edicine 9 (2008) 218–223
Cardiovascular Revascularization MRenal impairment is an independent predictor of adverse events postcoronary intervention in patients with and without drug-eluting stents
James A. Shawa,b,⁎, Nick Andrianopoulosc, Stephen Duffya,b, Anthony S. Waltona,d,David Clarke, Robert Lewf, Martin Sebastiang, Gishel Newh, Angela Brennanc,
Chris Reidc, Andrew E. Ajanic,i,j
On behalf of the Melbourne Interventional GroupaDepartment of Cardiology of The Alfred Hospital, Melbourne, Australia
bBaker Heart Research Institute, Melbourne, AustraliacDepartment of Epidemiology and Preventive Medicine, NHMRC Center of Clinical Research Excellence in Therapeutics,
Monash University, Melbourne, AustraliadDepartment of Cardiology of The Western Hospital, Melbourne, AustraliaeDepartment of Cardiology of The Austin Hospital, Melbourne, Australia
fDepartment of Cardiology of The Frankston Hospital, Melbourne, AustraliagDepartment of Cardiology of The Geelong Hospital, Melbourne, AustraliahDepartment of Cardiology of The Box Hill Hospital, Melbourne, Australia
iDepartment of Cardiology of The Royal Melbourne Hospital, Melbourne, AustraliajUniversity of Melbourne, Melbourne, Australia
Received 21 January 2008; received in revised form 6 May 2008; accepted 8 May 2008
Abstract Renal impairment (RI) is known to be an independent risk factor for the progression of cardiovascular
⁎ Corrresponding9076 3321; fax: +61 3
E-mail address: ja
1553-8389/08/$ – seedoi:10.1016/j.carrev.2
disease. Its impact, however, on the outcomes in patients undergoing percutaneous coronaryintervention (PCI) especially in the era of drug-eluting stents (DES) is not well known. We analyseddata from patients undergoing PCI from April 1, 2004, to September 30, 2006, who were part of theMelbourne Interventional Group registry. RI was defined as an estimated glomerular filtration rate(eGFR), calculated using Cockcroft–Gault formula, of b60ml/min.We compared outcomes at 30 daysand 12 months in patients with and without RI. Four thousand one hundred ninety-five patients (3043male) with an average age 65±12 years (mean±S.D.) underwent PCI. Twelve-month follow-up wasavailable in 3963 (95%) patients, and these were included in the analysis. One thousand twelve patients(26%) had RI; of these, 608 (60%) presented with an acute coronary syndrome. Both 30-day majoradverse cardiac events (MACE), 9.1% vs. 4.6% (Pb.01), and all-cause mortality, 4.5% vs. 0.7%(Pb.01), were significantly higher in those with RI compared to those without RI. Twelve-monthmortality (8.8% vs. 1.7%, Pb.01) and MACE (19.7% vs. 10.3%, Pb.01) were also significantly higherin those with RI. In multiple regression analysis, RI was an independent predictor of 12-month MACE[OR 2.0 (CI 1.6–2.6), Pb.01]. RI is an independent predictor of 30-day and 12-month MACE anddeath after PCI in patients with stable and unstable coronary syndromes, even with widespread use ofDES. eGFR should be used to help risk-stratify patients undergoing PCI.© 2008 Elsevier Inc. All rights reserved.
Keywords: Renal impairment; Percutaneous coronary intervention; Drug eluting stents
author. Department of Cardiology Alfred Hospital and Baker Heart Research Institute, Commercial Rd, Melbourne, Australia. Tel.: +61 [email protected] (J.A. Shaw).
front matter © 2008 Elsevier Inc. All rights reserved.008.05.002
219J.A. Shaw et al. / Cardiovascular Revascularization Medicine 9 (2008) 218–223
1. Introduction
Renal impairment (RI) is known to be an important riskfactor in the development and progression of atheroscleroticdisease [1–4]. Previous studies have shown that there is anindependent graded association between estimated glomer-ular filtration rate (eGFR) and risk of cardiovascular events[2]. Furthermore, in patients post myocardial infarction (MI),even mild renal disease increases patients' risk of cardio-vascular complications [3].
