Résumé : BACKGROUND: The effects of different types of hydroxylethyl starch (HES) on blood coagulation closely depend on their physicochemical properties. HES with lower molar substitution and a lower in vivo molecular weight interferes relatively little with hemostasis and therefore results in lower perioperative blood losses and red blood cell (RBC) transfusion. To test this hypothesis, we analyzed pooled data from all available studies in major surgery comparing 6% HES 130/0.4 and 6% HES 200/0.5 from waxy maize starch. METHODS: Estimated blood loss, drainage loss, calculated blood loss, transfused blood product volumes, and coagulation variables were examined for 24 h after the start of surgery. Groups were compared using analysis of variance, evaluating several covariates. RESULTS: Four-hundred-forty-nine patients from seven clinical trials were analyzed, 228 received HES 130/0.4, and 221 received HES 200/0.5. For HES 130/0.4 patients, when compared to HES 200/0.5 patients, the estimated blood loss was reduced by 404 mL [P = 0.006], drainage loss was 272 mL less [P = 0.009], and calculated RBC loss was 149 mL less [P = 0.003]. RBC transfusion volumes were also lower for HES 130/0.4 by 137 mL [P = 0.004]. In the early postoperative phase, HES 130/0.4 was found to exert significantly less effect on measures of coagulation, especially activated partial thromboplastin time and von Willebrand factor (antigen and ristocetin cofactor), than HES 200/0.5. CONCLUSIONS: Blood loss and transfusion requirements can be significantly reduced in major surgery when using third generation HES 130/0.4 (Voluven®) compared to second generation waxy maize starch HES 200/0.5. Since HES 130/0.4 and HES 200/0.5 were found similar regarding volume efficacy in other studies, HES 130/0.4 is recommended in this clinical setting. © 2008 International Anesthesia Research Society.