We evaluated the analytical characteristics and clinical usefulness of a commercially available IRMA kit for measuring plasma concentrations of atrial natriuretic peptide (ANP) in healthy subjects and in patients with heart failure. The method uses two monoclonal antibodies prepared against sterically remote epitopes of the ANP molecule; the first antibody is coated on the solid-phase beads, and the second is radiolabeled with 125I. Fifty-nine healthy subjects and 77 patients with heart failure were studied. After subjects had rested 20 min in a recumbent position, blood samples were collected from a brachial vein into ice-chilled disposable polypropylene tubes containing aprotinin and EDTA. Plasma samples were immediately separated by centrifugation and stored at -20 degrees C until assay. The working range (CV <15%) was 10-2000 ng/L. The detection limit (2.13 +/- 0.91 ng/L) was similar to those reported for other IRMAs but was much better than those of RIAs. For healthy subjects, the results of this method (18.0 +/- 10.6 ng/L, range 4.7-63 ng/L, median 16.7 ng/L, n = 59) were similar to those generally reported for the most accurate methods, i.e., those using preliminary extraction and chromatographic purification of plasma samples. Measured plasma ANP was significantly associated with the severity of clinical symptoms, i.e., NYHA class (ANOVA, P <0.0001), and with the left ventricular ejection fraction (n = 62, r = 0.618, P <0.0001). Patients with severe heart failure showed greatly increased values (NYHA III-IV: 257.4 +/- 196.6 ng/L, n = 23).