Giuseppe Berton, MD, Rocco Cordiano, MD, Fiorella Cavuto, MD, Giulia Giacomini, PhD,Renzo De Toni, PhD, and Paolo Palatini, MD
Abstract
The long-term event-free survival (EFS) after acute myocardial infarction (AMI) is largelyuninvestigated. We analyzed noninvasive clinical variables in association with long-termEFS after AMI. The present prospective study included 504 consecutive patients with AMIat 3 hospitals from 1995 to 1998 (Adria, Bassano, Conegliano, and Padova Hospitals [ABC]study).
Thirty-seven variables were examined, including demographics, cardiovascular riskfactors, in-hospital characteristics, and blood components. The end point was 10-year EFS.Logistic and Cox regression models were used to identify the predictive factors. Wecompared 3 predictive models according to the goodness of fit and C-statistic analyses. Atenrollment, the median age was 67 years (interquartile range 58 to 75), 29% were women,38% had Killip class>1, and the median left ventricular ejection fraction was 51%(interquartile range 43% to 60%).
The 10-year EFS rate was 19%. Both logistic and Coxanalyses identified independent predictors, including young age (hazard ratio 1.2, 95%confidence interval 1.1 to 1.3, p=0.0006), no history of angina (hazard ratio 1.4, 95%confidence interval 1.1 to 1.8, p=0.009), no previous myocardial infarction (hazard ratio1.4, 95% confidence interval 1.1 to 1.7, p0.01), high estimated glomerular filtration rate(hazard ratio 0.8, 95% confidence interval 0.7 to 0.9, p=0.001), low albumin/creatinineexcretion ratio (hazard ratio 1.2, 95% confidence interval 1.1 to 1.3, p<0.0001), andhighleft ventricular ejection fraction (hazard ratio 0.8, 95% confidence interval 0.7 to 0.9, p=0.006).These variables had greater predictive power and improved the predictive power of 2 othermodels, including Framingham cardiovascular risk factors and the recognized predictors ofacute heart damage. In conclusion, 10-year EFS was strongly associated with 4 factors (ABCmodel) typically neglected in studies of AMI survival, including estimated glomerular filtrationrate, albumin/creatinine excretion ratio, a history of angina, and previous myocardial infarc-tion. This model had greater predictive power and improved the power of 2 other models usingtraditional cardiovascular risk factors and indicators of heart damage during AMI.