Author + information
- Received February 4, 2008
- Revision received June 4, 2008
- Accepted June 13, 2008
- Published online January 1, 2009.
- Faisal Latif, MD⁎,
- Neal S. Kleiman, MD†,
- David J. Cohen, MD, MSc‡,
- Michael J. Pencina, PhD§,
- Chen-Hsing Yen, MS§,
- Donald E. Cutlip, MD§,
- David J. Moliterno, MD∥,
- Deborah Nassif, PhD§,
- John J. Lopez, MD¶,
- Jorge F. Saucedo, MD⁎,⁎ (, )
- EVENT Investigators
- ↵⁎Reprint requests and correspondence:
Dr. Jorge F. Saucedo, 920 Stanton L. Young Boulevard, WP 3010, Oklahoma City, Oklahoma 73104
Objectives This study sought to evaluate ischemic and bleeding outcomes in patients with chronic kidney disease (CKD) undergoing percutaneous coronary intervention (PCI) with drug-eluting stents (DES).
Background Previous studies have shown that CKD is associated with poor outcomes after PCI. However, these studies were largely conducted before the introduction of DES and aggressive antithrombotic therapy or were performed in the setting of randomized trials. With data from a contemporary registry, we evaluated the influence of CKD on major cardiovascular events and bleeding complications in unselected “real-world” patients undergoing PCI.
Methods Data from 4,791 patients enrolled in the EVENT (Evaluation of Drug Eluting Stents and Ischemic Events) Registry between July 2004 and September 2005 were analyzed. Patients were stratified into 4 groups: creatinine clearance (CrCl) >75, 50 to 75, 30 to 49 and <30 ml/min.
Results During the index hospital stay, there was a step-wise increase in bleeding complications with decreasing CrCl (3.3%, 5.0%, 8.8%, and 14.3%; p < 0.0001 for trend). Lower CrCl was also associated with more frequent death or myocardial infarction (MI) during the initial hospital stay (p = 0.001) and at 1 year (p < 0.001). These findings were confirmed in multivariate analyses that adjusted for baseline differences in demographic, clinical, and angiographic factors. Use of guideline-recommended medications at 1 year, including aspirin, clopidogrel, angiotensin-converting enzyme inhibitors, and statins, also decreased with declining renal function.
Conclusions Renal function is an independent and powerful predictor of bleeding and ischemic complications in the era of DES and contemporary antithrombotic therapy in patients undergoing PCI. The low use of guideline-recommended drugs among patients with CKD undergoing PCI might contribute to these adverse outcomes and warrants further evaluation.
More than 19 million people in the U.S. suffer from chronic kidney disease (CKD); cardiovascular disease is the major cause of mortality in these patients (1,2). Many clinical studies have demonstrated that patients with CKD have poor outcomes after percutaneous coronary intervention (PCI), including increased rates of death and myocardial infarction (MI) relative to patients with preserved renal function (3,4). However, most of these studies were performed before the introduction of drug-eluting stents (DES) and modern antithrombotic therapy (5,6). Moreover, major cardiovascular disease trials often exclude or limit enrollment of patients with significant kidney disease (7). As such, the clinical outcomes in this patient population are largely unexplored yet remain relevant to the practice of cardiology. The EVENT (Evaluation of Drug Eluting Stents and Ischemic Events) registry is an ongoing multicenter study that aims to evaluate the outcomes of unselected patients undergoing PCI in the era of DES (8,9) and thus provides a unique opportunity to evaluate ischemic and bleeding outcomes of patients with CKD after PCI in the “real-world” setting.
A detailed description of the patient population and the design of the EVENT registry have been published (8). Briefly, it is an ongoing multicenter registry that prospectively enrolls unselected patients undergoing PCI at 42 U.S. centers. Enrollment occurs in discrete “waves” of approximately 2,500 patients at approximately yearly intervals. The current analysis includes patients enrolled in the first 2 waves of the registry. Patients in whom stent implantation was not attempted during PCI or who had undergone PCI or cardiac surgery within the preceding 4 weeks were excluded. Patients underwent PCI for a variety of clinical indications, including ST-segment elevation myocardial infarction (STEMI), acute coronary syndrome (ACS), as well as chronic stable angina or a positive stress test. Data concerning patient characteristics, presentation, and treatment were collected prospectively on standardized case report forms and submitted to the data coordinating center. Protocol-mandated creatine kinase (CK) and creatine kinase-myocardial band (CK-MB) levels were assessed at baseline and every 8 h for a minimum of 2 samples after the procedure and assayed with each site's clinical laboratory and reference values. If MI was suspected clinically at a later point, additional biomarkers were obtained. The institutional review board of each hospital approved the facility's participation in the EVENT registry. Informed consent was obtained from each participating patient. Telephone follow-up was performed at 6 and 12 months after the PCI to ascertain the occurrence of major cardiovascular events.
