Prosthetic vascular graft infection: A risk factor analysis using a case–control study

Abstract

Several factors have been anecdotally reported to increase the risk of prosthetic vascular graft infection (PVGI), a serious complication with high mortality and limb amputation rates. The goal of this study is to confirm purported risk factors using a well-designed statistical model. Patients undergoing vascular graft placement at Mayo Clinic Rochester between January 1, 1982 and December 31, 2002 were retrospectively evaluated, excluding redo cases. PVGI cases had to meet one of the following criteria: perigraft air or fluid >8 weeks postoperatively, positive Gram stain or cultures of graft or perigraft material, intraoperative gross purulence, failure of graft incorporation, or exposed graft. Each case had two controls matched for age, gender and date of surgery. Duration of follow-up for a given control had to be at least equal to the time to infection of the corresponding case. Demographic and other clinical data were collected for each patient. Fifty-one cases and 102 controls met the inclusion criteria. Median age was 68 years. Staphylococcal species were identified in 45% of the total number of cases, and 60% of cases with a recovered organism. Univariate analysis identified groin incision, wound complication and wound infection as significant risk factors for PVGI (p<0.005 for all three factors). After adjusting for age and surgery date, multivariate analysis confirmed the statistical significance of groin incision and wound infection as risk factors for PVGI. This is the first published risk factor analysis that includes statistical evaluation in a case-control study of PVGI. Groin incision, wound complication, and wound infection are significantly associated with developing PVGI. These findings will assist in identifying patients at increased risk of developing subsequent infection. In order to confirm purported risk factors of prosthetic vascular graft infections (PVGI), a well designed case-control study was conducted at Mayo Clinic, Rochester, MN. Univariate and multivariate analyses of 51 cases and 102 controls identified three statistically significant risk factors for the development of PVGI: groin incision, wound complication, and wound infection.