To evaluate the extent of shed blood activation in two autotransfusion systems and the effect of circulating blood activation upon autotransfusion, we performed a prospective study in 18 patients undergoing internal mammary artery bypass operation and a control group of 10 patients. The autotransfusion systems were from Sorin (n = 9) consisting of a hard shell reservoir with a filter having a small contact area (0.32 m2), and from Dideco (n = 9) consisting of a hard shell reservoir with a filter having a larger contact area (4.64 m2). We found high concentrations of thromboxane, fibrinogen degradation products, complement split product C3a, and elastase in the shed blood and, with the exception of C3a, in the circulating blood of autotransfused patients. There was no such activation in control patients. The degree of the systemic inflammatory reaction was determined by the type of autotransfusion system and by the amount of infused shed blood. The Dideco system provoked more inflammatory response than did the Sorin. This was reflected by the larger shed blood loss during autotransfusion in the Dideco patients than in Sorin patients, resulting in infusion of more shed blood (means, 737 mL versus 566 mL; not significant). After autotransfusion, Dideco patients shed significantly more blood than did Sorin or control patients (p < 0.05). Dideco patients also needed more colloid/crystalloid solution per 24 hours than Sorin patients (p < 0.05). This became clinically relevant only after infusion of more than 800 mL of shed blood (p < 0.001): hemodilution indicated the need for packed cells in 4 Dideco patients and in 1 Sorin patient.(ABSTRACT TRUNCATED AT 250 WORDS)