High diagnostic accuracy of white blood cell scintigraphy for fracture related infections: Results of a large retrospective single-center study

Introduction: White blood cell (WBC) scintigraphy for diagnosing fracture-related infections (FRIs) has only been investigated in small patient series. Aims of this study were (1) to establish the accuracy of WBC scintigraphy for diagnosing FRIs, and (2) to investigate whether the duration of the time interval between surgery and WBC scintigraphy influences its accuracy. Patients and methods: 192 consecutive WBC scintigraphies with Tc-99m-HMPAO-labelled autologous leucocytes performed for suspected peripheral FRI were included. The golden standard was based on the outcome of microbiological investigation in case of surgery, or - when these were not available - on clinical follow-up of at least six months. The discriminative ability of the imaging modalities was quantified by several measures of diagnostic accuracy. A multivariable logistic regression analysis was performed to identify predictive variables of a false-positive or false-negative WBC scintigraphy test result. Results: WBC scintigraphy had a sensitivity of 0.79, a specificity of 0.97, a positive predicting value of 0.91, a negative predicting value of 0.93 and a diagnostic accuracy of 0.92 for detecting an FRI in the peripheral skeleton. The duration of the interval between surgery and the WBC scintigraphy did not influence its diagnostic accuracy; neither did concomitant use of antibiotics or NSAIDs. There were 11 patients with a false-negative (FN) WBC scintigraphy, the majority of these patients (n = 9, 82%) suffered from an infected nonunion. Four patients had a false-positive (FP) WBC scintigraphy. Conclusions: WBC scintigraphy showed a high diagnostic accuracy (0.92) for detecting FRIs in the peripheral skeleton. Duration of the time interval between surgery for the initial injury and the WBC did not influence the results which indicate that WBC scintigraphy is accurate shortly after surgery. (C) 2018 Elsevier Ltd. All rights reserved.