With an ageing population, there is an increasedprevalence of patients with renal impairment [5]. If onedefines chronic kidney disease as an eGFR of b60 ml/minper 1.73 m2, there is an estimated 11 million Americans withrenal impairment [5]. Many of these patients will havecoronary artery disease, and treatment strategies such ascoronary bypass surgery and percutaneous coronary inter-vention can further impact on renal function.
The use of percutaneous therapies to treat coronaryartery disease is well established. Their prognosticbenefits in patients with acute MI [6] and unstable angina[7] are well described as are the symptomatic benefits inthose with stable angina. Since the advent of drug-elutingstents (DES), there has been a significant reduction in theneed for target lesion revascularisation [8,9], and despitesome concerns about the issue of late stent thrombosis[10], DES are used in the majority of coronaryinterventions. The importance of RI on outcomes in theera of DES is unknown.
The aim of this study was to examine 30-day and 12-month outcomes in patients from a large multi centre registryand determine whether renal impairment, as defined as aneGFR b60ml/min, predicted adverse events in patientsundergoing percutaneous revascularisation.
2. Methods
All patients undergoing percutaneous coronary interven-tion (PCI) from April 2004 until September 2006 who wereenrolled in the Melbourne Interventional Group (MIG)registry were included in the study. The MIG registry is avoluntary collaborative venture between the interventionalcardiologists at seven government hospitals in Victoria,Australia [11,12]. All patients who undergo PCI atparticipating hospitals are included in the registry includingelective, urgent, and emergent cases. As previouslydescribed clinical, demographic, and procedural data arecollected for all patients on a standard case report form withstandardised definitions for all fields [12]. The case reportforms are faxed to a central office and electronicallyuploaded. The registry is coordinated by the Centre forClinical Research Excellence in Therapeutics a researchbody within the department of Epidemiology and PreventiveMedicine Monash University, Melbourne, Australia. Aquery system has been developed to identify missing data,
data inconsistencies, and out-of-range values. The databaseis built on a Microsoft SQL server platform with a Microsoftaccess user interface. An independent audit is conducted on3% of all patients to ensure data accuracy.
This study's primary aim was to compare the outcomes inpatients with and without RI, this being defined as acreatinine clearance of b60 ml/min per 1.73 m2. CrCl wascalculated using the Cockcroft Gault formula:
For males:weight ðkgÞ � ð140−ageÞ72� Se creatinine ðmg=dlÞ
For females:weight ðkgÞ � ð140−ageÞ72� Se creatinine ðmg=dlÞ � 0:85
In Australia, the creatinine levels are measured in mmol/l;this is converted to mg/dl, as shown:
mmol=l� 1000 ¼ lmol=l
Se Crðlmol=lÞ88:4
¼ Se Cr ðmg=dlÞ
2.1. Subjects and clinical and procedural variables
The case report forms for patients undergoing PCI includeinformation about the participants' coexisting illnesses aswell as cardiac risk factors and procedural informationincluding type and size of stent, lesion type, and adjunctivepharmacotherapy. Final stent size is used as a surrogate forreference vessel diameter; lesion morphology was classifiedaccording to the American College of Cardiology/AmericanHeart Association (ACC/AHA) classification task force withB lesions further classified into B1 and B2 lesions [13].Inhospital follow-up was completed by the treating physicianor trained cardiac nurse. At 30 days and 12 months postprocedure, follow-up was performed by telephone contactfor all patients by a trained cardiac nurse. Where contactdirectly with the patient could not be made, attempts weremade to contact the treating primary care physician and/orcardiologist, and matches were made with the national deathindex to determine if the patients were still alive. Duringtelephone contact, information was obtained about currentmedications, hospital admissions, and major adverse cardiacevents (MACE), which included MI, readmission to hospitalfor repeat cardiac procedures including percutaneouscoronary intervention and/or coronary artery bypass surgery.Patients receiving repeat PCI were recaptured in the databaseunless they underwent treatment in a hospital which was notpart of the registry. When patients were readmitted tohospital, their records were reviewed to confirm thediagnosis and ensuing management. The registry wasapproved by the ethics committee at the seven Victorianhospitals; opt-out informed consent was obtained from allparticipants as previously described [12].
