The objective of our study was to evaluate impedance cardiography (IMP) as a noninvasive method to determine the left ventricular ejection fraction (LVEF). METHODS. A total of 24 patients, 8 men and 16 women, aged 45.0 +/- 12.9 years, participated in the study. They used cardiotoxic chemotherapeutic drugs or suffered from cardiac failure. LVEF was measured by means of IMP (LVEFimp) and radionuclide ventriculography (LVEFnuc). LVEFimp was calculated in three ways. Capan and colleagues [13] proposed a formula in which LVEF (LVEFCap) can be calculated from the systolic time intervals, namely, left ventricular ejection time and preejection time. Judy and colleagues [14] described a systolic (S) and a diastolic (D) part in the first derivative curve of the impedance signal. The ratio S/D might equal the LVEF (LVEFJud). A new LVEF calculation was introduced (LVEFimp) in this study based on the first derivative of the impedance signal, the thoracic impedance, and heart rate.Mean LVEFCap was 59.9 +/- 8.4%, which did not differ from LVEFnuc (59.9 +/- 7.1%). However the correlation between both methods was not significant (r = 0.29). Mean LVEFJud was 63.9 +/- 17.4%, which was not significantly different from LVEFnuc, with a fair correlation (r = 0.55). Mean LVEFimp was 59.2 +/- 9.4%, with a better correlation with radionuclide ventriculography (r = 0.75).The results of this study indicate that the equations that have been used until now can be improved. The new equation provides reliable LVEF values in this group of patients.