PHYSIOLOGY AND PATHOPHYSIOLOGY: Total body water represents 60% of body weight (BW), consisting of 40% BW intra- and 20% BW extracellular fluid. Extracellular fluid is divided into 16% BW interstitial fluid and 4% BW plasma volume (Fig. 1). The colloid oncotic pressure (COP) of the plasma proteins, which is about 25 mmHg (Fig. 2), is the main factor for the retention of intravascular volume and the prevention of interstitial edema. Within a defined range, oxygen transport capacity can be improved by normovolaemic haemodilution. Under strictly normovolaemic conditions ("controlled haemodilution"), the critical haemoglobin concentration for intensive care patients is about 10.0 g/dl, and 8-6 g/dl in patients with satisfactory compensatory mechanisms under stable clinical conditions. Larger blood volume deficits are replaced step by step with volume replacement solutions (crystalloids and synthetic colloids), packed red cells, fresh-frozen plasma, and platelets (Fig. 3). VOLUME REPLACEMENT WITH CRYSTALLOID AND COLLOID SOLUTIONS: Crystalloid solutions do not contain any macromolecules. Due to their lack of intrinsic COP, they spread rapidly over the intravascular and interstitial space. To achieve a comparable volume effect like colloid solutions, a fourfold infusion volume is necessary. Thus, crystalloids should be used in addition to colloid solutions to compensate the interstitial fluid deficit. Hyperosmotic-hyperoncotic solutions have not yet been established, and their benefit seems doubtful. Synthetic colloid solutions contain gelatin, dextran, or hydroxyethyl starch (HES) molecules. Due to their intrinsic COP, fluid is fixated in the intravascular space (Fig. 4). Solutions with high COP increase the intravascular volume due to resorption of interstitial fluid (plasma expanders). Pharmacological characterisation of synthetic colloids includes concentration [%], mean molecular weight [x1,000 Dalton] and degree and position of substitution (HES only). Main clinical features are the maximal volume effect and the duration of a 100% and a 50% volume effect (Fig. 5). For economic reasons, 5% albumin should not be used for volume replacement. The use of 20% albumin in intensive care patients is also limited and recommended only if a capillary leck is unlikely and the dose limits of synthetic colloids are reached. Gelatin is a polypeptide of bovine origin and achieves a shortlasting isovolaemic volume effect. When compared with dextran or HES, negative effects on haemostatis are less, and the renal function is not impaired. Thus, gelatin is first of all indicated in patients with limited volume demand, and secondly in situations with massive blood losses, when the dose limits of HES are reached. Dextran has no specific benefits. HES is a polysaccharide of maize or potato origin. By substitution of glucose molecules with hydroxyethyl groups, starch molecules are protected against fast amylase degradation. Metabolism depends on the degree and position of substitution and mean molecular weight. Smaller molecules are eliminated via the kidneys, but a certain amount of larger molecules is stored in the reticulo-endothelial system. HES is available in very different preparations (concentration 3-10%, mean molecular weight 70,000-450,000 Dalton, substitution 50-70%). Special indications of 10% HES 200/0.5 are rapid hypervolaemic replacement of massive blood losses and increase of COP in intensive care patients without capillary leak. Synthetic colloids as well as albumin may lead to adverse reactions, which are generally very rare. In large-scala studies, no significant differences have been found with regard to incidence and severity.