Assessment of renal function and stratification
Serum creatinine levels were recorded at baseline for all patients. For the purposes of this study, renal function was assessed according to creatinine clearance (CrCl), which was calculated on the basis of age, gender, ideal body weight (kg), and serum creatinine (mg/dl) with the Cockcroft-Gault equation (10). Patients were separated into 4 groups on the basis of CrCl: >75 ml/min, 50 to 75 ml/min, 30 to 49 ml/min, and <30 ml/min. Because the focus of this analysis was on renal function, patients for whom baseline CrCl could not be calculated were excluded from the current study.
All deaths, MIs, and stent thromboses were adjudicated centrally by 2 cardiologists blinded to CrCl. Myocardial infarction was defined as elevation of CK or CK-MB at least 3× the local upper limit of normal (or CK if CK-MB levels were not available). Among patients in whom CK or CK-MB were elevated at baseline, an increase of at least 2× compared with the baseline was necessary as well. Lesion complexity was assessed according to the modified American College of Cardiology/American Heart Association classification system (11). Target lesion revascularization (TLR) was defined as any repeat percutaneous intervention or bypass surgery to treat recurrent stenosis of the original target lesion. Stent thrombosis was defined as angiographic thrombus or subacute closure within the stented vessel at the time of coronary angiography for documented ischemia (chest pain and electrocardiographic changes) as well as any death not attributed to a noncardiac cause within the first 30 days in the absence of documented angiographic stent patency. Bleeding was defined using Thrombolysis In Myocardial Infarction (TIMI) major and minor criteria (12). Vascular access site complications included retroperitoneal bleeding, thrombosis of the access vessel, femoral arterial aneurysm or hematoma necessitating surgical repair, blood transfusion, and increase in the length of hospital stay by at least 1 day. Composite bleeding included access site complication, TIMI major and minor bleeding, and blood transfusion.
The principal end points for this study were all cause mortality, MI, the composite of death or MI, and a composite of bleeding complications consisting of access site complications, TIMI major and minor bleeding, or blood transfusion during the index hospital stay. One-year outcomes included death, MI, and the composite of death or MI. Secondary outcomes included TIMI major and minor bleeding and post-procedure length of stay during index hospital stay, whereas TLR and stent thrombosis were evaluated at 1 year.
For baseline, angiographic, and procedural characteristics, continuous variables are described as mean ± 1 SD and compared among CrCl groups with analysis of variance. Categorical variables are described as proportions and compared by chi-square testing. Clinical outcome and follow-up medication variables are reported as proportions or mean ± 1 SD and compared with logistic, exact logistic, or linear regression as appropriate with a test for trend across the CrCl categories. Time-to-event curves were constructed with the Kaplan-Meier approach. For the analyses of the association between CrCl and clinical outcomes, CrCl was analyzed as a categorical predictor with the 4 categories defined in the preceding text, and a test for trend in categories was conducted. In secondary models it was analyzed as a continuous log-transformed variable. Logistic regression analysis was used for all in-hospital outcomes, and the Cox proportional hazards model was employed for 1-year outcomes, censoring individuals at the last known date of event-free status. In cases where there were <5 events in 1 or more CrCl strata, exact logistic regression was used.
In addition to the univariate models, multivariable regression analyses were conducted (logistic or proportional hazards, as appropriate) adjusting for baseline differences in demographic, clinical, and angiographic factors. In addition to age and gender, the analysis of bleeding complications was further adjusted for hypertension, clinical indication for the procedure (STEMI, ACS or other), hemoglobin, and the use of antithrombotic agents (heparin alone, heparin plus glycoprotein [GP] IIb/IIIa inhibitors, direct thrombin inhibitor [DTI] alone, or heparin/DTI plus GP IIb/IIIa inhibitors). The variables used for the adjustment in the analysis of 1-year ischemic outcomes (death, death or MI) included age, gender, clinical indication for the procedure (STEMI, ACS, chronic stable angina or positive stress test, or other), presence of significant left anterior descending artery disease, as well as composite bleeding.
All statistical analyses were performed with SAS 8.2 (SAS Institute, Cary, North Carolina). Unless specified otherwise, a 2-tailed p value of ≤0.05 was considered statistically significant.
