Preoperative Diagnosis of Infection in Patients with Nonunions

Background: The surgical treatment of a fracture nonunion is complicated in the presence of infection. The purpose-of the present study is to report on the utility of a standardized protocol to rule out infection in high-risk patients and to evaluate the efficacy of each component of the protocol. Methods: A single protocol of preoperative laboratory tests (white blood-cell count, C-reactive protein level, and erythrocyte sedimentation rate) and a combined white blood cell/sulfur colloid scan were performed for patients with a high risk of fracture nonunion. Infection was diagnosed on the basis of positive intraoperative cultures, evidence of gross infection at the time of the procedure, or evidence of gross infection during the immediate postoperative period. With use of infection as the end point, univariate analysis and multiple logistic regression analysis were used to compare tests. A risk stratification method was used to combine tests. Results: Ninety-three patients with ninety-five nonunions were evaluated. Thirty of the ninety-five nonunions were ultimately diagnosed as being infected. With use of a combination of elevated white blood-cell count, erythrocyte sedimentation rate, and C-reactive protein level and a positive scan, the predicted probabilities of infection associated with zero, one, two, and three positive tests were 18%, 24%, 50%, and 86%, respectively. With the elimination of the nuclear scan, the predicted probabilities for zero, one, two, and three risk factors were 20%, 19%, 56%, and 100%. Conclusions: The erythrocyte sedimentation rate and the C-reactive protein level were both independently accurate predictors of infection. Use of a risk stratification method showed that the likelihood of infection increased with each additional positive test. A combined white blood cell/sulfur colloid scan was the least predictive method of revealing infection and is not cost effective, even as part of a stratification scheme. Level of Evidence: Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.