The mechanism of anastomotic thrombosis in microvascular surgery remains poorly understood. We hypothesized that thrombin activity at anastomoses plays a major role in this process. To study this, a surgically relevant human artery anastomosis model was used to (i) measure surface thrombin activity on anastomoses and on intact vessel, (ii) determine the inhibitability of surface thrombin by heparin and recombinant hirudin (r-hirudin), and (iii) determine the anastomotic and intact vessel binding capacity for additional thrombin. Human placental artery segments were placed in chambers in which 0.2 cm2 of luminal surface was exposed to citrated platelet-poor plasma for 10 min at 37 degrees C. The fibrinopeptide A (FPA) concentration (indicating the action of thrombin on fibrinogen) in the supernatant was then measured using an ELISA assay. Intact vessels and anastomoses expressed equivalent thrombin activity that could not be inhibited by heparin at a concentration (0.3 U/ml) that is sufficient to prolong the activated partial thromboplastin time two-fold. Conversely, the concentration of heparin routinely used in intraoperative vessel irrigation solutions (50 U/ml) was able to completely block thrombin activity at both sites. r-Hirudin (0.3 heparin equivalent anti-IIa U/ml) was able to inhibit nearly all of the thrombin activity on each site. Each site was able to bind and express the activity of additional thrombin, indicating the potential for increased vessel thrombogenicity after local clot has formed and has been removed. These data indicate the presence of thrombin on dissected human vessels and its presence in equal amounts on intact and anastomosed vessels when measurement is made before blood flow resumes. Furthermore, vessel-associated thrombin is resistant to a standard systemic concentration of heparin but is susceptible to the much higher heparin concentration that can be delivered locally by the surgeon during vessel irrigation.