Author + information
- Received February 20, 2004
- Revision received July 6, 2004
- Accepted July 12, 2004
- Published online November 2, 2004.
- Patricia J.M. Best, MD*,* (, )
- Peter B. Berger, MD†,
- Barry R. Davis, MD, PhD‡,
- Cindy L. Grines, MD§,
- H. Mehrdad Sadeghi, MD§,
- Brent A. Williams, MS*,
- James T. Willerson, MD∥,
- Jeffrey R. Granett, MD¶,
- David R. Holmes Jr, MD*,
- PRESTO Investigators
- ↵*Reprint requests and correspondence:
Dr. Patricia J. M. Best, Mayo Clinic, 200 First Street SW, Rochester, Minnesota 55905
Objectives The goal of this study was to determine if restenosis is increased in mild and moderate chronic kidney disease (CKD) patients after percutaneous coronary intervention (PCI).
Background Mortality is increased in CKD after PCI. Restenosis may contribute to increased late mortality.
Methods We analyzed 11,187 patients with a creatinine <1.8 mg/dl from the Prevention of REStenosis with Tranilast and its Outcomes (PRESTO) trial, grouped by estimated creatinine clearance (CrCl) (<60, 60 to 89, >89 ml/min). The Cox proportional hazards models investigated the association between CrCl group and death, myocardial infarction, and target vessel revascularization (TVR). Generalized estimating equation regression models determined the association between CrCl group and lesion-specific restenosis.
Results At 30 days, there was no difference in myocardial infarction, death, or TVR between the CrCl groups. At nine months, mortality was higher in the lowest CrCl group (2.2%, 1.2%, 0.8%; p < 0.001), which was no longer significant after adjusting for confounding variables. Myocardial infarction and TVR were not different between the groups. In patients undergoing protocol follow-up angiography, restenosis (≥50%) was not increased with CKD (32%, 32%, 37%; p = 0.02).
Conclusions Mortality nine months after PCI is mildly increased in mild or moderate CKD patients. However, restenosis is not and does not account for the increased mortality.
Mild chronic kidney disease (CKD) is associated with lower procedural success and increased in-hospital complications after percutaneous coronary intervention (PCI) (1). Even if the PCI procedure is successful, cardiovascular events and mortality are increased up to 50% compared with patients having normal renal function (2). The cause of the increased mortality is unknown, but could relate to increased restenosis or accelerated atherosclerosis progression elsewhere in the coronary circulation. No studies in the current era evaluated both clinical and angiographic restenosis in mild and moderate CKD patients. Thus, this prospectively defined substudy of the Prevention of REStenosis with Tranilast and its Outcomes (PRESTO) trial was designed to address this issue.
In the PRESTO trial, 11,484 patients were enrolled from 432 centers after successful PCI of ≥1 vessel (3,4). Patients were excluded if a myocardial infarction (MI) occurred ≤48 h, if the PCI was staged, if a prior PCI was ≤3 months, if the creatinine was ≥1.8 mg/dl, if anemia or chronic liver disease was present, or if creatinine was not available at study entry (n = 297). This analysis included all remaining 11,187 patients.
Randomization and treatment
Patients were randomly assigned to one of five parallel groups including placebo or tranilast. Because tranilast did not affect any outcome measured (death, MI, restenosis, and target vessel revascularization [TVR]), regardless of renal function, patients from all groups were analyzed together (4).
Creatinine clearance (CrCl) was estimated using the Cockroft-Gault formula (5). Patients were divided into three CrCl groups: >89 ml/min, 60 to 89 ml/min, and <60 ml/min.
Clinical follow-up occurred at one, three, and nine months. A prespecified subgroup of patients underwent nine-month angiographic follow-up or sooner if indicated. Complete revascularization was defined as no remaining stenosis ≥70% by visual analysis and restenosis was defined as a luminal diameter stenosis ≥50% at follow-up on a per-lesion basis.
