An IBM compatible impedance cardiac output monitoring prototype system has been developed for use at the bedside on patients in the ICU, CCU, ER, Cath. Lab, and OR, etc. This impedance cardiographic (ICG) system, whose operation is completely technician-free, provides a continuous display with digital results and four channel color waveforms on an Enhanced Graphics Display screen. The software is written in C language with several special segments in assembly code where speed is essential. In this prototype system, a real-time algorithm was introduced to modify the ensemble averaging technique so that it averages nonperiodic signals such as: ECG, dZ/dT, delta Z, etc. Also, a real-time algorithm was developed to adaptively detect R spikes from conventional ECG signals. A signal preprocessor was developed to process signals digitally before any further work is done. This procedure reduces muscle noise, 60 Hz interference, and ventilatory movement. A special digital filter was designed to cope with the cases in which pacemakers are used. A special algorithm was also developed to further reduce the ventilation artifacts so that a period of apnea is unnecessary during the performance of the measurements. An anatomically specified electrode configuration has been defined enabling precise and reproducible positioning of the electrodes--hopefully leading to electrode standardization. At the present time, this prototype system has been compared with standard hand calculation and correlated with the clinical "gold standard," the Swan-Ganz thermodilution cardiac output. Using 144 sets of data from 10 healthy volunteers, 4 critically ill patients, and 8 healthy exercising volunteers, calculations of cardiac output were made using our system and the standard hand calculation of stroke volume, based upon Kubicek's equation; there was a relatively high and stable correlation: r = 0.93, p less than 0.005 (healthy); r = 0.94, p less than 0.002 (ill), r = 0.95, p less than 0.002 (exercise). From 20 patients at two different hospitals all with Swan-Ganz catheters in their hearts, 65 correlation studies between our system and the standard thermodilution technique were performed; the results were encouraging in terms of accuracy and consistency (r1 = 0.84, p less than 0.01, n = 10 CCU patients), and (r2 = 0.93, p less than 0.01, n = ICU patients). These results along with a growing body of data from other investigators indicate that this noninvasive and technician-free system for measuring cardiac output could have a significant role in patient care.