Central venous blood temperature fluctuations and thermodilution signal processing in dogs

Abstract

Pulmonary artery blood temperature fluctuations are a principle source of variability of thermodilution cardiac output measurements. Two signal processing strategies were investigated to decrease this variability; a subtractive noise cancellation algorithm, developed from a heat balance model of heat transport through the right heart, and a band pass integration technique. Blood temperature fluctuations were recorded from the inferior and superior vena cavae and the pulmonary artery in four dogs. The recorded temperature signals were processed off line to investigate the performance of the signal processing strategies. The findings indicate that the accuracy of the heat transport model, while dependent on the accuracy of the sensing of caval blood temperature, is sufficient for cancellation of a significant fraction of the pulmonary artery thermal noise power if the amplitude of the fluctuations is large. The use of the subtractive cancellation algorithm improved the standard deviation of thermodilution curve area determinations by a factor of five, relative to area estimates made without noise cancellation. Band pass integration produced a more modest improvement in area estimate reproducibility. These results demonstrate that it is possible to reduce, but not eliminate, the variability of thermodilution cardiac output measurements with the described processing strategies.