The primary end points for this analysis were TVR and restenosis at nine-month follow-up. Target vessel revascularization was defined as repeat PCI or coronary artery bypass graft (CABG) surgery performed for revascularization of a stenosis in the vessel originally treated. Target vessel revascularization-free survival time was measured in days from the date of randomization to the date of TVR or last known date the patient was known to be alive without having a TVR. Patients known to be TVR-free were censored. For nine-month TVR, event-free survival time was truncated at 270 days; for 30-day TVR event-free survival was truncated at this time point. Secondary end points include the composite of death/MI/TVR as well as death and MI individually.
Hazard ratios (HRs) for outcomes were estimated using the Cox proportional hazards models comparing the mild and moderate CKD groups with the normal group. Adjusted HRs were estimated after including potential confounding variables in the model, demonstrating a statistically significant association with the specific outcome using a stepwise variable selection procedure. The CKD variables, age, and gender were forced into all of the Cox models.
In order to account for multiple treated lesions per patient, generalized estimating equation regression models were developed to assess the association between CKD groups and restenosis at nine-month follow-up using the SAS procedure GENMOD (SAS Institute Inc., Cary, North Carolina). An exchangeable correlation matrix for the multiple lesions was specified, which assumes a constant correlation among all pairs of lesions within an individual. Odds ratios were estimated comparing the CKD groups with the normal group. Adjusted odds ratios were estimated after including other covariates significantly associated with restenosis in the multivariable model.
Data are presented as the mean value ± 1 standard deviation for continuous variables and percentages for dichotomous variables. All analyses were performed using SAS software (SAS Institute Inc.). Comparisons between groups for patient-specific variables were made with the one-way analysis of variance, chi-square test, or log-rank test. All tests of significance were two-tailed. Lesion-specific characteristics were compared across groups using generalized estimating equations and repeated measures analysis of variance for categorical and continuous variables, respectively.
Baseline patient characteristics
The baseline patient characteristics are shown in Table 1.Frequency of tranilast treatment was not different between CrCl groups (80%, p = 0.92).
The lowest CrCl group had a greater number of diseased coronary vessels than the higher CrCl groups, but had a similar number of vessels treated (Table 2).Stents were used in 75% of the lesions in the lowest CrCl group, 76% in the middle group, and 78% in the highest CrCl group (p = 0.003). Patients in the lowest CrCl group more frequently had calcified (19%, 17%, 13%; p < 0.001), restenotic (15%, 13%, 13%; p = 0.039), and tortuous lesions (9%, 7%, 7%; p = 0.032). The lowest CrCl group less commonly had bifurcation lesions (10%, 11%, 12%; p = 0.09) and total occlusions treated (5%, 6%, 8%; p < 0.001). Angiographic follow-up was lowest in the lowest CrCl group (67%, 70%, 74%; p = 0.004). The angiographic lesion characteristics (Table 3)demonstrate that the highest CrCl group had a greater pretreatment lesion stenosis.
Clinical and angiographic outcomes
At 30 days, there were no differences in death, MI, repeat revascularization, or need for CABG surgery between the renal function groups (Table 4).However, at nine-month follow-up (mean 249 days), mortality was nearly three times greater in the lowest CrCl group (2.2%) than the highest CrCl group (0.8%). There were no differences between the CrCl groups in frequency of MI, PCI, CABG surgery, or TVR. Moreover, after protocol-driven follow-up angiography (n = 2,556), restenosis was not increased with CKD (32%, 32%, 37%; p = 0.02) and actually highest in the highest CrCl group (Table 5).
Using a Cox proportional hazards model, the risk of death at nine months was lower in the highest CrCl group compared with the lowest CrCl group (Table 6).There was no difference in the risk of MI or TVR. When the Cox proportional hazards model was adjusted for age, gender, diabetes, hypertension, smoking status, prior CABG surgery, peripheral vascular disease, congestive heart failure, and unstable angina, the risk of death was no longer different between the CrCl groups (Table 7).Additionally, in the angiographic substudy of 2,556 successfully treated lesions, after adjustment, creatinine was not independently associated with a higher frequency of restenosis as was hypothesized and was actually associated with a lower risk of restenosis (Table 5).
