Twenty-nine patients, subjected to major abdominal surgery, were randomized to different types of postoperative fluid replacement, given on the basis of subcutaneous oxygen tension measurements (PscO2). One regimen consisted of crystalloids, the other of a combination of crystalloids and colloids (dextran 70). Perioperative fluid replacement was given according to clinical criteria and postoperatively according to polarographic PscO2 measurements. PscO2 was measured by a silicone catheter implanted in the upper arm. Two tubes of expanded polytetrafluoroethylene (ePTFE) were implanted next to the silicon catheter. They were removed on day 5 and 7, and analysed for hydroxyproline content. Total postoperative fluid support was equal in the two groups, and so was the hydroxyproline content on day 7. In conclusion, postoperative fluid substitution with dextran has no advantages over crystalloids only, with regard to granulation tissue formation if postoperative fluid support is optimum, according to PscO2 measurements.