Abstract
AIMS: Nonunion after foot and ankle arthrodesis is a major complication, which can lead to revision arthrodesis with increased risk of morbidity. Multiple factors can contribute to developing a nonunion, including a low-grade infection. The aim of this study is to investigate the rate of unsuspected low-grade infection in revision surgery for nonunion after foot and ankle arthrodesis. We also analyzed the difference in outcome of revision arthrodesis for patients with or without infection. METHODS: We conducted a retrospective study in the Sint Maartenskliniek, the Netherlands. All patients who underwent a revision arthrodesis (index surgery) for assumed aseptic nonunion after foot and ankle arthrodesis between January 2019 and July 2021 were included. Patients were excluded if fewer than five tissue samples were obtained during index surgery, or if infection was suspected before or during the index surgery. The causative microorganisms, antibiotic susceptibility, treatment, and outcome were assessed. RESULTS: In total, 105 revision arthrodeses due to nonunion were performed, of which 77 patients were included. In 17 patients (22.1%), an infection was diagnosed based on culture results of tissue samples taken during index arthrodesis. The causative bacteria identified were Cutibacterium acnes (n = 11), Staphylococcus spp. (n = 7), and Acinetobacter spp. (n = 1). After two years of follow-up, the success rate of the infection group was 77% compared with 88% for the patients with negative cultures (p = 0.451). CONCLUSION: Low-grade infection should be considered as a possible explanation of a nonunion, despite the lack of local inflammatory signs. As 22% of the nonunions were unexpectedly caused by low-grade infection, we strongly recommend obtaining at least five tissue samples for culture during revision arthrodesis. The success rate of revision surgery followed by targeted antibiotics in patients with a low-grade infection is 77%, which is lower than the outcome in revision surgery for patients without infection (88%), although not significant (p = 0.451).