Comparison of endovascular and conventional vascular prostheses in an experimental infection model

Introduction: The causes and management of prosthetic graft infections have been extensively studied for conventional bypass grafts; however, the infectivity and therapy for endovascular graft infections are completely unknown. The aim of this study was to compare the biologic properties of infected aortic grafts when inserted by endoluminal or standard transabdominal techniques. Methods: Eighteen dogs underwent placement of polytetrafluoroethylene grafts in their infrarenal aortas either by an endovascular technique (8) or a standard interposition technique (10). Endovascular grafts were constructed from polytetrafluoroethylene (3 cm) and two balloon-expandable stents coaxially mounted onto a balloon catheter delivery system. The grafts were inserted through a left carotid arteriotomy under fluoroscopic control. Initially, seven grafts were infected with decreasing inocula of Staphylococcus aureus, starting at 10⁷ organisms per ml for 30 minutes and then rinsed briefly (10 seconds) in normal saline solution, until a 50% infective dose for the standard grafts was determined to be 10² organisms per ml. After this initial experiment, a second group of 11 dogs were compared at a concentration of 10² S. aureus per ml. Five dogs underwent endovascular repair, and six dogs had standard graft interpositions after an identical period of bacterial exposure. All grafts were removed at 2 weeks under sterile conditions and were submitted for quantitative culture analysis. Results: Three of the six dogs (50%) with standard grafts appeared to clear their infections, whereas only one of the five dogs (20%) with an endovascular graft was free of organisms at 14 days. This result was further manifested by statistically significant lower postmortem colony counts in the standard grafts (p < 0.01). Conclusions: The endoluminal position of the graft and its proximity to the arterial wall do not appear to provide protection against infection. These data suggest that if endovascular grafts become infected, they may be in a disadvantaged position for host defense mechanisms to be effective.