Table 2Clinical presentation and procedural characteristics
RI (n = 1012) No RI (n = 2951) P
ACS 608 (60) 1800 (61) nsIIb/IIIa 220 (21) 840 (28) b .01DES 549 (54) 1471 (50) b .01Reference vessel diameter (mm) 2.9±0.5 3.0±0.5 b .01Stent length (mm) 18.3±8.4 18.7±8.5 nsLesion Type A 152 (12) 474 (13) nsLesion Type B1 436 (35) 1358 (38) nsLesion Type B2 433 (35) 1236 (34) nsLesion Type C 203 (17) 510 (14) ns
220 J.A. Shaw et al. / Cardiovascular Revascularization Medicine 9 (2008) 218–223
2.2. Statistical analysis
All statistical analyses were performed using SPSS forWindows, Version 14.0 (SPSS, Chicago, IL, USA).Continuous variable are expressed as mean±S.D.; categoricalvariables, as percentage. Student's t test and Pearson chi-squared analysis were carried out for comparison ofcontinuous and categorical variables respectively. A valueof Pb.05 was considered significant. Univariate predictors atPb.100 have been entered into the multivariate modellooking at 30-day and 12-month MACE as outcomes.
Chronic occlusion 9 (1) 25 (1) ns
ACS, acute coronary syndrome.
3. Results3.1. Baseline characteristics
A total of 4195 patients were included in the registryduring the specified timeline, i.e., between April 2004 andSeptember 2006. Of these, 12-month follow-up was availablein 3963 patients, and this is the cohort from whom thesubsequent data has been obtained. There were 1012 patientswith RI, as defined as a CrCl of b60 ml/min per 1.73 m2 andthe baseline characteristics of the two groups (those with andwithout RI) are shown (Table 1). Of note, the average age inthose with RI was significantly higher than in those without RI(75.5±8.1 vs. 61.2±10.9 years, Pb.01). In the entire cohort,2888 (73%) were male, while in those with RI, only 54%weremale, compared to 79% in those without RI (Pb.001). Asexpected, there was a higher incidence of diabetes amongstthose with RI (29% vs. 21% Pb.01) and hypertension (75%vs. 58% Pb.01), and more patients with RI had a history ofcerebrovascular and peripheral vascular disease.
3.2. Procedural characteristics
A similar percentage of patients in the 2 groups presentedwith acute coronary syndromes. There was a higher use ofDES in the patients with RI, although with less IIb/IIIainhibitor use, compared to those without RI. The reference
Table 1Baseline clinical characteristics in patients with and without RI
RI (CrCl b60 ml/min)[no. (%), n=1012]
No RI (CrCl N60 ml/min)[number (%), n=2951] P
Age 75.5±8.1 61.2±10.8 b .01Male 544 (54) 2344 (79) b .01Diabetes 290 (29) 632 (21) b .01↑ Cholesterol 725 (68) 2247 (72) b .05Smokers 84 (9) 775 (27) b .01Family Hx 279 (30) 1195 (42) b .01HT 756 (75) 1703 (58) b .01Previous MI 351 (35) 799 (27) b .01Hx of CVD 83 (8) 128 (4) b .01Hx of PVD 115 (11) 150 (5) b .01
Hx, history; CVD, cerebrovascular disease; PVD, peripheral vasculardisease.
vessel diameter was smaller 2.9 vs. 3.0 mm (Pb.05) in thepatients with RI, but there was no difference in stent lengthor in ACC/AHA lesion morphology classification betweenthose with and without RI (see Table 2).
