Cardiac dysfunction may be suspected in patients with cardiovascular disease but identifying those with the highest risk is problematic. B-type natriuretic peptide (BNP) is a strong marker of heart failure in un-treated patients. This study evaluates a combined BNP and clinical algorithm for detecting cardiac dysfunction and the risk of death, in patients receiving cardioactive medication.
Methods: 459 stable general practice patients, who were taking typical heart failure drugs for any indication, were included. Echocardiography, ECG, and assay of NT-proANP and BNP (two methods) were performed. Regression models were used to identify items in a Risk Score to detect cardiac dysfunction.
Results: A Risk Score based on history of myocardial infarction (1 point), abnormal ECG (2 points), atrial fibrillation (1 point) and raised BNP (1-2 points) detected cardiac dysfunction with an AUC of ROC of 0.85. A Risk Score > or = 2 had a sensitivity of 90%, specificity of 58%, and positive and negative predictive values of 37% and 96%. Risk Score and LVEF<0.36 also predicted mortality. Abnormal BNP defined as either >100 pg/ml (Shionogi), or as age and sex related values, had similar predictive value.
Conclusion: In patients on cardioactive medication, a structured Risk Score based on raised BNP, history of MI, atrial fibrillation and abnormal ECG was useful for identifying patients for immediate further examination and those who could be evaluated later.