Cardiac output (CO) is an important hemodynamic measure that helps to guide the therapy of critically ill patients. Invasive CO assessment in infants and children is often avoided because of the inherent risks. A noninvasive CO monitor that uses partial rebreathing has been recently developed to determine CO via the Fick principle for carbon dioxide. There have been no clinical studies confirming its accuracy in pediatric patients. This is a prospective observational study of 37 children <12 yr of age who underwent cardiac catheterization. Under general anesthesia via an endotracheal tube without a leak, we made multiple CO measurements using thermodilution and compared them with noninvasively determined CO measurements. Paired measurements were analyzed for bias, precision, and correlation via Bland-Altman plot and linear regression. Noninvasive measurements showed a linear correlation with thermodilution CO assessment with an r value of 0.83 (P < 0.03). Bland-Altman analysis yielded a bias of -0.27 L/min and a precision +/-1.49 L/min. Cardiac index measurements demonstrated a decreased r value of 0.67 (P = 0.15) and a bias of -0.18 L . min(-1) . m(-2) and precision of +/-2.13 L . min(-1) . m(-2). Differences between partial rebreathing measurements and thermodilution measurements were largest in children with a body surface area of </=0.6 m(2) ventilated with tidal volumes <300 mL. Based on these findings, noninvasive CO measurement using partial rebreathing may be clinically acceptable in children with >0.6 m(2) body surface area and >300 mL tidal volume.