Comparative bone healing with induced membrane technique (IMT) versus empty defects in septic and aseptic conditions in a novel rabbit humerus model

Background Long bone defects resulting from primary trauma or secondary to debridement of fracture-related infection (FRI) remain a major clinical challenge. One approach often used is the induced membrane technique (IMT). The effectiveness of the IMT in infected versus non-infected settings remains to be definitively established. In this study we present a new rabbit humerus model and compare the IMT approach between animals with prior infection and non-infected equivalents. Methods A 5 mm defect was created in the humerus of New Zealand White rabbits ( n = 53) and fixed with a 2.5 mm stainless steel plate. In the non-infected groups, the defect was either left empty ( n = 6) or treated using the IMT procedure (PMMA spacer for 3 weeks, n = 6). Additionally, both approaches were applied in animals that were inoculated with Staphylococcus aureus 4 weeks prior to defect creation ( n = 5 and n = 6, respectively). At the first and second revision surgeries, infected and necrotic tissues were debrided and processed for bacteriological quantification. In the IMT groups, the PMMA spacer was removed 3 weeks post implantation and replaced with a beta-tricalcium phosphate scaffold and bone healing observed for a further 10 weeks. Infected groups also received systemic antibiotic therapy. The differences in bone healing between the groups were evaluated radiographically using a modification of the radiographic union score for tibial fractures (RUST) and by semiquantitative histopathology on Giemsa-Eosin-stained sections. Results The presence of S. aureus infection at revision surgery was required for inclusion to the second stage. At the second revision surgery all collected samples were culture negative confirming successful treatment. In the empty defect group, bone healing was increased in the previously infected animals compared with non-infected controls as revealed by radiography with significantly higher RUST values at 6 weeks ( p = 0.0281) and at the end of the study ( p = 0.0411) and by histopathology with increased cortical bridging (80% and 100% in cis and trans cortical bridging in infected animals compared to 17% and 67% in the non-infected animals). With the IMT approach, both infected and non-infected animals had positive healing assessments. Conclusion We successfully developed an in vivo model of bone defect healing with IMT with and without infection. Bone defects can heal after an infection with even better outcomes compared to the non-infected setting, although in both cases, the IMT achieved better healing.