A comparison of bolus versus continuous cardiac output in an experimental model of heart failure

The majority of studies examining cardiac output measurement have been done in physiologically stable models with low thermal background noise. Research comparing continuous cardiac output (CCO) with bolus thermodilution cardiac output (COTD) measures in human and animal models have reported high correlations, negligible bias, but large limits of agreement. The purpose of this project was to compare CCO with COTD measures in an experimental model of heart failure where the cardiac output values were low and the range was narrow.

A one-group experimental design with preintervention control measures and repeated CCO and COTD measures across nine time points.

Cardiovascular research laboratory.

Thirty dogs.

Univariate and multivariate versions of repeated-measures analysis were used to assess the influences of temperature, weight, and stage of the experimental protocol on CCO, COTD, and the differences between them. The two measures CCO and COTD were assessed for agreement by using methods proposed by Bland and Altman. Two hundred and fifty pairs of measurements were obtained during sinus rhythm. The range for COTD measures was 0.5-4.67 L/min and for CCO measures 1.0-5.40 L/min. Of the 250 cardiac outputs estimated by the continuous method, 73.4% of the values were within +/-15% of that estimated by the repetitive, single thermodilution method. The mean bias for the entire protocol was 0.01 (SD = 0.51) with a range of 4.33 L/min.

Agreement between the two measures may be the function of biological variability, responses to anesthesia, and technique. Bland and Altman evaluation demonstrated low bias and precision and similar levels of agreement when compared with previous studies in an experimental model where the cardiac output was low and the range was narrow.