Abstract
Background
Preoperative diagnosis of fracture related infections can be challenging, especially when confirmatory criteria such as sinus tract and purulent discharge are absent. Although serum parameters, such as CRP and white blood cell count (WBC), showed poor accuracy in the literature, they are still often used in clinical practice. The European Bone and Joint Infection Society (EBJIS) defined evidence-based criteria for fracture related infection. Elevated serum inflammatory markers were regarded as suggestive criteria only, as the literature was of limited quality. This study assessed the diagnostic value of the serum parameters CRP, WBC and differential cell count in the diagnosis of fracture related infections defined by the EBJIS-criteria for fracture related infections.
Methods
In this retrospective cohort study, 94 patients who underwent surgical treatment for suspected infected non unions after failed fracture fixation were included. Preoperatively, blood samples including serum inflammatory markers were taken. For this study, cut-offs of 5 mg/L for CRP, 10×10⁁9 cells/L for WBC, and >70% for the percentage of neutrophils were regarded as positive for infection. All patients had intraoperative samples taken for microbiology and histology. Analysis of diagnostic accuracy was based on the receiver-operating characteristic (ROC).
Results
Based on the EBJIS criteria, 40 patients (43%) were diagnosed with a fracture related infection. 11/94 (12%) patients had an elevated serum WBC count, 13/94 (14%) an increased percentage of neutrophils, and 43/82 (52%) an elevated serum CRP. The mean values of CRP concentration, WBC count, and percentage of neutrophils in the infection group were 7.9 mg/L (IQR:6.4 – 9.7), 18.3 G/l (IQR: 3.9 – 24.9), and 63% (IQR: 58 – 67%), respectively. The sensitivity, specificity, and area under the curve of serum WBC count were 20% (95% CI: 10 −35%), 94.4% (84 −99%), and 0.57 (0.50 – 0.64), respectively; of percentage of neutrophils 12.5% (5 – 27%), 85.2% (73 −93%), and 0.49 (0.42 – 0.56); and of serum CRP 67.6% (51 – 90%), 60.0% (45 – 73%), and 0.64 (0.53 – 0.74), respectively. A statistically significant difference between the AUCs of all three serum parameters and AUC of tissue culture as well as AUC of histology was shown (p <0.0001). A simple decision tree approach using only low WBC and CRP may allow identification of aseptic cases.
Conclusion
Based on the standardized and evidence-based EBJIS criteria, the three inflammatory serum markers showed an insufficient accuracy for the diagnosis of fracture related infections. They also correlate poorly with culture or histological diagnosis. Therefore, they should not be used alone as a confirmatory test.
