BACKGROUND

Newly discovered circulating peptides, N-terminal pro-brain natriuretic peptide (N-BNP) and adrenomedullin (ADM), were examined for prediction of cardiac function and prognosis and compared with previously reported markers in 121 patients with myocardial infarction.

METHODS AND RESULTS

The association between radionuclide left ventricular ejection fraction (LVEF) and N-BNP at 2 to 4 days (r=-.63, P<.0001) and 3 to 5 months (r=-.58, P<.0001) after infarction was comparable to that for C-terminal BNP and far stronger than for ADM (r=-.26, P<.01), N-terminal atrial natriuretic peptide (N-ANP), C-terminal ANP, cGMP, or plasma catecholamine concentrations. For prediction of death over 24 months of follow-up, an early postinfarction N-BNP level > or = 160 pmol/L had sensitivity, specificity, positive predictive value, and negative predictive values of 91%, 72%, 39%, and 97%, respectively, and was superior to any other neurohormone measured and to LVEF. Only 1 of 21 deaths occurred in a patient with an N-BNP level below the group median (Kaplan-Meier survival analysis, P<.00001). For prediction of heart failure (left ventricular failure), plasma N-BNP > or = 145 pmol/L had sensitivity (85%) and negative predictive value (91%) comparable to the other cardiac peptides and was superior to ADM, plasma catecholamines, and LVEF. By multivariate analysis, N-BNP but not ADM provided predictive information for death and left ventricular failure independent of patient age, sex, LVEF, levels of other hormones, and previous history of heart failure, myocardial infarction, hypertension, or diabetes.

CONCLUSIONS

Plasma N-BNP measured 2 to 4 days after myocardial infarction independently predicted left ventricular function and 2-year survival. Stratification of patients into low- and high-risk groups can be facilitated by plasma N-BNP or BNP measurements, and one of these could reasonably be included in the routine clinical workup of patients after myocardial infarction.