Evaluation of three techniques for documenting Staphylococcus epidermidis vascular prosthetic graft infections.

Staphylococcus epidermidis (S. epidermidis) vascular prosthetic graft infections are notoriously hard to detect. Three different techniques of determining whether vascular prosthetic grafts were infected using a dog model were evaluated. Aortic angiograms were compared with nuclear magnetic resonance (NMR) imaging and systemic norepinephrine (NE) kinetics to determine if either newer technique would be more reliable than standard angiograms. Twelve dogs were randomized to control (n = 6) or infected groups (n = 6). All dogs had a 5 cm section of their infrarenal aorta replaced with knitted Dacron vascular prosthetic graft. The grafts in the infected group were contaminated by soaking them in a broth containing S. epidermidis. NE production and clearance rates were calculated for all animals after an infusion of 3H-NE using the steady-state radionuclide tracer methodology. One week following graft insertion, dogs were reanesthetized, and the 3H-NE infusion and measurements were repeated. Standard angiograms and NMR imaging were also performed. Once all tests were performed, the prosthetic grafts were removed for cultures. Comparisons between the initial and final norepinephrine measurements for each group were made using the nonparametric Wilcoxon two-sample test, while comparisons between the groups were made by chi square or the Student's t test. Angiogram results were similar for control and infected animals. Angiograms missed disruption of the proximal anastomosis found in three of the six infected dogs at graft removal. None of the six control animals, while five of the six infected animals, had localized areas of high signal intensity on NMR imaging (P less than 0.01) suggesting abscess formation.(ABSTRACT TRUNCATED AT 250 WORDS)