Abstract
Monitoring cardiac output (CO) is helpful in the perioperative management of the patient with severe mitral regurgitation (MR). We assessed the accuracy and precision of the Cheetah CO monitor in patients with moderate or severe MR undergoing right and left heart catheterization as part of their pre-operative evaluation for mitral valve surgery. Cheetah CO was obtained concurrently with thermodilution CO (TD CO). Bias data was non-normally distributed; therefore, a non-parametric equivalent to Bland and Altman limits of agreement was used. Additionally, the proportions of differences between the experimental and reference method that were ≤ 0.5 L/min, ≤ 1 L/min, and >1 L/min were calculated. Twenty-seven subjects were enrolled and completed the study. The median difference between Cheetah and TD CO measurements was - 0.82 L/min, and the 5th and 95th centiles were - 6.05 L/min and 3.25 L/min, respectively. Of all differences, 25.9%, 51.9%, and 48.1% were ≤ 0.5 L/min, ≤ 1 L/min, and > 1 L/min. No proportional bias was present. We conclude that the Cheetah CO measurements in patients with moderate to severe MR cannot be used interchangeably with TD CO due to a large bias and imprecision.