Current methods of measuring cardiac output require the invasive insertion of a thermodilution catheter with its concomitant risks and complications. We examined the noninvasive method of transthoracic electrical bioimpedance (TEB) in comparison with thermodilution cardiac outputs in a sample of 44 critically ill patients with poor left ventricular function (left ventricular ejection fraction less than 30%) and with either ischemic or idiopathic dilated cardiomyopathy. Dyspnea, mitral regurgitation, tricuspid regurgitation, and difference between real and ideal weight had the most marked effects on the correlation between the two methods, with lesser influence by left ventricular ejection fraction, height, weight, hemoglobin, hematocrit, and aortic regurgitation. TEB and thermodilution cardiac outputs were correlated, at r = 0.51 (p less than 0.00009), but the low reliability and low percentage of TEB readings within 0.5 L/min of thermodilution cardiac outputs (31%) renders TEB inadequate for clinical measurement of cardiac outputs in this patient population.