Background: Colloid fluids influence the coagulation system by diluting the plasma and, potentially, by exerting other effects that are unique for each fluid product. We hypothesised that changes in the coagulation measured at the end of surgery would be mainly governed by differences in half-life between the colloid fluids. Methods: Eighty-four patients were randomised to receive one of four colloids: HES 130/0.42/6 : 1 (Venofundin (R)), 130/0.4/9 : 1 (Voluven (R)), 200/0.5/5 : 1 (Haes-steril (R)) and 6% dextran 70 (Macrodex (R)). Blood samples were taken just before and after a preoperative 500ml bolus, and also after subsequent elective hip replacement surgery. Volume expansion was estimated from the blood dilution and coagulation assessed by ROTEM, activated partial thromboplastin time, prothrombin international normalised ratio (PT-INR), D-dimer and thrombin-antithrombin complex (TAT). Results: The blood volume expansion amounted to approximately 600 ml for all four colloids directly after infusion. Voluven (R) and Haes-steril (R) prolonged the aPT time and Venofundin (R) increased TAT. Although all colloids increased PT-INR and D-dimer, the ROTEM analyses showed that they consistently shortened the clotting time and weakened the clot strength. These effects were mainly unchanged after surgery, during which the haemorrhage averaged 500-600 ml. Macrodex (R) produced a stronger volume support at the end of the surgery (91% of infused volume; Pless than0.001) than the three starch solutions (42-60%). Conclusions: All tested colloid fluids induced a mild hypercoagulable state with faster clotting, but with weaker clot strength. The additive influence of surgery was relatively small, and postoperative changes in coagulation were mainly due to differences in the half-life of each colloid.