Reliability of thermodilution derived cardiac output with different operator characteristics

Abstract

Background and Aims: The thermodilution method (TM) performed during right heart catheterisation (RHC) is used to determine cardiac output (CO). The calculation of CO presumes: no significant injectate recirculation, smooth injectate delivery, and no intra-cardiac blood flow perturbations caused by the injection. Studies validatingTMnever specified number of injectate-delivering operators and presumptions of the calculation leave potential for significant inter-operator variability. We aimed to clarify this. Methods: RHC experienced physicians were asked to determine 4 different CO settings of a previously validated mock circulation loop using the TM. CO of the mock circulation loop was randomly adjusted to different values with participants blinded to the CO. Participants were instructed to perform a minimum of 3 measurements of each CO and ensure that each were within 10% of each other and that they were satisfied with the result. Speciality (intensive care or cardiology), age, gender, body mass index (BMI), grip strength and number of previous RHC were recorded. A standard linear regression model was used in the analysis. Results: There were 15 participants; 10 male, 5 female (results presented as: mean (SD)): Age 37.7yrs (4.4yrs), BMI 26.4kgm-2 (5.4kgm-2), grip strength 42.3kg (10.1kg). Overall R2 = 0.85 for actual vs measured CO. Operator experience was the only variable predictive of CO measurement accuracy (p = 0.001). The kappa value for inexperienced (10 - 25 prior RHC) = 0.01 vs. very experienced operators (> 100 prior RHC) = 0.37. Conclusion: RHC operator experience is the only variable predictive of CO measurement accuracy.<br />