The ability to measure cardiac output accurately and reproducibly at frequent intervals remains elusive to the clinician caring for critically ill pediatric patients although a large proportion of these children are known to have hemodynamic compromise as a result of their illnesses. I propose to define the accuracy of clinician estimates of cardiac output in these patients and to test the ability of a new device that measures central blood volume to predict preload in this setting. I will study 100 patients in two groups. The first group will consist of critically ill pediatric patients at the time of admission to the PICU.

These 50 patients will have cardiac output estimated by three methods: Estimation by a "panel of experienced clinicians"; Calculation of the arteriovenous content difference; and Direct measurement using a new minimally invasive system based on assessing changes in ultrasound transmission in blood after intravenous administration of a small (< 5 ml) bolus of normal saline. Group two will consist of 50 PICU patients who therapeutically require either a fluid bolus (25 patients) or a single intravenous dose of diuretic (25 patients). These patients will be assessed in the same fashion at the outset, but will have ongoing measurements of cardiac output and central blood volume using the minimally invasive system and central venous pressure using conventional techniques at 3 intervals within the first hour. Data will be analyzed to evaluate the strength of correlations between expert assessment and objective measures of cardiac output in all patients (linear regression with calculation of correlation coefficient). Group two patients will be divided according to therapeutic intervention. The strength of correlation between changes in central blood volume and changes in central venous pressure and cardiac output will be assessed using similar statistical techniques.