The validity of cardiac output measurement by thoracic impedance: a meta-analysis.

A structured review of existing literature was undertaken to provide an objective evaluation of the validity of cardiac output measurement by thoracic impedance. Relevant studies were obtained from journals, advertising literature, and N.A.S.A. contract publications, using MEDLINE retrieval and references from other studies. Studies describing concurrent measurement of impedance cardiac output with other techniques were selected if they met certain technical and methodological criteria known to improve the validity of impedance cardiac output measurement. All studies provided Pearson's correlation coefficient as a measure of association. A weighted mean coefficient was calculated for selected studies using Fisher's Z transformation of the data. Correlation of impedance with thermodilution gave r = 0.82; with dye dilution, r = 0.83; with Fick technique, r = 0.80; and with radionuclear angiography, r = 0.65. Studies done on ICU patients showed a lower correlation than studies done on non-ICU patients (p = 0.003). A moderately good correlation exists between impedance cardiac output measurement and other techniques, although correlation is not so good when ICU patients are studied. Better techniques and methodology may improve this correlation.