Interaction with diabetes
All follow-up outcomes were analyzed across diabetics and non-diabetics, respectively. The interaction terms between CrCl groups and diabetes were not significant for any of the outcomes: follow-up stenosis ≥50% (p = 0.37), death (p = 0.52), MI (p = 0.72), TVR (p = 0.95), and death/MI/TVR (p = 0.98). These results suggest that the effect of CrCl group on outcome does not differ between diabetics and non-diabetics.
Interaction with stenting
All follow-up outcomes were analyzed across stenting and non-stenting groups, respectively. The interaction terms between CrCl groups and stent use were not significant for any of the outcomes: follow-up stenosis ≥50% (p = 0.16), death (p = 0.91), MI (p = 0.44), TVR (p = 0.83), and death/MI/TVR (p = 0.92). These results suggest the effect of CrCl group on outcome does not differ between those receiving stents and those not receiving stents.
In this, the largest PCI restenosis study ever reported, we found that patients with mild or moderate CKD, restenosis, and TVR are not increased after PCI in the current stenting era. Myocardial infarctions were not increased during the nine months after PCI in CKD patients. The lowest CrCl group had a 2.7-fold increased risk of death within nine months compared with the highest CrCl group. However, after adjusting for other variables, CrCl group was no longer independently associated with mortality.
CKD and mortality
Cardiovascular events are 3.5 to 100 times greater in dialysis-dependent CKD patients than the general population (6,7). Recently, we and others (2,8) demonstrated that mild CKD is associated with a doubling of mortality. The degree of CKD appears to be directly proportional to the risk for cardiovascular events and mortality. The current study demonstrates that the risk of death at nine months in the lowest CrCl group was 2.7 times greater than the highest CrCl group, and there was an intermediate risk of 1.5 in the middle CrCl group.
Possible explanations for the increased mortality associated with mild CKD include increased classic cardiovascular risk factors in these patients, accelerated atherogenesis, and increased complications from cardiovascular disease when it develops. Not surprisingly, CKD patients have increased age, diabetes, and hypertension compared with those with normal renal function (2,8). In PRESTO, patients with CKD were older, with a greater frequency of diabetes and hypertension. The lowest CrCl group also had more prior PCI and CABG surgery. Remarkably, despite this increased risk profile of the CKD population, there was lower use of statin therapy. Moreover, the lowest CrCl group had lower complete revascularization than the highest CrCl group (22% vs. 30%), which is another risk for subsequent cardiovascular events and mortality. Unlike previous studies, in the PRESTO trial, CKD was no longer independently associated with mortality after adjusting for the other variables. This difference may be due to the shorter follow-up of nine months in our study compared with previous studies or greater adjustment for covariates in our study. Because CKD is associated with a lower frequency of success and a greater complication rate during PCI, the difference in our study may in part be due to the fact that patients were enrolled in the PRESTO trial only if the PCI procedure had been successful. Moreover, because of the homogenous nature of trial participants owing to strict inclusion and exclusion criteria, these patients may differ from the general population.
CKD and MIs
Nearly one-half of the deaths in the dialysis-dependent CKD population are caused by cardiovascular disease (9). Patients with CKD not only have a much greater frequency of MIs, but also have a higher mortality from MIs (10,11). Among those receiving dialysis, the mortality is nearly 60% in the first year after a first MI. In-hospital mortality after a MI is 2% with normal renal function, 6% with mild CKD, 14% with moderate CKD, 21% with severe CKD, and 30% in dialysis patients (11). Despite the high risk of death in CKD patients with a MI, proven therapies for treatment, including reperfusion therapy, beta-blockers, and aspirin, are all underutilized (11,12).