3.3. 30 Day results
Thirty-day follow up was available in 3963 patients.MACE (including death, MI, and target lesion revascular-isation) was nearly twofold higher in the patients with RI9.1% vs. 4.6% (Pb.001). All-cause mortality was alsosignificantly higher (sixfold) in the RI group 4.5% vs. 0.7%(Pb.001). Patients with an eGFR of b30 ml/min, comparedto those with a eGFR 30–60 ml/min, had a higher mortality(12.1% vs. 3.3%, Pb.05) and higher rate of MI (7.1% vs.2.9%, Pb.05). In a multiple logistic regression model takinginto account multiple variables including age, use of DES,and diabetes, predictors of MACE included RI (OR 1.6, CI:1.2–2.3, Pb.01), presence of diabetes (OR 1.5, CI: 1.1–2.0),and using DES (OR 0.6. CI: 0.5–0.9, Pb.01) (see Table 3).
In the RI group at 30 days, there was a higher incidence ofdeath (6.0% vs. 2.9%, Pb.05) and MACE (10.3% vs. 6.2%,Pb.05) in the patients who received Bare metal stent(BMS) vs. DES.
3.4. 12 Month results
Twelve-month follow-up was available in 3963 patients.MACE (19.7% vs. 10.3%, Pb.01) and mortality (8.8% vs.1.7, Pb.01) were significantly higher in those patients withRI. When the RI patients were analysed separately, was no
Table 3Multiple logistic regression predictors of 30 day MACE
Variable OR 95% CI P Value
Use IIb/IIIa 0.6 0.4-0.8 b .01Diabetes 1.5 1.1-2.0 b .05Small vessel 1.6 1.2-2.1 b .01Age 1.0 1.0-1.0 nsUse of DES 0.6 0.5-0.9 b .01Presents ACS 1.2 0.9-1.7 nsCrCl 1.6 1.2-2.3 b .01
Table 4Multiple logistic regression predictors of 12 month MACE
Variable OR 95% CI P Value
Heart failure 1.8 1.2-2.8 b .01Diabetes 1.5 1.2-1.9 b .05Small vessel 1.5 1.2-1.8 b .01Age 1.0 1.0-1.0 nsUse of DES 0.6 0.5-0.7 b .01Presents ACS 1.2 0.9-1.5 nsCrCl 2.0 1.6-2.6 b .01
Fig. 2. Kaplan–Meier Curve showing time to death in three groups. Therewas a significant difference in events between three groups (Pb.01, logrank). The pattern seen is very similar to that in Fig. 1, with the b30-ml/mingroup doing worst and the 30–60-ml/min group doing slightly worse thanthe N60-ml/min group.
221J.A. Shaw et al. / Cardiovascular Revascularization Medicine 9 (2008) 218–223
difference in 12-month MACE (20.4% vs. 16.9% P=ns) ormortality (9.8% vs. 7.5% P=ns) in those who received BMSvs. DES.
There was no difference in target vessel revascularisation(TVR) between those with and without RI (7.8% vs. 6.6%,P=ns). In the logistic regression model, predictors of 12-month MACE included diabetes (OR 1.5, CI: 1.2–1.9,Pb.05), using DES (OR 0.6, CI: 0.5–0.7, Pb.01), vesseldiameter ≤2.5 mm (OR 1.5, CI: 1.2–1.8, Pb.05), knownheart failure (OR 1.8, CI: 1.2–1.8, Pb.01) and creatinineclearance (OR 2.0, CI: 1.6–2.6, Pb.01) (Table 4).
