277

Abstract

Introduction: Pulmonary artery catheters and thermodilution (PAC-TD) remain the gold standard for hemodynamic monitoring, particularly after cardiac surgery. However, the invasive nature of PAC-TD, combined with its lack of clinical efficacy and declining use, calls for increased use of non-invasive hemodynamic monitoring techniques. Electrical cardiometry(EC), a form of thoracic electrical bioimpedance, provides completely non-invasive hemodynamic monitoring. This study evaluated the reliability of EC compared to PAC-TD for continuous hemodynamic monitoring in post-operative patients. Hypothesis: EC provides reliable continuous cardiac index (CI) monitoring compared to PAC-TD in post-operative patients. Methods: Post-operative patients with indwelling PACs (Continuous Cardiac Output catheters, Vigilance monitor, Edwards) underwent simultaneous, continuous hemodynamic monitoring with EC (Aesculon monitor, Cardiotronic). Comparisons were made between hourly PAC-TD and EC determinations of CI. EC CI values were averaged from one minute readings ±6 minutes centered on the hour. Bland-Altman analysis was performed for each individual patient’s data. Bias (mean difference between PAC-TD- and EC-derived CI) and percentage error (2SD of bias/mean CI) for each patient were calculated. Variance (coefficient of variation) over the entire monitoring period for each patient was calculated. Data are mean±SD. Results: 30 patients (cardiac surgery 29, vascular surgery 1) were monitored for 33 ± 19 hours. Bias was -0.7 ± 1 L/min/m2 and percentage error was 39 ± 12%. 23 (77%) of 30 patients had a percentage error < 40%. Coefficients of variation for PAC-TD and EC were 0.15 ± 0.04 and 0.14 ± 0.05, respectively (P = NS). Conclusions: EC monitoring of CI demonstrated good correlation with PAC-TD in post-operative patients. While a percentage error < 30% represents the ideal, clinically-acceptable limit of agreement between two hemodynamic monitoring techniques (Critchley and Critchley), many studies report values up to 40%. Our relatively small bias would generally be considered clinically acceptable and the low variance indicates reliable measurements. Our study supports EC as a reliable non-invasive method of continuous hemodynamic monitoring.