A total of 5,053 consecutive patients were enrolled in Waves 1 and 2 of the EVENT registry between July 2004 and September 2005. After excluding 262 patients with missing baseline renal function assessment, 4,791 patients constituted the study cohort. Approximately 59% (2,827) of patients had normal renal function (CrCl >75 ml/min), whereas 26% (1,253) had mild renal impairment (CrCl 50 to 75 ml/min), 12% (571) had moderate renal impairment (CrCl 30 to 49 ml/min), and 3% (140) had severe renal impairment (CrCl <30 ml/min). Of the 140 patients with severe renal impairment, 51 (approximately 1% of the total sample) were dependent on renal dialysis.
On average, patients with a CrCl below 75 ml/min were older than those with a CrCl above 75 ml/min and were more likely to be female and have a history of congestive heart failure, coronary artery bypass grafting, diabetes, hypertension, stroke, and peripheral arterial disease. Patients with lower CrCl were less likely to be current smokers and to have presented with a STEMI. Other potential indications for PCI did not differ according to renal function. Of note, among patients with at least moderate renal insufficiency, 24% had only a slight increase in serum creatinine (level between 1.5 and 2.0 mg/dl) at baseline (Table 1).
Angiographic characteristics of the study population are displayed in Table 2. A total of 6,655 lesions were treated in 4,791 patients, with a similar number of lesions treated/patient across groups. Approximately 15% of patients underwent multivessel PCI in each group. Percutaneous coronary intervention was performed more frequently in left main and saphenous vein graft lesions in the lowest CrCl group. Additionally, patients with worse renal function had PCI performed in more complex lesions (B2/C per American College of Cardiology classification).
Procedural characteristics of the study population are summarized in Table 3. Most patients received only DES. The use of anticoagulants varied significantly among the groups (p < 0.0001) as shown in Table 3. Unfractionated heparin or DTI alone was used more frequently in patients with lower CrCl. Conversely, GP IIb/IIIa inhibitors were used less frequently in patients with greater renal impairment. Eptifibatide was the predominant GP IIb/IIIa inhibitor used, irrespective of renal function. However, use of abciximab increased as renal function worsened, (CrCl >75 ml/min, 83% eptifibatide vs. 15% abciximab; CrCl <30 ml/min, 57% eptifibatide vs. 41% abciximab). In patients receiving heparin, higher activated clotting time was seen in patients with CrCl between 30 and 75 ml/min (p = 0.03).
Impaired renal function was associated with a graded increase in both ischemic and bleeding complications during the index hospital stay (Table 4). Of the 4,791 patients, 227 developed the composite end point of bleeding complications. There was a progressive increase in the frequency of the composite end point of death or MI during the index hospital stay from 5.8% to 10% with worsening renal function (p = 0.0016), driven primarily by a greater incidence of in-hospital MI among patients with lower CrCl (p = 0.002). Finally, post-procedure length of hospital stay was greater in patients with worse renal function at baseline (Table 4).
During 1 year of follow-up, 139 patients (2.9%) died and 395 (8.2%) experienced an MI. The Kaplan-Meier curves for survival are displayed in Figure 1. With decreasing CrCl, there was a stepwise increase in mortality across all groups (unadjusted p value for trend <0.0001) (Table 4). The incidence of MI also increased with worsening renal function (unadjusted p value for trend 0.0007). Stent thrombosis (p = 0.99) and repeat revascularization (unadjusted p value for trend 0.51) were not different among the various groups.
Trends in cardiovascular medication usage at 6 and 12 months showed a significant decline in use of guideline-recommended pharmacotherapy with decreasing CrCl (e.g., CrCl >75 ml/min vs. CrCl <30 ml/min) at 1 year, aspirin (93% vs. 73%), clopidogrel (72% vs. 63%), hydroxymethylglutaryl-coenzyme A reductase inhibitors (86% vs. 65%), and angiotensin-converting enzyme inhibitors (52% in vs. 34%) as shown in Table 5. The follow-up rates were between 95% and 98% at 6 months and between 92% and 94% at 1 year for all CrCl groups.
Creatinine clearance was a powerful predictor of bleeding complications. This observation held true even when multiple variables including antithrombotic use and baseline hemoglobin were progressively forced into the model (Table 6). The multivariable-adjusted odds of bleeding complications were more than 3× as high in the <30 ml/min CrCl group compared with the referent group of CrCl >75 ml/min (odds ratio: 3.23, 95% confidence interval: 1.72 to 6.04) (Table 6).
