Haemostatic properties of aprotinin could be associated with an increased risk of thrombosis. A randomized, blinded study was conducted to consider the potential thrombogenicity of aprotinin, using the Folts' model on femoral arteries in 12 pigs. The flow variations were measured by a pulsed Doppler in anaesthetised animals. Ear immersion bleeding time was performed. During the first part of the study, a stenosis was performed successively on both femoral arteries, each for a period of 30 min, without prior injury, to assess the integrity of the vessel, and to check that the arteries did not develop cyclic flow reductions (CFR), permanent cessation of flow (PCF) or partial thrombosis, when a stenosis is applied. Then the clamp was released and a bolus of placebo (saline), or aprotinin (4 millions KIU, followed by a continuous infusion of 1 million KIU.h-1), was administered. At the end of the bolus, the second part of the study began. Stenosis was applied to the arteries. If CRF, PCF, or partial thrombosis were observed without prior injury then the infused drug (aprotinin or saline) was considered a prothrombotic drug, and the opposite artery was studied. For each animal, right and left femoral artery segments were fixed and studied (morphologic study). Eighteen arteries were studied. In the aprotinin group, 6 arteries out of 8 developed an unexpected thrombosis, as compared with only 2 out of 10 arteries in the control group (p = 0.02). The morphologic study confirmed the occurrence of thrombosis in 4 out of 7 arteries in the aprotinin group, as compared with only 1 out of 9 in the control group.(ABSTRACT TRUNCATED AT 250 WORDS)