OBJECTIVE To evaluate the direct action of PAF on the pro-coagulant activity of cultured vascular endothelial cells; to analyse by photometric methods the thrombogenic effect of PAF on platelets, and to assess platelet deposition on vascular endothelium. METHODS Human endothelial cells were isolated from umbilical cord vein and incubated for three minutes at 22 degrees C with different PAF concentrations (10(9) M to 10(-7) M) in order to assess the influence of this lipidic mediator on the procoagulant activity of the cells. The effect of PAF on platelet aggregation was assessed by aggregation studies using arachidonic acid (AA) and different PAF and Lyso-PAF concentrations (10(-8) M to 10(-4) M). Serotonin (5-HT) release was tested in platelet rich plasma (PRP) samples highly sensitive to PAF. PRP samples were incubated for 30 minutes at 22 degrees C with 100 microCi 3H-5-HT. Platelets were activated with 10(-7) M to 10(-9) M PAF concentrations, the percentage of 3H-5-HT released into the extra-platelet medium being calculated. Baumgartner's continuous perfusion model was used to study platelet deposition on the vascular endothelium. The morphometric evaluation was carried out by a planimeter assembled to a data processor with a programme devised for analysing the platelet-subendothelium interaction. RESULTS Evaluation of the procoagulant activity on the cell surface: The expression of the procoagulant activity showed no variations with respect to the controls under different concentrations of PAF. Platelet aggregation and release studies: Normal values of platelet aggregation (80% to 100%) were found when using AA, however, there were great case to case variations under different PAF concentrations. Serotonin release was less marked than aggregation itself. The use of Lyso-PAF failed to elicit platelet aggregation and serotonin release in any case. Morphometric evaluation: The results attained showed that perfusion carried out with 10(-8) M PAF concentration showed contact, adhesiveness and thrombus formation figures similar to those of control perfusion. CONCLUSIONS Human platelets are not too sensitive to PAF activity, only high PAF concentrations being capable of inducing platelet aggregation and 5-HT release with ample variability. This suggests the existence of a heterogeneous platelet population with PAF receptors. Low PAF concentrations do not modify the haemostatic function, and only those PAF concentrations inducing maximal release and aggregations could reduce the interaction of platelets with vascular subendothelium and the formation of thrombi.