Stroke volume variation during acute normovolemic hemodilution

The intravascular volume of surgical patients should be optimized to avoid complications associated with both overhydration and underresuscitation. In patients undergoing intraoperative acute normovolemic hemodilution, we investigated whether stroke volume variation (SVV) derived from an arterial pressure-based cardiac output (CO) monitor system (FloTrac/Vigileo, Edwards Lifesciences, Irvine, CA) tracked the changes associated with blood removal and replacement. We further evaluated the correlations between SVV and 3-dimensional (3D) transesophageal echocardiographic (TEE) left ventricular (LV) volume measurements.

Twenty-five patients had procedures during which acute normovolemic hemodilution was a planned part of the intraoperative management. We defined 7 measurement timepoints: baseline, after the removal of 5%, 10%, and 15% of the estimated blood volume (EBV) and after replacement with an equal volume of 6% hetastarch to -10%, -5%, and baseline EBV. At each timepoint, heart rate and systolic, diastolic, and mean arterial blood pressure were obtained from standard monitors, CO and SVV measurements were obtained from the FloTrac/Vigileo monitor, and TEE images were recorded for subsequent off-line reconstruction and determination of LV end-systolic and end-diastolic volumes. For statistical evaluations, we used a mixed models analysis of variance and Dunnett's test for post hoc comparisons with baseline values. Pearson's correlation was used to examine the relationships between SVV and LV volume.

Analysis of variance demonstrated no significant change in heart rate or mean arterial blood pressure over the duration of study. CO decreased from 4.9 +/- 0.3 to 4.5 +/- 0.3 L/min after removal of 15% of the EBV and then increased to a final value of 5.4 +/- 0.3 L/min after replacement of 15% of the EBV. SVV increased from 9.2% +/- 0.9% to 20.3% +/- 2.0% (P < 0.001) after removal of 15% of the EBV and returned to a final value of 7.2% +/- 0.9% after replacement of 15% of the EBV. The indexed LV end-diastolic volume decreased from 42.1 +/- 8.3 to 36.9.3 +/- 8.3 mL/m(2) (P < 0.001) after removal of 15% of the EBV and then returned to a final volume of 45.9 +/- 10.3 mL/m(2) after replacement of 15% of the EBV. The measurements of SVV correlated inversely with the 3D TEE LV volume measurements.

The SVV derived from the FloTrac/Vigileo system changes significantly as blood is removed and replaced during hemodilution. These changes correlate with 3D TEE measurements of LV volume. The utility of SVV in guiding optimization of intravascular volume merits further study.