When the DES group were examined separately at 12months, RI was a predictor for MACE (OR 2.1, CI: 1.4–3.0,Pb.01). Other significant predictors in the DES group at 12months were:
Presence of cerebrovascular disease (OR 2.6, CI: 1.6–4.4,Pb.01)Use of IIb/IIIa blockers (OR 1.8, CI: 1.3–2.6, Pb.01)Reference vessel diameter ≤2.5 mm (OR 1.7, CI: 1.2–2.2, Pb.01)
Kaplan–Meier curves for freedom from MACE and deathshowed significantly less events in those with CrCl N60 ml/
Fig. 1. Kaplan–Meier Curve showing time to MACE in three groups asdefined by their estimated GFR b30, 30-60, and N60. There was asignificant difference in events between three groups (Pb.01, log rank). Thecurves for each group show a consistent pattern across the 12 months, wherethe 30–60-ml/min curve lies slightly beneath the N60-ml/min curve, whilethe b30-ml/min curve starts increasing its distance beneath the 30–60-ml/min curve from approximately 30 days onwards.
min compared to those whose CrCl was in the 30–60-ml/minrange and those b30 ml/min (see Figs. 1 and 2).
4. Discussion
In this multicenter PCI registry, RI was an independentpredictor of MACE at 30 days and 12 months. Interestingly,there was no difference in TVR between the two groupsdespite a higher use of DES in the RI group. This study, toour knowledge, is the first large study showing theimportance of RI as a risk factor for adverse events inpatients undergoing PCI in the current era of DES.
Numerous studies have shown the importance of renalimpairment, defined as a reduced glomerular filtration rate,in predicting both death and cardiovascular events [2,3]. Inpatients post MI, reduced eGFR was seen to be a majorrisk factor for cardiovascular complications with an inverserelationship between GFR and rate of cardiovascularevents [3].
Previous studies have looked at the effects of impairedrenal function on patients undergoing PCI. Rubinstein et al.[14] compared immediate and long-term outcomes inpatients undergoing percutaneous revascularisation withand without renal impairment. In that study, RI was definedas serum creatinine N1.5 mg/dl, and patients with RI hadreduced procedural success and increased MACE. A majorlimitation of the study was that it does not reflect currentpractice, with only 32% of patients receiving stents, andatherectomy was used more frequently than in currentpractice [14]. Further analysis on the same patient populationshowed no difference in outcomes between those with mildRI and those on dialysis [15]. The use of serum creatinine asa marker of RI has limitations because of its nonlinearassociation with GFR that varies according to age, sex, race,and lean body mass [16,17]. Thus, eGFR, which takes into
222 J.A. Shaw et al. / Cardiovascular Revascularization Medicine 9 (2008) 218–223
consideration patient age, body weight, and serum creati-nine, is more sensitive for detecting patients with RI. Thus,in our study, 26% of patients had RI compared to 12% in thestudy by Rubinstein et al. [14]. Other studies have shownsimilar results with the definition of renal impairment beingan elevated serum creatinine. In one study, the patients withrenal impairment (defined as those with an elevatedcreatinine in the setting of a history of renal failure treatedwith medications, dialysis, or a low-protein diet) hadincreased in hospital and 1-year mortality [18]. Inthe registry used in that study, only 192 (4.2%) of the 4602patients had renal impairment, significantly less than inour study.
The pathophysiology underlying the increased risk ofcardiovascular events in patients with RI is not entirelyunderstood. There are, however, numerous changes, whichoccur in patients with renal impairment, which couldpotentially increase their cardiovascular risk. Reducedrenal function is associated with inflammation, anaemia,oxidative stress, derangements in calcium-phosphate home-ostasis with an increase in arterial calcification, andconditions which promote coagulation [1,5]. It has beenpreviously shown that inflammation, as measured withhighly sensitive C-reactive protein, is an independentpredictor of a worse prognosis in patients undergoing PCI[19]. The other physiological changes that occur with RI areassociated with accelerated atherosclerosis and endothelialdysfunction, a precipitant of the development and progres-sion of atherosclerosis. With a decline in renal function, thereis an increase in albuminuria, proteinuria, homocysteinemia,and uric acid levels. These factors may also contribute to theincreased risk of events in patients with renal impairmentundergoing PCI. In our cohort, the patients with RI wereolder, had more diabetes and hypertension, and had a higherincidence of previous MI and cerebrovascular and peripheralvascular disease.
