Estimation of Errors in Determining Intrathoracic Blood Volume Using the Single Transpulmonary Thermal Dilution Technique in Hypovolemic Shock

Abstract

Background The transpulmonary thermal dilution technique has been widely adopted for monitoring cardiac preload and extravascular lung water in critically ill patients. This method assumes intrathoracic blood volume (ITBV) to be a fixed proportion of global end-diastolic volume (GEDV). This study determines the relation between GEDV and ITBV under normovolemic and hypovolemic conditions and quantifies the errors in estimating ITBV. Methods Nineteen pigs allocated to control (n = 9) and shock (n = 10) groups were studied. Shock was maintained for 60 min followed by volume resuscitation. The dual dye-thermal dilution technique was used to measure GEDV and ITBV (ITBVm) at baseline (time 0), shock phase (30 and 90 min), and after resuscitation (150 min). The regression equations estimated from paired GEDV and ITBVm measurements under normovolemic and hypovolemic conditions were used to estimate ITBV from the corresponding GEDV, and the estimation errors were quantified. A more simplified equation, used in a commercially available clinical monitor (ITBV = 1.25 x GEDV), was then used to estimate ITBV. Results The regression equation in the control group was ITBVm = 1.21 x GEDV + 99 (r = 0.89, P < 0.0001) and in the shock group at 30 and 90 min was ITBVm = 1.45 x GEDV + 0.6 (r = 0.95, P < 0.0001). The 95% confidence interval for the y-intercept was relatively wide, ranging from 31 to 168 and -47 to 49, respectively, for the two equations. The equation estimated in the control group led to overestimation of ITBV and a significant (P < 0.05) increase in errors in the shock group at 30 and 90 min. Errors in estimating ITBV using the simplified commercial algorithm were less than 15% under normovolemic and hypovolemic conditions. Conclusions The linear relation between GEDV and ITBV is maintained in hypovolemic shock. Even though the relation between GEDV and ITBV is influenced by circulatory volume and cardiac output, the mean errors in predicting ITBV were small and within clinically tolerable limits.