RESUMO
BACKGROUND: Intramedullary nailing (IMN) has been frequently indicated to treat long bone open and closed fractures, but IMN infection (IMNI) may have devastating consequences. Sonication has been regarded as an important add-on for microbial identification on a variety of orthopaedic implant-associated infections, but its role in the IMNI is poorly studied. We aim at evaluating the accuracy obtained by conventional peri-implant tissue culture (TC) samples with sonication fluid cultures (SCs) of IMNI. METHODS: Longitudinal prospective cohort study ongoing since June 2014, which included patients with indication for IMN removal due to any reason. Clinical diagnosis of INMI was defined according to publication addressing fracture-related infections. Minimal of two samples from TC were cultured. SCs followed the protocol previously published. Statistical analysis was performed using McNemar's test for related proportions. RESULTS: We included 54 patients submitted to IMN retrieval, of whom 47 presenting clinical signs of IMNI. Sensitivity for detecting microorganisms using TC and SC was 89.4% (42/47) and 97.6% (40/41), and specificity was 71.4% (5/7) for both TC and SC (p = 1.00). Positive and negative predictive values for TC and SC were 95.5% (42/44), 95.2% (40/42), 50% (5/10), and 83.3% (5/6), respectively. The most frequent organisms isolated in both TC and SC were Staphylococcus aureus, S. epidermidis, and Enterococcus sp. Polymicrobial infection was diagnosed in 14.8% (8/54) and 25% (12/48) by TC and SC, respectively (p = 0.19). CONCLUSION: Sonication fluid and tissue samples presented similar accuracy on the diagnosis of IMNI, but SC was advantageous of detecting polymicrobial infection.