A model of in vivo human venous thrombosis that confirms changes in the release of specific soluble cell adhesion molecules in experimental venous thrombogenesis

Purpose: The mechanisms of venous thrombogenesis have been studied by using animal models and cells in culture. The results from these systems may not , however, be relevant to the human condition. The aim of this study was to develop a method by which thrombus could be safely produced in a human vei n in vivo. The model that was developed was used as a means of studying the changes in soluble adhesion molecule expression in human venous thrombogen esis. Methods: An autologous thrombin extract tvas used to generate experimental thrombi in the disconnected portion of the long saphenous veins of 30 patie nts who were undergoing routine bilateral varicose vein surgery. The contra lateral vein was perfused with thrombin extract diluent buffer to act as th e control. The concentration of soluble P-, E- and L-selectin, intercellula r adhesion molecule 1 (ICAM-1), and vascular cell adhesion molecule-1 were measured by means of specific enzyme-linked immunosorbent assays in samples of blood taken from veins in which thrombus had formed and in contralatera l control veins. Results: Thrombosis invariably formed when at least 100 IU of thrombin acti vity was administered. Thrombus formation was independent of the time that the thrombin extract was allowed to remain within the emptied vessel. Throm bosis never developed in control vessels that were similarly treated with t he buffer used to dilute the thrombin extract. Experimental thrombi were co mposed mainly of red cells, with layers of fibrin next to platelet and leuk ocyte packages. These findings are similar to those observed in samples of established human venous thrombi. There were small but significantly higher levels of the adhesion molecules, soluble P-selectin, and vascular cell ad hesion molecule-1 in blood taken from veins in which experimental thrombi h ad formed, compared with controls (P =.015 and .007, respectively; Wilcoxon signed rank test). Serum levels of soluble L-selectin, E-selectin, and ICA M-1 were not affected by thrombosis. Conclusion: This model is safe and reproducible. It produces thrombi with a morphology similar to that described for established human deep venous thr ombi. The model may be appropriate for the study of the early changes that occur during human venous thrombogenesis and may also be of value in testin g the efficacy of novel antithrombotic agents.