In the PRESTO trial, there was no such association between CKD and MI in the nine months after successful PCI. Myocardial infarction frequency was only 1.5% in the lowest CrCl group, compared with 1.3% in the highest CrCl group. The most likely explanation for the difference in our study compared with previous studies is a shorter follow-up period. However, other factors such as less multivessel disease, the exclusion of patients with a creatinine ≥1.8 mg/dl, and the requirement for enrollment after a successful PCI procedure may have contributed.
CKD and restenosis
Enthusiasm for performing PCI in patients with severe CKD has been limited by the results of studies with balloon angioplasty alone, which demonstrated significantly lower procedural success rates and markedly higher restenosis rates, as high as 81% (13–15). As in the general population, stenting improves procedural success and reduces clinical cardiovascular events in end-stage CKD patients by one-half at one year (16). Conflicting results exist regarding the frequency of restenosis in CKD patients receiving stents; some studies demonstrated that TVR was not higher in this population (17), whereas others still found a doubling of the TVR in CKD patients (35% vs. 16%, p < 0.05) (18). However, repeat angiography was not performed after PCI in any of these studies.
Understanding the impact of mild CKD on restenosis is important because mild CKD is more than 10 times more prevalent than end-stage CKD and affects approximately 8% of the population (19). Previous database studies showed that mild CKD is not associated with increased TVR (2,20). Because these studies did not prospectively perform follow-up angiography, it is unknown whether silent restenosis is increased, accounting for the higher mortality in this population, as seen in diabetic patients. Our study demonstrates not only that there was no increased risk of TVR associated with mild or moderate CKD, but also demonstrates for the first time that angiographic restenosis is also not increased. Thus, in the stenting era, PCI appears to have the same favorable results in this population as in those with normal renal function.
The Cockroft-Gault formula was used to estimate CrCl, which is a reliable but surrogate assessment of renal function. The mild inaccuracies of this method introduced into the study should not significantly alter the interaction of renal function and restenosis, in part because of the large patient population studied. Additionally, because of the exclusion criteria of the PRESTO trial, only patients with a creatinine <1.8 mg/dl were included in the study. Thus, few patients with severe CKD were evaluated and the results of this study cannot be applied to patients with more severe CKD. Lastly and perhaps most importantly, this was a study of patients in whom PCI was successful, not of patients undergoing PCI. Because success rates are lower and complications are higher in CKD, caution must be taken when comparing these results with those from other studies.
After successful PCI, mortality is increased in patients with mild or moderate CKD over a nine-month follow-up period, although the association is no longer significant after adjustment for other risk factors known to be associated with mortality. There is no increased risk of MI or TVR among CKD patients at nine months. Furthermore, angiographic restenosis was not increased in CKD patients, demonstrating that the increased mortality among CKD patients is not due to silent restenosis. These data suggest that PCI is associated with similar outcomes in patients with mild or moderate CKD as in those without CKD.
This study was supported by grants from GlaxoSmithKline Pharmaceuticals.
- Abbreviations and acronyms
- coronary artery bypass graft
- congestive heart failure
- chronic kidney disease
- creatinine clearance
- hazard ratio
- myocardial infarction
- percutaneous coronary intervention
- Prevention of REStenosis with Tranilast and its Outcomes trial
- target vessel revascularization
- Received February 20, 2004.
- Revision received July 6, 2004.
- Accepted July 12, 2004.
- American College of Cardiology Foundation
- Rubenstein M.H.,
- Harrell L.C.,
- Sheynberg B.V.,
- Schunkert H.,
- Bazari H.,
- Palacios I.F.
- Best P.J.,
- Lennon R.,
- Ting H.H,
- et al.
- Holmes D.R. Jr..,
- Savage M.,
- LaBlanche J.M,
- et al.
- Szczech L.A.,
- Best P.J.,
- Crowley E,
- et al.
- ↵Anonymous. Excerpts from the United States Renal Data System 1998 annual data report: V. Patient mortality and survival. Am J Kidney Dis 1999;32 Suppl 1:S69–80.
- Schoebel F.C.,
- Gradaus F.,
- Ivens K,
- et al.
- Le Feuvre C.,
- Dambrin G.,
- Helft G,
- et al.