Interestingly, there was no difference in lesion morphol-ogy between those with and without RI, and the slightdifferences in reference vessel diameter and stent length wereunlikely to account for the large differences in outcomes seenbetween the two groups. Thus, it appears that it is not thelesion and the complexity of it per se but rather the systemicstate of the patient with RI which increases their risk ofcomplications after percutaneous coronary interventions.
The current registry commenced when DES wereavailable in the public health system in Australia. However,due to the cost constraints, DES were only available inspecific patient groups, including those with small vessels,i.e., (diameter ≤2.5 mm), long lesions (≥20 mm), diabeticsubjects, restenotic lesions, chronic total occlusions, bifurca-tion lesions, and ostial lesions. DES were used in 54% of thepatients with RI, which was higher than in the non-RI groupwhere 50% of the subjects received DES. Although there hasbeen concern regarding the long-term safety of DES with anincreased incidence of late stent thrombosis, the slightlyhigher use of DES in the RI group would be highly unlikely
to account for the increased MACE and mortality seen in theRI group. The differences were already evident at 30 days,and the magnitude of the increased risk in the RI group wasfar greater than the small difference in DES use. Interest-ingly, there was no difference in TVR between the 2 groupsdespite the higher use of DES in the renal group; however,the TVR was low being approximately 8% in the RI groupand 7% in those without RI. The selective use explains whyTVR rates are similar, i.e., the higher-risk lesions got DES.
5. Limitations
This study has certain limitations. Firstly, it has theinherent limitations of a registry as compared to arandomised trial. While the data in this registry was collectedprospectively, it is analysed retrospectively. Follow-up in thepatients is also an issue; however, of the 4195 patients, wehad 12-month follow-up available in 3963 patients (94% oforiginal cohort), which we believe to be an excellent follow-up rate and significantly reduces the risk of selection biasconfounding our results. Finally, it is possible that patientswith RI are more closely followed up especially in the earlypostprocedure period, and this may influence their outcomes.
6. Conclusions
This study is from a large prospective registry in thecurrent era of DES and showed that the incidence of renalimpairment, as defined as eGFR b60 ml/min calculatedusing the Cockcroft Gault equation, is high — 26% ofpatients. Furthermore, patients with RI have significantlyworse outcomes at both 30 days and at 12 months. Webelieve, therefore, that eGFR should be used to help riskstratify patients undergoing PCI.
Reference
[1] Luft FC. Renal disease as a risk factor for cardiovascular disease. Basicresearch in cardiology 2000;95(Suppl 1):I72–6.
[2] Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidneydisease and the risks of death, cardiovascular events, and hospitaliza-tion. N Engl J Med 2004;351(13):1296–305.
[3] Anavekar NS, McMurray JJ, Velazquez EJ, Solomon SD, Kober L,Rouleau JL, White HD, Nordlander R, Maggioni A, Dickstein K,Zelenkofske S, Leimberger JD, Califf RM, Pfeffer MA. Relationbetween renal dysfunction and cardiovascular outcomes after myo-cardial infarction. N Engl J Med 2004;351(13):1285–95.
[4] Foley RN, Parfrey PS, Sarnak MJ. Epidemiology of cardiovasculardisease in chronic renal disease. J Am Soc Nephrol 1998;9(12 Suppl):S16–S23.
[5] McCullough PA. Cardiorenal risk: an important clinical intersection.Rev Cardiovasc Med 2002;3(2):71–6.
[6] Keeley EC, Boura JA, Grines CL. Primary angioplasty versusintravenous thrombolytic therapy for acute myocardial infarction: aquantitative review of 23 randomised trials. Lancet 2003;361(9351):13–20.
