BACKGROUND Thrombin-mediated vascular smooth muscle cell proliferation has been implicated in coronary restenosis. Attempts to inhibit this mitogenic activity have recently focused on non-physiologic direct thrombin inhibitors, whereas endogenous thrombin inhibitors such as antithrombin III (ATIII) have received little attention. ATIII is the main physiologic inhibitor of thrombin and may thus be a potential therapeutic agent for prevention of restenosis. METHODS Human ATIII (125 U/kg) and heparin (200 U/kg) were administered to 12 atherosclerotic swine 30 min prior to inducing restenosis by oversized stent (left anterior descending and right coronary arteries; stent-to-artery ratio approximately 1.2) and balloon injury (circumflex; balloon artery ratio approximately 1.2). Eleven control swine received only heparin every 6 h for 24 h and were subjected to similar stent and balloon injury. Quantitative coronary angiography [change in minimal lumen diameter (delta MLD)] and morphometric analysis [percentage area stenosis (PAS)] were performed 4 weeks later. RESULTS ATIII activity (mean +/- SD) of treated swine increased from a baseline of 103 +/- 10% to a peak of 266 +/- 48%, whereas trough levels were maintained at 259 +/- 55% for 72 h by drug infusions every 6 h. The delta MLD, the primary angiographic endpoint in the balloon injured vessel was -0.57 +/- 0.33 mm in heparin group versus -0.26 +/- 0.27 mm in the ATIII group (P < or = 0.03). For stented vessels the delta MLD was -0.61 +/- 0.33 mm in the heparin group versus -0.41 +/- 0.37 mm in the ATIII group (P < or = 0.06). The PAS for the balloon injured vessels was 30 +/- 12% in the heparin group versus 19 +/- 14 in the ATIII group (P < or = 0.06). In stented vessels the PAS was 45 +/- 16% in the heparin group versus 38 +/- 16% in the ATIII group (P < or = 0.1). CONCLUSIONS Supraphysiologic ATIII levels in combination with heparin inhibits the reduction in MLD in coronary arteries subjected to oversized balloon injury and demonstrates a beneficial trend in arteries subjected to oversized stent injury. These data provide cautious optimism for further investigation with ATIII to prevent coronary restenosis.