BACKGROUND: The implementation of intraoperative normovolemic hemodilution is a strategy used in an attempt to diminish the need for or obviate allogeneic transfusion and to avert the potential complications. The goal of this study was to evaluate the safety and efficacy of moderate intraoperative normovolemic hemodilution. METHODS: Fifteen patients scheduled for posterolateral spinal fusion underwent intraoperative normovolemic hemodilution with 10% pentastarch to a target hematocrit level of 25% (hemodilution group). All units of blood procured by hemodilution and additional allogeneic blood was transfused in the perioperative period to maintain the hematocrit level of>25%. We investigated the effect of hemodilution on whole blood coagulation as measured by the thromboelastography and, evaluated its efficacy of decreasing the need for allogeneic blood transfusion as compared to the nonhemodilution group, retrospectively. RESULTS: Hemodilution with pentastarch caused a decrease in response and coagulation time (p<0.01) but did not influence on the alpha angle and maximum amplitude on thromboelstogram. Amount of transfusion of allogeneic blood was 4.1 1.7 units for the hemodilution group and 5.0 1.5 units for the nonhemodilution group. Net red blood cell volume ""saved"" from hemodilution was about 120 ml. CONCLUSION: Moderate intraoperative normovolemic hemodilution with pentastarch does not affect the coagulability of whole blood. But its efficacy of decreasing the need for allogeneic blood transfusion is minimal.