The present report aims to generate a simple and efficient non-invasive prediction system of myocardial infarction as well as evaluating of the new generated score.
Methods: 208 patients (both men and women) aged 37-75 admitted at Centre Hospitalier Coulommiers, France, were included in the study. Patients were divided into two groups, according to presence (104 patients, 63.7 +/- 9.2 years) and absence of myocardial infarction (104 patients, 58.1 +/- 12.4 years).
Results: Based on univariate and multivariate analyses the following parameters were selected to form the prediction model: age, lipoprotein (a), C-reactive protein, systolic blood pressure, non-HDL-Cholesterol. A new score was obtained and it was validated by the area under the ROC curve of 0.745 +/- 0.034 (95% CI = 0.681-0.804), p < 0.0001. For the newly generated score, at the cut-off point (> -0.3), we obtained the following accuracy parameters: sensibility=80%, specificity=62.5%, positive predictive values=65%, negative predictive values =78%. The mentioned parameters were superior to values of each component of the score individually analyzed.
Conclusions: The newly generated system can become a useful tool in coronary risk evaluation, with direct application in clinical practice.