Heart failure (HF) has been classified as systolic and diastolic based on the left ventricular ejection fraction. We hypothesized that left ventricular diastolic dysfunction is an important element of HF regardless of ejection fraction. Two hundred six patients who had clinical HF were compared with 72 age-matched controls. Diastolic dysfunction, as assessed by the mitral filling pattern and tissue Doppler imaging, was present in >90% of patients who had HF regardless of ejection fraction and was more frequent and severe than in age-matched controls (p <0.001). In patients who had HF, B-type natriuretic peptide correlated with diastolic dysfunction (r = 0.62, p <0.001) but not with ejection fraction or end-diastolic volume index (EDVI). The degree of diastolic dysfunction influenced survival rate (risk ratio 1.64, p <0.05), whereas ejection fraction and EDVI did not. Systolic function measured by systolic mitral annular velocity was decreased in patients who had HF and an ejection fraction </=0.40 (4.8 +/- 1.1 cm/s) and, to a lesser extent, an ejection fraction >/=0.50 (6.6 +/- 1.8 cm/s) compared with control subjects (8.0 +/- 2.1 cm/s, p <0.01). Patients who had HF and an ejection fraction >/=0.50 had an increased ratio of ventricular mass to EDVI. Patients who had HF and an ejection fraction </=0.40 had increased left ventricular EDVI. Thus, regardless of ejection fraction, patients who have HF also have diastolic dysfunction. Diastolic dysfunction is a better predictor of B-type natriuretic peptide levels and mortality than ejection fraction or left ventricular EDVI. In addition to diastolic dysfunction, HF with an ejection fraction >/=0.50 is associated with mild systolic dysfunction and an increased ratio of left ventricular mass to EDVI. In HF with an ejection fraction </=0.40, systolic dysfunction and left ventricular dilation accompany diastolic dysfunction.