At 1 year, renal dysfunction emerged as an independent predictor of mortality after adjustment for the risk factors listed earlier (Table 6). The composite of death and MI also increased across all CrCl groups (p < 0.0001). Individuals with CrCl <30 ml/min had more than 5× the risk of dying than those with CrCl >75 ml/min (multivariable-adjusted hazard ratio: 5.35, 95% confidence interval: 2.86 to 10.02). Noticeably, TLR did not increase with renal dysfunction. Analysis of CrCl as a continuous, log-transformed predictor demonstrated similar findings for in-hospital and 1-year outcomes (data not shown).
Our study generated several important results. Renal insufficiency was associated with increased in-hospital death and MI as well as bleeding complications, even after adjustment for other clinical factors that could potentially confound these associations. One-year ischemic outcomes including death and nonfatal MI also increased in a step-wise fashion as CrCl decreased. Despite these important clinical outcomes, renal dysfunction was not associated with an increased risk of stent-specific events, including stent thrombosis or clinical restenosis (i.e., TLR). We also found that guideline-recommended medications were used less frequently as renal function worsened. We found that, consistent with other reports, more than 40% of patients undergoing PCI have impaired renal function. We also found that almost one-fourth of patients with significant renal insufficiency undergoing PCI have only a mild increase in serum creatinine (between 1.5 to 2.0 mg/dl), emphasizing the advantage of using CrCl over serum creatinine, as the dosage of certain antithrombotic medications need to be adjusted on the basis of CrCl.
Bleeding complications in CKD
Renal dysfunction has been shown to be a risk factor for bleeding complications in patients undergoing PCI in the setting of STEMI as well as ACS (13,14). Few previous studies have demonstrated a linear increase in bleeding after PCI with a concomitant decrease in CrCl (15,16). However, these analyses were either performed in the setting of randomized controlled trials or were in the era of bare-metal stents and non-contemporary antithrombotic therapy, in contrast to our study, which was performed in the real-world setting.
In our study, patients with CrCl 30 to 49 ml/min had higher peak activated clotting times and were particularly more susceptible to hemorrhagic complications. Whereas there has been increasing emphasis on proper medication dosing in patients with renal insufficiency, one possible explanation for our finding is that moderate degrees of renal insufficiency might have been overlooked in some patients. In particular, estimation of renal function on the basis of serum creatinine alone might have led to inadvertent overdosing of antithrombotic medications that are renally metabolized, including GP IIb/IIIa inhibitors such as eptifibatide. Recently, Alexander et al. (17,18) reported that renal dysfunction was an independent predictor of excessive dosing of heparins and GP IIb/IIIa inhibitors and of TIMI major bleeding among patients presenting with ACS and that this practice was more common in women than in men. The appropriate dosing of heparin in patients with CKD has not been studied very well (19). Although excess dosing of antithrombotic agents could also have played a role in increased bleeding complications of patients with CKD, we were unable to prove this association in the EVENT registry, because doses of GP IIb/IIIa antagonists were not systematically collected in our study cohort. It is also likely that patients with CKD are at increased risk of bleeding, even in the absence of dosing errors. For example, renal insufficiency has been associated with several abnormalities of primary hemostasis, particularly platelet dysfunction characterized by decreased release of adenosine triphosphate and decreased serotonin content in dense granules (20).
Blood transfusions, even after adjusting for baseline hemoglobin, were more often administered in patients with decreased renal function, a finding that could be related to a lower threshold for transfusion in patients with severe CKD. Previous studies have shown an increase in blood transfusions in patients with CKD after PCI (3,21).
Ischemic outcomes in CKD
Our study showed a marked increase in in-hospital and 1-year cardiovascular events after PCI as renal function worsened. Other smaller and retrospective studies have shown increased mortality and MI within 1 year after PCI with DES (22), even when the degree of renal insufficiency is mild (23). However, the large population of patients with varying degrees of renal dysfunction in our study provides greater power to examine a broader spectrum of outcomes, including death as well as increased long-term risk of MI, in this patient population.
One of the strengths of our study was that systematic collection of pre- and post-procedure cardiac markers allowed for reliable detection of periprocedural ischemic events without ascertainment bias. Indeed, we observed a stepwise increased incidence of periprocedural MI with worsening renal function. Of note, the use of GP IIb/IIIa antagonists decreased with worsening renal function, which could have played a role in the increased incidence of post-procedural MI in patients with CKD, given their proven role in reducing these complications (24).
Studies have shown that the need for repeat revascularization in patients with CKD has decreased with the advent of DES (21,25). However, a recent study showed that the need for repeat revascularization increases with renal dysfunction, even in patients receiving DES (26). Renal failure has also been implicated as an independent predictor of stent thrombosis with DES, being second only to premature discontinuation of clopidogrel (27). However, repeat revascularization or stent thrombosis was not found to be increased in patients with CKD at 1-year follow-up in our analysis, questioning earlier findings.
