To determine whether clinical parameters alone can differentiate normal versus decreased systolic left ventricular function in patients with heart failure. Detailed clinical data were collected prospectively from 225 consecutive patients who were hospitalized with heart failure. Findings in patients with normal (ejection fraction > or =45%) or decreased (ejection fraction <45%) left ventricular function were compared. Systolic function was normal in 104 patients (46%) and decreased in 121 patients (54%). Patients with normal function were older (mean [+/- SD] age, 59 +/- 13 years vs. 54 +/- 13 years, P = 0.007) and more likely to be female (56% vs. 35%, P = 0.001), obese (body mass index > or =30 kg/m(2), 62% vs. 48%, P = 0.04), have marked systolic (> or =160 mm Hg, 50% vs. 27%, P <0.001) and diastolic (> or =110 mm Hg, 25% vs. 13%, P = 0.02) hypertension, and use calcium antagonists (34% vs. 14%, P = 0.001). Patients with decreased function were more likely to use alcohol (37% vs. 20%, P = 0.007), angiotensin-converting enzyme (ACE) inhibitors (85% vs. 62%, P <0.001), and digoxin (57% vs. 27%, P <0.001); and more likely to have tachycardia (51% vs. 32%, P = 0.004), rales (89% vs. 80%, P = 0.05), electrocardiographic left ventricular hypertrophy (42% vs. 22%, P = 0.002), left atrial abnormality (52% vs. 22%, P <0.001), or flow cephalization on chest radiograph (91% vs. 79%, P = 0.02). Only sex, tachycardia, and use of digoxin and ACE inhibitors were associated with ventricular function in multivariable analysis. However, the sensitivity, specificity, and predictive values for all clinical variables were low. Differences in clinical parameters in heart failure patients with decreased versus normal systolic function cannot predict systolic function in these patients, supporting recommendations that heart failure patients should undergo specialized testing to measure ventricular function.