0.013).
Conclusions The presence of PAD is associated with higher rates of post-PCI death and MI, and is an independent predictor of short- and long-term mortality.
Peripheral arterial disease (PAD) is prevalent, currently estimated to affect 8 to 12 million Americans and 27 million people in Europe and North America (1). Up to 20% of Americans ≥65 years of age suffer from PAD, and this prevalence is estimated to double by 2050 (1,2). Despite its widespread prevalence, this atherosclerotic manifestation is frequently neglected, with only a quarter of patients undergoing treatment (3). Peripheral arterial disease often coexists with other manifestations of the systemic atherosclerotic process, including coronary artery disease (CAD) and cerebrovascular disease (4). In fact, approximately 2 out of 3 patients with PAD have concomitant CAD, and up to 1 out of 3 elderly patients with CAD have concomitant PAD (5,6). Furthermore, patients with concomitant CAD and PAD who undergo coronary revascularization have higher periprocedural and long-term complications compared with those without PAD (7). This may be due to more extensive CAD or vascular complications related to PAD. Earlier studies which evaluated CAD patients undergoing percutaneous coronary interventions (PCI) involved either relatively small numbers of PAD patients or only single-center registries (8–10). Therefore, we performed a pooled analysis of 8 large multicenter randomized PCI trials to further characterize the impact of PAD on short-term and long-term ischemic events following PCI.
The primary reports of the EPIC (Evaluation of c7E3 for the Prevention of Ischemic Complications), EPILOG (Evaluation of PTCA to Improve Long-Term Outcome by c7E3 GPIIb/IIIa receptor blockade), EPISTENT (Evaluation of IIb/IIIa Platelet Inhibitor for Stenting), CAPTURE (c7E3 Fab Antiplatelet Therapy in Unstable Refractory Angina), RAPPORT (ReoPro and Primary PTCA Organization and Randomized Trial), IMPACT-II (Integrilin to Minimize Platelet Aggregation and Coronary Thrombosis II), TARGET (Do Tirofiban and ReoPro Give Similar Efficacy Trial), and CREDO (Clopidogrel for the Reduction of Events During Observation) trials have been previously reported (11–18). These 8 large multicenter randomized trials were selected because they studied PCI patients and recorded the presence of PAD as a baseline characteristic. The EPIC, EPILOG, EPISTENT, CAPTURE, RAPPORT, and IMPACT-II trials randomized patients to a glycoprotein (GP) IIb/IIIa inhibitor versus placebo. The TARGET study randomized patients to abciximab versus tirofiban, and the CREDO study randomized patients to clopidogrel versus placebo. All patients received aspirin and heparin. Baseline characteristics and outcome data from these 8 trials were extracted from the original trial datasets and pooled for the present analysis. End points of death, myocardial infarction (MI), and target vessel revascularization (TVR) were pooled from studies to assess 7-day, 30-day, and 6-month events. One-year events (of which only mortality data were collected) were available for EPIC, EPILOG, EPISTENT, CAPTURE, TARGET, and CREDO.
The presence of PAD was ascertained and recorded by site investigators at the time of patient enrollment. However, a specific definition of PAD was not stipulated by individual studies. Death was defined as all-cause mortality. Myocardial infarction was typically defined as creatine kinase (CK) or CK-MB elevation of at least 3 times the upper limit of normal in hospital or at least 2 times the upper limit of normal at follow-up or as development of significant Q waves on electrocardiogram. Target vessel revascularization was defined as either percutaneous revascularization or bypass surgery to a vessel that had been treated during the index procedure. Bleeding complications were defined according to the individual studies; the majority used the criteria defined by the Thrombolysis In Myocardial Infarction (TIMI) Study Group (19).
Baseline characteristics and outcomes were stratified by the presence of PAD. Baseline characteristics were compared with chi-square test for discrete variables and Wilcoxon rank-sum test for continuous variables. Hypothesis testing was done using 2-sided tests at the 5% significance level. Survival estimates were computed using Kaplan-Meier estimates and compared using the log-rank test. Multivariable analyses were performed with stepwise logistic regression for 7- and 30-day outcomes and stepwise Cox regression for 30-day, 6-month, and 1-year events. The regression models adjusted for age, body mass index, gender, diabetes mellitus, smoking, hypertension, prior MI, prior stroke, prior congestive heart failure, and baseline creatinine clearance. Hazard ratios (HR) are reported with 95% confidence intervals (CI). In a subanalysis including only the 6 trials that randomized PCI patients to GPIIb/IIIa inhibitor versus placebo, the outcomes evaluating the randomized treatment were stratified according to the presence of PAD. The Breslow-Day test was performed to ensure homogeneity of the 8 randomized trials (p > 0.05 for all outcomes assessed). All statistical analyses were performed with SAS (Version 8; SAS Institute, Cary, North Carolina).
