931C orrection of hypovolemia is fundamental, but it is one of the most important factors in managing critically ill patients. In patients with sepsis, intensivists and investigators are generally interested in the type of colloids and crystalloids, as well as the total volume of fluids required to overcome hypovolemia. However, they may have been overlooking the infusion rate of administered fluids, since it is believed that hypovolemia should be treated as early as possible to improve survival (1). In this issue of Critical Care Medicine, Bark et al (2) from Lund University Hospital in Sweden report their experiment highlighting the importance of the infusion rate of different types of fluids for maintaining normal plasma volume in septic rats. A fixed volume of different types of colloids and a crystalloid including 5% albumin, 6% hydroxyethyl starch 130/0.4, 4% gelatin, and 0.9% NaCl was administered via fast or slow infusion. In the fast-infusion group, fluids were administered over 15 mins, while in the slow-infusion group, fluids were administered over 3 hrs. Measurement of plasma volume was carefully performed at 3 hrs after the start of volume resuscitation using human iodine-125 ( 125 I) albumin, even though this volume marker is known to overestimate plasma volume in critical illness associated with increased capillary protein leakage (3). Although the interval between the end of fluid administration and measurement was apparently different between the two infusion groups, the authors demonstrate the negligible effect of the interval by additional experiments (Fig. 4C). The results show that plasma volume expansion by a colloid is greater when infused slowly, an effect more pronounced with albumin. The authors also speculate the mechanism mediating the increased transcapillary fluid loss due to the fast infusion of colloids in the presence of increased permeability. A transient increase in arterial pressure and hemo dilution would have a significant impact on transcapillary fluid loss, but this mechanism was not observed in the absence of increased permeability (4). In a clinical setting, Wills et al (5) showed that the effect of fast infusion of hydroxyethyl starch or dextran is obvious only during the first 2 hrs in children with dengue shock syndrome characterized by severe vascular leakage as judged by changes in hematocrit. Considering the results of the experimental study by Bark et al (2) as well as human studies indicating that a lower net fluid balance is associated with fewer days of ventilator use and fewer days in the ICU (6, 7), a slow rather than a fast infusion of colloids, particularly albumin, would have the potential to improve survival, even though evidence that colloids provide better survival is lacking (8).Bark et al (2) gave three-fold more 0.9% NaCl than colloids to achieve a similar effect on plasma volume expansion but resultant expansion was poor regardless of the infusion rate because crystalloids were rapidly distributed from the intravascular to the interstitial fluid comp...