SummaryAlbumin is often cited in textbooks as the gold standard for fluid replacement in paediatrics, but in practice artificial colloids are more frequently used. Although one concern with the use of artificial colloids is their intrinsic action on haemostasis, the available data in children are inconclusive for 6% hydroxyethyl starch 130 ⁄ 0.4 (HES) and no data exist for gelatine solution with respect to coagulation. A total of 42 children (3-15 kg) undergoing surgery and needing colloid replacement were randomly assigned to receive 15 ml.kg )1 of either albumin 5%, 4% modified gelatine solution or 6% hydroxyethyl starch 130 ⁄ 0.4 solution. Standard coagulation tests and modified thrombelastography (ROTEM) were performed. After colloid administration, routine coagulation test results changed significantly and comparably in all groups, although activated partial thromboplastin time values increased more with gelatine and HES. Coagulation time was unchanged in the children who received albumin or gelatine but other activated modified thrombelastography values were significantly impaired in all groups. After gelatine and after albumin the median clot firmness decreased significantly but remained within the normal range. Following HES, coagulation time increased significantly, and clot formation time, a angle, clot firmness, and fibrinogen ⁄ fibrin polymerisation were significantly more impaired than for albumin or gelatine, reaching median values below the normal range. From a haemostatic point of view it might be preferable to use gelatine solution as an alternative to albumin; HES showed the greatest effects on the overall coagulation process. Hypovolaemia is the most common cause of circulatory failure in children and may lead to critical tissue perfusion and eventually multiple-organ failure. Administration of albumin has been regarded as the gold standard for maintaining adequate colloid osmotic pressure in infants thereby minimising oedema formation following highdose crystalloid resuscitation. Albumin has been shown to exhibit plasma-expanding effects [1]; and with increased capillary leakage, interstitial oedema might develop earlier and persist even longer than with administration of artificial colloids. This, and the considerable cost of albumin administration, might explain why various hydroxyethyl starch (HES) or gelatine solutions are more frequently used than albumin in paediatric practice [2]. All intravenous fluids provoke dilutional coagulopathy and colloids interact specifically with the coagulation system. It remains unclear which is the ideal fluid for children. Available data on the influence of colloids on haemostasis in infants are inconclusive because of the limited significance of routine coagulation tests or native thrombelastography assays in estimating fibrinogen ⁄ fibrin polymerisation. To our knowledge, no data on haemostatic changes exist for children receiving gelatine solution.To test the hypothesis that artificial colloids impair the clot formation process significantly more than...