E data have suggested that renal dysfunction is an important predictor of outcome in patients with cardiovascular disease, including acute myocardial infarction (AMI).1–9 Among patients with AMI, those with end-stage renal disease on dialysis are known to have a particularly poor long-term survival.6 Although these patients have usually been excluded from trials of reperfusion therapy, many patients with AMI have evidence of mild or moderate renal insufficiency due to diabetic or hypertensive kidney disease. The degree of renal impairment is often under-recognized, because most clinicians rely on serum creatinine alone as a measure of renal function. Recent studies have shown that even mild or moderate renal impairment is associated with adverse outcomes in patients with hypertension or heart failure, but few data exist regarding this relation in patients with AMI.2–5,8,9 This study investigates the effect of baseline renal function on early and late clinical outcome after primary angioplasty for AMI. • • • Nine hundred patients with AMI presenting within 12 hours of symptom onset were randomized to primary angioplasty or stenting in the Stent Primary Angioplasty in Myocardial Infarction (Stent-PAMI) study.10 All patients had either ST-segment elevation of 1 mm in 2 contiguous electrocardiographic leads or a nondiagnostic electrocardiogram (left bundle branch block or posterior AMI) with documentation of AMI in the catheterization laboratory (with a high-grade stenosis and associated left ventricular wall motion abnormality). Patients with known renal impairment (serum creatinine 2.7 mg/dl) or cardiogenic shock were not eligible for enrollment. Other exclusion criteria have been previously described.10 The catheterization and study procedure have been reported in detail.10 In brief, all patients received aspirin, ticlopidine, and heparin before catheterization. Diagnostic coronary angiography was performed with a low osmolar ionic contrast medium (Ioxaglate, Mallinckrodt, St. Louis, Missouri). After initial balloon inflation, patients were randomized to implantation of a heparin-coated stent or angioplasty alone. The infarct vessel had to be a native vessel with a reference diameter of 3.0 to 4.5 mm and with 1 lesions that could be covered with 1 or 2 stents of 15 mm in length. Administration of glycoprotein receptor antagonists was discouraged. Postprocedural heparin was recommended after primary angioplasty alone. Clinical follow-up was obtained at 6 and 12 months. The baseline serum creatinine on admission was available for 847 patients. Creatinine clearance (CrCl) was calculated using the Cockcroft-Gault formula: CrCl ([140 age] weight)/(72 sCr) , where weight is measured as kilograms and serum creatine (sCr) is measured as milligrams per deciliter.11 CrCl was corrected by a factor of 0.85 for women to account for differences in lean body mass. Major adverse cardiac events (MACEs) were defined as death, nonfatal reinfarction, disabling stroke, and ischemia-driven target vessel revascularization. Reinfarction was defined as the recurrence of clinical symptoms or new electrocardiographic changes accompanied by an increase in creatine kinase-MB levels. Disabling stroke was defined as a stroke that resulted in severe limitation in the ability to perform daily activities or the inability to live independently. An independent clinical events committee adjudicated all events. Data analysis was performed using SAS software (version 6.18, SAS Institute, Cary, North Carolina). Results are expressed as mean SD or percentages. Categorical variables were examined using a Pearson chi-square test or Fisher’s 2-sided exact test. Continuous variables were examined using a Student’s t test. Univariate analysis was performed using serum creatinine and CrCl as continuous and dichotomous variables. For the latter, a serum creatinine of 1.5 mg/dl and CrCl of 75 ml/min were chosen to provide a clinically useful index of renal dysfunction. Multivariate predictors of in-hospital, 6-, and 12-month mortality were identified using stepwise logistic regression analysis. Because there is a strong correlation between age, gender, and the calculated CrCl, we performed models with and without these specific variables. A Kaplan-Meier curve was calculated for freedom from death according to baseline CrCl. A p value of 0.05 was considered statistically significant. From William Beaumont Hospital, Royal Oak, Michigan; Lennox Hill Heart and Vascular Institute, New York, New York; Le Bauer Health Care, Greensboro, North Carolina; Mid Carolina Cardiology, Charlotte, North Carolina; Hospital Gregorio Maranon, Madrid, Spain; Instituto Dante Pazzanese de Cardiologia, San Paulo, Brazil; and L’Institut Cardiovasculaire Paris Sud, Antony, France. Dr. C. Grines’ address is: Division of Cardiology, William Beaumont Hospital, 3601 West 13 Mile Road, Royal Oak, Michigan 48073. E-mail: cgrines@ beaumont.edu. Manuscript received December 27, 2002; revised manuscript received and accepted February 27, 2003.