Reversal of systemic heparinization with protamine is problematic during perfusion with heparin surface coated devices because protamine reacts with both circulating and surface bound heparin. Hence, the development of a deheparinization device allowing for ex vivo heparin absorption by the means of an immobilized polycation is of prime interest for a number of indications. To assess the coagulation patterns during ex vivo deheparinization, a heparin surface coated venovenous pump loop including a plasma separator with immobilized polycation was studied in a bovine model (n = 6, body weight 71 +/- 5 kg). After systemic heparinization with 300 IU of heparin/kg body weight, spontaneous evolution (control) of coagulation parameters was compared to ex vivo deheparinization with a mean pump flow of 500 ml/min. No device failure occurred during the procedures and all plasma separators remained patent. These baseline levels were measured for control versus ex vivo deheparinization: activated coagulation time 158 +/- 6 sec (161 +/- 4 sec: NS), antithrombin III 101 +/- 5% (108 +/- 8%: NS), fibrinopeptide A 3.4 +/- 1.7 ng/ml (4 +/- 1.7 ng/ml: NS). After heparin application mean activated coagulation time was longer than 1000 sec in both groups. Sixty minutes later, the activated coagulation time was 757 +/- 43 sec (184 +/- 5 sec: P < 0.05), antithrombin III was 96 +/- 12% (99 +/- 2%: NS), and fibrinopeptide A was 2.7 +/- 0.7 ng/ml (9.5 +/- 3.5 ng/ml: P < 0.05). It is concluded that ex vivo deheparinization resulted in significant acceleration of activated coagulation time normalization. Fibrinopeptide A production in the group with ex vivo deheparinization appeared to be higher. As antithrombin III levels were close to normal in both groups, allowing for adequate function of circulating as well as surface bound heparin, and the coagulation process was blocked by significant heparin levels in the control group, the difference for FPA may be due to activation of the coagulation process in the surgical field (sternotomy).