Recently, more emphasis is being laid on bleeding complications as a determinant of mortality in patients after PCI. However, our results indicate that even after correcting for bleeding complications, outcomes remain worse in patients with lower CrCl. Therefore, in patients with CKD, additional factors attributable to CKD as well as underuse of recommended cardioprotective medications could account for worse outcomes. It has been reported that cardioprotective medications are less often used in the management of patients with CKD (28). To our knowledge, our study is the first to show the paradoxical underuse of cardioprotective drugs after PCI in a CKD population. Use of medications such as aspirin, clopidogrel, statins, and angiotensin-converting enzyme inhibitors decreased in a step-wise fashion with the drop in CrCl at 1 year. Although further studies are needed to evaluate the safety and efficacy of guideline-recommended therapies in patients with CKD, it seems paradoxical that cardiovascular pharmacotherapy is being underused in patients who might benefit the most.
Although our data were collected prospectively, the current analysis was not specified prospectively. Consequently, the data collected did not include whether the anticoagulants were dosed on the basis of CrCl or serum creatinine or whether renal function was considered at all in dose selection. Similarly, transfusion was a major component of the composite bleeding end point, and we did not collect data regarding indication for transfusion. We had limited power to evaluate some end points such as stent thrombosis and to evaluate dialysis patients separately. Similarly, we did not collect information on post-procedure renal failure or dialysis as complications of PCI.
The Cockcroft-Gault equation was used for estimation of CrCl. This formula might overestimate CrCl in obese patients. Although the number of patients who received bare-metal stents was small, we did not compare the outcomes of patients receiving DES with those of patients receiving bare-metal stents.
Renal dysfunction, as determined by CrCl, is a powerful independent predictor of in-hospital bleeding complications and 1-year ischemic events including death and nonfatal MI in the current era of DES and antithrombotic therapy. Paradoxically, suboptimal long-term use of guideline-recommended cardiovascular pharmacotherapy is observed in patients with CKD after PCI. With the aging of the population and the increase in the number of patients with CKD, the need for their inclusion in ongoing and future cardiovascular trials to determine whether current therapeutic approaches are effective is as great as ever.
Funding for the EVENT registry and its analysis was provided by grants from Millennium Pharmaceuticals and Schering Plough Inc. Dr. Saucedo reports research support from Eli Lilly, Schering Plough, Medicines Co., and Abbott and served as consultant for Schering Plough. Dr. Kleiman reports research support from Cordis Inc., Eli Lilly, Medtronic, Sanofi-Aventis, and Medicines Co.; received honoraria for speaking engagements from Sanofi-Aventis, Cordis, Medtronic, and Medicines Co., and served as a consultant to Boston Scientific and Medtronic. Dr. Cohen received grant support from Cordis, Boston Scientific, Bristol-Myers Squibb, and Eli Lilly. Dr. Lopez received educational grants from Cordis; served on an advisory board for Medicines Co.; and received research support from Conor, Cordis, Medicines Co., and Guidant.
- Abbreviations and Acronyms
- acute coronary syndrome
- creatine kinase
- creatine kinase-myocardial band
- chronic kidney disease
- creatinine clearance
- drug-eluting stent(s)
- direct thrombin inhibitor
- myocardial infarction
- percutaneous coronary intervention
- ST-segment elevation myocardial infarction
- Thrombolysis In Myocardial Infarction
- target lesion revascularization
- Received February 4, 2008.
- Revision received June 4, 2008.
- Accepted June 13, 2008.
- American College of Cardiology Foundation
- U.S. Renal Data System
- Rubenstein M.H.,
- Harrell L.C.,
- Sheynberg B.V.,
- Schunkert H.,
- Bazari H.,
- Palacios I.F.
- Best P.J.M.,
- Lennon R.,
- Ting H.H.,
- et al.
- Smith S.C. Jr..,
- Dove J.T.,
- Jacobs A.K.,
- et al.
- Bovill E.G.,
- Terrin M.L.,
- Stump D.C.,
- et al.
- Sadeghi H.M.,
- Stone G.W.,
- Grines C.L.,
- et al.
- Mathis A.S.,
- Gugger J.J.
- Alexander K.P.,
- Chen A.Y.,
- Newby L.K.,
- et al.
- Das P.,
- Moliterno D.J.,
- Charnigo R.,
- et al.
- Karvouni E.,
- Kastritis D.G.,
- Ioannidis J.P.
- Nakazawa G.,
- Tanabe K.,
- Aoki J.,
- et al.