Postoperative deep infection in tumor endoprosthesis reconstruction around the knee

Background. Although deep infection remains one of the most diffi cult complications to manage in the treatment of musculoskeletal tumor reconstructed with an endoprosthesis, limited information with respect to its incidence and risk factors has been reported. Methods. This multicenter, retrospective, uncontrolled study reviewed the medical records of 82 patients who underwent reconstruction with an endoprosthesis or temporary spacer for bone-immature patients after resection of malignant bone tumor around the knee. Risk factors for deep infection and the impact of deep infection on prosthesis survival and oncological outcomes were analyzed. Deep infection was defi ned according to the Centers for Disease Control and Prevention (CDC) guidelines with minor modifi cation. Results. Deep infection occurred in 14 cases (17%), identifi ed at a mean of 10.9 months (range <1 to 48 months) after initial surgery. Univariate analysis identifi ed surface infection (P < 0.001) and skin necrosis (P < 0.001) as risk factors associated with deep infection. Conversely, tumor origin, chemotherapy, number of postoperative antibiotics, and length of bone resection were not associated with infection. Subclass analysis in femur cases identifi ed a correlation between infection and the extent of partial resection of the quadriceps muscle (P = 0.04). In the multivariate analysis, surface infection represented an independent risk factor for deep infection (P = 0.03). Deep infection was a risk for endoprosthesis survival (P = 0.003) but did not affect the oncological outcome. Conclusions. A strong correlation between the condition of soft tissue and establishment of deep infection is suggested in this study. Although practical options for preventing deep infection seem limited, the present data allow a form of perioperative evaluation for patients with a higher risk of deep infection.