Improvement of mini-bolus thermodilution using an optimal subspace-based matrix filter

Cardiac output estimation by thermodilution is the cornerstone of hemodynamic monitoring in the intensive-care setting. However, the clinical utility of thermodilution has been hampered by the application of heat as the indicator. The major source of error in thermodilution is the pulmonary artery thermal noise arising from respiration, cardiac contractions and Mayer waves. To obtain an acceptable signal-to-noise ratio, improve precision and bias, 3-7 measurements using 10 cc iced or room-temperature injectate are averaged. This limits applicability of thermodilution to intermittent CO monitoring due to possible risks to the patient. The authors describe the application of a subspace decomposition based optimal signal enhancement method to mini-bolus thermodilution signals as small as 1 cc