OBJECTIVE: to study the histopathological characteristics of prosthetic vascular graft infection. DESIGN: prospective clinical study over 2 years. SETTING: University Hospital. MATERIALS: 36 infected and 29 uninfected (control) chronically implanted vascular prostheses (half aortic) were removed and 352 sections prepared. CHIEF OUTCOME MEASURES: light microscopy (multiple stains), scanning electron microscopy (SEM), and multiple culture techniques to identify characteristics of healing, infection, and microorganisms. MAIN RESULTS: Acute inflammation (AI) (neurophils, granulocytes and necrosis) were seen in 75% of infected grafts, were most prominent in the perigraft tissue and rarely seen on the luminal surface. These were usually well localised, leaving the remainder of a graft well incorporated with no signs of infection. In 25% of clinically infected, culture-positive grafts there was no significant acute inflammation. Chronic inflammation (CI) (macrophages, lymphocytes, monocytes, giant cells) was seen in 70% of both control and infected grafts. In 50% of both groups a significant lymphocytic population was composed exclusively of T-lymphocytes suggesting a true host vs graft response. Unincorporated chronically implanted grafts (> 1 yr) were seen with equal frequency in the two groups although more diffusely unincorporated grafts were infected. Microorganisms were cultured from 23 infected grafts (64%) and were, at microscopy, mostly found outside the graft and nerves on the luminal side. CONCLUSIONS: This data cast doubt on criteria commonly used to distinguish graft infections and host vs. graft reactions from normal graft healing. Acute and chronic inflammation are not predictive of infection.
OBJECTIVE: to study the histopathological characteristics of prosthetic vascular graft infection. DESIGN: prospective clinical study over 2 years. SETTING: University Hospital. MATERIALS: 36 infected and 29 uninfected (control) chronically implanted vascular prostheses (half aortic) were removed and 352 sections prepared. CHIEF OUTCOME MEASURES: light microscopy (multiple stains), scanning electron microscopy (SEM), and multiple culture techniques to identify characteristics of healing, infection, and microorganisms. MAIN RESULTS: Acute inflammation (AI) (neurophils, granulocytes and necrosis) were seen in 75% of infected grafts, were most prominent in the perigraft tissue and rarely seen on the luminal surface. These were usually well localised, leaving the remainder of a graft well incorporated with no signs of infection. In 25% of clinically infected, culture-positive grafts there was no significant acute inflammation. Chronic inflammation (CI) (macrophages, lymphocytes, monocytes, giant cells) was seen in 70% of both control and infected grafts. In 50% of both groups a significant lymphocytic population was composed exclusively of T-lymphocytes suggesting a true host vs graft response. Unincorporated chronically implanted grafts (> 1 yr) were seen with equal frequency in the two groups although more diffusely unincorporated grafts were infected. Microorganisms were cultured from 23 infected grafts (64%) and were, at microscopy, mostly found outside the graft and nerves on the luminal side. CONCLUSIONS: This data cast doubt on criteria commonly used to distinguish graft infections and host vs. graft reactions from normal graft healing. Acute and chronic inflammation are not predictive of infection.