In this pooled analysis of 8 randomized trials, of 19,867 patients who underwent PCI, 1,602 (8.1%) were previously diagnosed with PAD. The baseline characteristics are described in Table 1.Patients with PAD were older, more frequently female, had lower body mass index, more atherosclerotic risk factors (diabetes, hypertension, current smoker), more frequent history of MI, stroke, and congestive heart failure, and lower baseline creatinine clearance.
In the unadjusted analyses, patients with PAD had higher mortality at 7-day (1.0 vs. 0.4%; p < 0.001), 30-day (1.7% vs. 0.7%; p < 0.001 [HR 2.18, 95% CI 1.36 to 3.52; p = 0.001]), 6-month (4.2%, vs. 1.5%; p < 0.001 [HR 2.68, 95% CI 2.01 to 3.59; p < 0.001]), and 1-year (5.0% vs. 2.1%; p < 0.001 [HR 2.25, 95% CI 1.69 to 3.01; p < 0.001]) follow-ups compared with patients without PAD (Fig. 1).The 1-year Kaplan-Meier curve for unadjusted mortality is shown in Figure 2(log-rank chi-square = 48.1). Patients with PAD also had higher incidences of MI at 7-day (6.8% vs. 5.6%; p = 0.047), 30-day (7.4% vs. 6.1%; p = 0.05), and 6-month (9.1% vs. 7.7%; p = 0.048) follow-ups. There were no significant differences on TVR between patients with or without PAD at 7-day (3.9% vs. 3.4%, respectively; p = 0.28), 30-day (4.9% vs. 4.4%; p = 0.37), or 6-month (16.0% vs. 14.9%; p = 0.22) follow-ups. However, PAD was associated with a higher composite end point of death, MI, or TVR at 7-day (10.0% vs. 8.0%; p = 0.004), 30-day (11.7% vs. 9.5%; p = 0.003), and 6-month (24.5% vs. 20.5%; p < 0.001) follow-ups. There was a trend toward a higher incidence of major bleeding (4.8% vs. 3.9%; p = 0.062) and minor bleeding events (8.0% vs. 6.8%; p = 0.066) among patients with PAD.
After multivariable analyses adjusting for all baseline characteristics, the presence of PAD remained an independent predictor of mortality but not of MI events. Peripheral arterial disease was a significant predictor of mortality at 30-day (HR 1.67, 95% CI 1.03 to 2.7; p = 0.039) (Fig. 3A),6-month (HR 1.76, 95% CI 1.31–2.37; p < 0.001) (Fig. 3B), and 1-year (HR 1.46, 95% CI 1.08–1.96; p = 0.013) (Fig. 3C) follow-ups. Independent predictors of 1-year mortality were history of congestive heart failure, diabetes mellitus, presence of PAD, hypertension, prior MI, increasing age, and worse creatinine clearance (Table 2).There was no heterogeneity in the excess mortality among the trials (p > 0.05 for all outcomes). Peripheral arterial disease was a significant predictor of the composite death, MI, or TVR event rate at 6 months (HR 1.17, 95% CI 1.05 to 1.31; p = 0.005).
Randomized trials of GPIIb/IIIa inhibitor versus placebo
In a subanalysis of the 6 randomized PCI trials (n = 12,238) comparing GPIIb/IIIa inhibitor versus placebo (EPIC, EPILOG, EPISTENT, CAPTURE, RAPPORT, and IMPACT-II), we stratified outcomes of randomized treatment according to PAD presence. Among patients with PAD (n = 880), the use of GPIIb/IIIa inhibitor was associated with lower composite death, MI, or TVR event rate at 7 days (9.1% vs. 14.9%; p = 0.009) and 30 days (11.0% vs. 15.6%; p = 0.05) compared with placebo. However, this was no longer significant at 6 months (27.1% vs. 30.2%; p = 0.33) (Fig. 4).There appeared to be a higher risk of major bleeding with the use of GPIIb/IIIa inhibitors in PAD patients, although this was not statistically significant (7.1% for GPIIb/IIIa inhibitors vs. 4.6% placebo; p = 0.15). The use of GPIIb/IIIa inhibitors during PCI confers similar benefit irrespective of the presence of PAD (Fig. 4).
The present analysis evaluates the impact of PAD on the outcomes of patients undergoing PCI. The principal findings are that PAD was independently associated with higher post-PCI mortality at multiple follow-up intervals (30 days, 6 months, and 1 year). Peripheral arterial disease was associated with a trend toward higher major bleeding complications after PCI, particularly with the use of GPIIb/IIIa inhibitors compared with placebo. Furthermore, when compared with patients without PAD, patients with PAD had similar reductions in death, MI, or TVR event rates at 7 and 30 days when randomized to GPIIb/IIIa inhibitors compared with placebo.
