Blood volume measured by ultrasound and radioisotope dilution in critically ill subjects.

BACKGROUND: Accurate bedside assessment of circulating blood volume (BV) continues to challenge clinicians in their attempt to implement goal-directed therapy in the critically ill subject. The aim of this investigation was to comparatively evaluate BV measurements obtained by ultrasound and radioisotope dilution methodologies in adult subjects admitted to a surgical intensive care unit. MATERIALS AND METHODS: Fifty subjects with concurrent central venous catheters and peripheral arterial lines underwent measurement of BV using both ultrasound and radioisotope dilution (BV-RD) methods. The ultrasound dilution method was performed using a 30-mL injectate (BV-UD30) and a 60-mL injectate (BV-UD60) of isotonic saline. RESULTS: There were 24 paired data points for the BV-UD30 and 40 paired data points for the BV-UD60 measurements. Spearman's rank-order correlation demonstrated a positive relationship comparing both the BV-UD30 (r = 0.46, P = 0.0249) and the BV-UD60 (r = 0.80, P < 0.0001) to values obtained by radioisotope measurements. Bland-Altman analysis showed a mean bias of 1329 mL with limits of agreement (LOA) ± 2559 mL comparing BV-RD and BV-UD30, and a mean bias of 62 mL with LOA ±1353 mL for BV-RD and BV-UD60. CONCLUSIONS: This preliminary investigation shows that the BV-UD60 had better agreement with BV-RD, compared with the BV-UD30, but its utility appears limited by a large LOA. As this technology continues to evolve, the ultrasound dilution approach may potentially become a feasible means to calculate BV in critically ill surgical subjects.