Non-invasive estimation of cardiac output in mechanically ventilated patients

Measurement of cardiac output provides an indication of ventricular function making the monitoring of this parameter an important component in the hemodynamic management of both critically ill patients and patients with suspected cardiovascular disease. The knowledge of cardiac output value also helps to guide therapy to maintain adequate tissue perfusion in the high-risk surgical patient. The gold standard in the assessment of cardiac output consists in measuring the blood flow directly in the aorta requiring a very invasive procedure. Lots of methods have been implemented in order to indirectly estimate this parameter including: invasive methods, semi-invasive and non-invasive ones. Two invasive methods are considered "practical gold standards": the thermodilution and the Fick method. In this work of thesis a thorough analysis of the accuracy of these two methods have been performed, showing their strengths and weaknesses. This thesis introduces and validates a new non-invasive method for the estimation of cardiac output in mechanically ventilated patients: a technological transfer to Cosmed s.r.l. is expected at the end of this study. The method is based on prolonged expiration, and relies on measurement of gas concentrations and flow rate. The sensors of a metabolic monitor (Quark RMR - Cosmed s.r.l., Italy) are employed with the purpose to integrate its functions and extend its use to the Intensive Care Unit. The monitor has been previously validated both in-vitro and in-vivo with the aim to test its functioning on mechanically ventilated patients and to identify the limits of accuracy of its measurements. The monitor's performance resulted unaltered by the presence of the mechanical ventilator into the breathing circuit, with or without bias flow. A pneumatic system, with an ad hoc designed orifice resistance, has been realized and experimentally characterized to adapt the ventilatory patient circuit to the estimation of cardiac output by prolonging expiration. The use of this system results safe and risks of volutrauma, barotrauma, hypoxia and hypercapnia can be excluded. Cardiac output is calculated by using two different elaboration algorithms employing data of gas concentration and flow rate, acquired during normal breathing and prolonged expiration. With the aim to quantify the agreement of the proposed method with the thermodilution, twenty mechanically ventilated patients, who have undergone cardiac surgery, have been enrolled. Good correlation with thermodilution is found for both algorithms (R >0.8). The application of the first algorithm gives mean cardiac output values slightly lower than those obtained by thermodilution (-6%), whilst the application of the second algorithm gives higher values (+30%). The standard deviations of the differences between paired measurements are: 0.72 L/min for the first algorithm and 1.07 L/min for the second one. Standard deviation obtained by the application of the first algorithm is slightly lower than those relative to other minimally invasive techniques. Prolonged expiration, standardization and automation of the procedure on mechanically ventilated patients all appear feasible with the proposed system, in order to obtain a non-invasive estimation of cardiac output.