Our data are concordant with recent nonrandomized studies evaluating outcomes of PAD patients undergoing PCI (8–10). Chiu et al. (8) performed an analysis of TARGET (n = 4,809), addressing the outcomes of 469 PAD patients (∼10%) undergoing PCI. These patients had 2.3 times higher 1-year mortality compared with those without PAD. In a retrospective, single-center Mayo Clinic PCI registry by Singh et al. (9), 18% (1,397 of 7,696) of patients undergoing PCI had concomitant PAD. These patients were older than non-PAD patients and had higher frequencies of diabetes, hypertension, smoking, and multivessel CAD. Peripheral arterial disease patients had higher in-hospital death, MI, and need for blood transfusion. At a median follow-up of 3.1 years, PAD was associated with higher mortality (HR 1.48; p < 0.001) despite adjusting for baseline variables. Similarly, in another retrospective registry study by Nikolsky et al. (10), the outcomes of 1,969 PAD (18.9% of 10,440) patients undergoing PCI were evaluated. Patients with symptomatic PAD had higher in-hospital complications, 1-year mortality, MI, and TVR. With multivariate analysis, PAD remained an independent predictor of 1-year mortality (OR 1.71; p < 0.001).
Our study extends the observations made in earlier retrospective studies. The data for our pooled analysis is extrapolated from the original databases of large multicenter randomized trials. The data collection (of both baseline characteristics and clinical outcomes) in randomized controlled trials is generally more complete compared with registry studies. Our study population included an aggregate of 19,867 CAD patients undergoing PCI, which represents one of the largest PCI populations to date. Furthermore, the inclusion of 6 randomized trials (n = 12,238) comparing GPIIb/IIIa inhibitors to placebo enabled a pooled and stratified subanalysis of the benefit of GPIIb/IIIa inhibitors for PAD patients, which had not been previously evaluated.
There are several potential mechanistic etiologies for the higher mortality observed in our patient subset with concomitant PAD and CAD. This is likely predominantly attributed to the more extensive systemic atherosclerotic burden in this patient population. Such systemic burden may culminate in not only cardiovascular, but also cerebrovascular complications. Both studies by Rihal et al. (7) and Nikolsky et al. (10) have shown higher prevalence of neurologic complications (stroke or transient ischemic attacks) following PCI in patients with PAD. Furthermore, patients with PAD undergoing PCI were also more likely to develop vascular complications such as retroperitoneal hemorrhage, femoral hematoma, limb ischemia, bleeding, and requirement for blood transfusion (10). In addition, in a recent biochemical study by Narins et al. (20), post-MI patients who have symptomatic intermittent claudication (78 of 1,045 patients) were shown to have enhanced proinflammatory and procoagulant states. Measurements of D-dimer, fibrinogen, C-reactive protein, and serum amyloid A were significantly higher among claudicants versus nonclaudicants.
There are several limitations associated with this study. Despite inclusion of only randomized controlled trials, the data derived from pooled analyses are usually only hypothesis generating. The definition of PAD used in our analysis was not unified in the studies included. The diagnosis of PAD was ascertained and recorded by site investigators; however, the definitions used at each site were not specified in the study protocols. Given the universal underdiagnosis of PAD, it is most likely that our study included higher-risk PAD patients, with symptomatic claudication or critical limb ischemia or who had undergone vascular surgery. Nevertheless, our broad inclusion of PAD patients allows extrapolation of our findings to a more “real-life” PCI population. Although the baseline characteristics of our PAD population were significantly different from the non-PAD population (older, higher incidence of atherosclerotic risk factors, worse renal function, prior MI, stroke, and congestive heart failure), multivariable analyses adjusting for these characteristics were performed. Peripheral arterial disease remained a significant predictor of mortality at 30 days, 6 months, and 1 year. Of note, we did not evaluate the use of beta-blockers in our PAD population (because this was not universally collected in the studies included), which may be a confounder for our study, because claudicants who were not treated with beta-blockers had 3-fold mortality excess (20). Data from randomized controlled PCI trials specifically including a PAD population are necessary to provide definitive conclusions; however, they are unlikely to be pursued.
The presence of concomitant PAD among patients undergoing PCI is associated with higher rates of post-PCI death and MI and is an independent predictor of short- and long-term mortality.
- Abbreviations and Acronyms
- coronary artery disease
- confidence interval
- glycoprotein IIb/IIIa
- hazard ratio
- myocardial infarction
- odds ratio
- peripheral arterial disease
- percutaneous coronary intervention
- target vessel revascularization
- Received October 15, 2005.
- Revision received February 27, 2006.
- Accepted March 21, 2006.
- American College of Cardiology Foundation
- ↵(2003) Heart disease and stroke statistics—2004 update (American Heart Association, Dallas, TX).
- Dieter R.S.,
- Tomasson J.,
- Gudjonsson T.,
- et al.
- Rihal C.S.,
- Sutton-Tyrrell K.,
- Guo P.,
- et al.
- Nikolsky E.,
- Mehran R.,
- Mintz G.S.,
- et al.
- Brener S.,
- Barr L.,
- Burchenal J.,
- et al.
- Rao A.K.,
- Pratt C.,
- Berke A.,
- et al.