[7] Cannon CP, Weintraub WS, Demopoulos LA, Vicari R, Frey MJ,Lakkis N, Neumann FJ, Robertson DH, DeLucca PT, DiBattiste PM,
223J.A. Shaw et al. / Cardiovascular Revascularization Medicine 9 (2008) 218–223
Gibson CM, Braunwald E. Comparison of early invasive andconservative strategies in patients with unstable coronary syndromestreated with the glycoprotein IIb/IIIa inhibitor tirofiban. N Engl J Med2001;344(25):1879–87.
[8] Moses JW, Leon MB, Popma JJ, Fitzgerald PJ, Holmes DR,O'Shaughnessy C, Caputo RP, Kereiakes DJ, Williams DO, TeirsteinPS, Jaeger JL, Kuntz RE. Sirolimus-eluting stents versus standardstents in patients with stenosis in a native coronary artery. N Engl JMed 2003;349(14):1315–23.
[9] Stone GW, Ellis SG, Cox DA, Hermiller J, O'Shaughnessy C, MannJT, Turco M, Caputo R, Bergin P, Greenberg J, Popma JJ, Russell ME.A polymer-based, paclitaxel-eluting stent in patients with coronaryartery disease. N Engl J Med 2004;350(3):221–31.
[10] Mauri L, Hsieh WH, Massaro JM, Ho KK, D'Agostino R, Cutlip DE.Stent thrombosis in randomized clinical trials of drug-eluting stents. NEngl J Med 2007;356(10):1020–9.
[11] Yan BP, Ajani AE, Duffy SJ, New G, Horrigan M, Szto G, Walton A,Eccleston D, Lefkovits J, Black A, Sebastian M, Brennan AL, ReidCM, Clark DJ. Use of drug-eluting stents in Victorian public hospitals.Med J Aust 2006;185(7):363–7.
[12] Ajani AE, Szto G, Duffy SJ, Eccleston D, Clark DJ, Lefkovits J, ChewDP, Warren R, Black A, New G, Walton A, Lew R, Shaw J, HorriganM, Sebastian M, Yan BP, Brennan A, Meehan A, Reid C, Krum H. Thefoundation and launch of the Melbourne Interventional Group: acollaborative interventional cardiology project. Heart Lung Circ2006;15(1):44–7.
[13] Ellis SG, Vandormael MG, Cowley MJ, DiSciascio G, Deligonul U,Topol EJ, Bulle TM. Coronary morphologic and clinical determinantsof procedural outcome with angioplasty for multivessel coronarydisease. Implications for patient selection. Multivessel AngioplastyPrognosis Study Group. Circulation 1990;82(4):1193–202.
[14] Rubenstein MH, Harrell LC, Sheynberg BV, Schunkert H, Bazari H,Palacios IF. Are patients with renal failure good candidates forpercutaneous coronary revascularization in the new device era?Circulation 2000;102(24):2966–72.
[15] Rubenstein MH, Sheynberg BV, Harrell LC, Schunkert H, Bazari H,Palacios IF. Effectiveness of and adverse events after percutaneouscoronary intervention in patients with mild versus severe renal failure.Am J Cardiol 2001;87(7):856–60.
[16] Walsh CR, O'Donnell CJ, Camargo CA, Giugliano RP, Lloyd-JonesDM. Elevated serum creatinine is associated with 1-year mortality afteracute myocardial infarction. Am Heart J 2002;144(6):1003–11.
[17] Shlipak MG. Pharmacotherapy for heart failure in patients with renalinsufficiency. Ann Intern Med 2003;138(11):917–24.
[18] Naidu SS, Selzer F, Jacobs A, Faxon D, Marks DS, Johnston J, DetreK, Wilensky RL. Renal insufficiency is an independent predictor ofmortality after percutaneous coronary intervention. Am J Cardiol2003;92(10):1160–4.
[19] Chew DP, Bhatt DL, Robbins MA, Penn MS, Schneider JP, Lauer MS,Topol EJ, Ellis SG. Incremental prognostic value of elevated baselineC-reactive protein among established markers of risk in percutaneouscoronary intervention. Circulation 2001;104(9):992–7.
