Background: Intravascular volume expansion is followed by loss of fluid from the circulation. The extravasation of albumin in this readjustment is insufficiently known. Methods: Twelve male volunteers participated, each in three separate sessions, in a controlled, randomised, open fashion. They received one of the following: albumin 40 g/L,(7.1 mL/kg, i.e. 500 mL per 70 kg); Ringer's acetate (21.4 mL/kg), or dextran 30 g/L (7.1 mL/kg). The fluids were infused during 30 min and the subjects were followed for 180 min. ECG, arterial oxygen saturation and non-invasive arterial pressure were recorded. Haemoglobin, haematocrit, serum albumin and osmolality, plasma colloid osmotic pressure and hyaluronan concentration were determined in venous samples. Results: The serum albumin concentration decreased (P < 0.05, anova) following Ringer's acetate or dextran, whereas serum osmolality was unchanged in all groups. The colloid osmotic pressure decreased (P < 0.05) after the Ringer solution. The blood volume increase was estimated from the decrease in haemoglobin concentration and did not differ between the three fluids. The cumulated extravasation of albumin was largest following albumin (10.4 ± 5.4 g, mean ± SD), less following dextran (5.6 ± 5.0 g) and negligible in the Ringer group (0.5 ± 10.0 g; P < 0.05 against albumin). However, the Ringer solution increased the plasma concentration of hyaluronan drastically. Conclusions: Infusion of hypotonic colloidal solutions entails net loss of albumin from the vascular space. This is not the case after Ringer's acetate. Increased interstitial hydration from the latter fluid is followed by lymphatic wash out of hyaluronan.