Abstract
Introduction: Diaphyseal fracture at a cortical perforation is the commonest postoperative complication of hips revised with impacted morsellised bone. To reduce fracture risk, surgeons can apply mesh, augment the bone with plate or strut graft, or bypass the perforation with a longer stem. No biomechanical data exists to choose between these alternatives. The objective of this study was to compare the above methods of cortical repair in terms of (i) bone fracture risk and (ii) stem migration.
Methods: Fourteen large composite femora (Sawbones, Malmö, Sweden) were prepared to simulate cavitary defects. An 18×40 mm lateral cortical perforation was made in 12 diaphyses. These diaphyses were repaired with mesh only, mesh and plate, or mesh and strut graft (n=4 each). Strut graft and plate were fixed with cables. Porcine cancellous bone was morsellised and impacted into each cavitary defect. Simplex P bone cement was injected. In the 12 femora with repaired perforation, a standard or a long Exeter prosthesis, bypassing the perforation 2 cortical diameters, was implanted. Thus, 6 methods of defect repair were created (mesh, plate and strut, combined with either long or short stem, each n=2). Standard stems impaction-grafted in the two femora without perforation served as control (n=2). Femora were placed in a testing machine and loaded at 1 Hz with 100 cycles of joint and abductor force. Peak joint force was 2,500 N. Strain amplitudes at the perforation and stem migration were determined. Statistical analysis was by 2-way and 1-way ANOVA, and the Student-Newman-Keuls (SNK) post-hoc test.
Results: Stem length did not affect average defect strain if used with plate or strut graft (2x2 ANOVA, p=0.62). Four combinations remained for further analysis: standard stem with mesh, long stem with mesh, plate, and strut graft, with defect strains of 5250, 3620, 2940, and 2480 μstrain. In controls, strains were 1750 μstrain. Defect strains differed significantly (ANOVA, p=0.0004), with strains for standard stems with mesh significantly higher than all other groups, those for long stems with mesh significantly higher than controls, and those for plate or strut graft no different from controls (SNK). Maximum permanent subsidence was 0.71 mm and retroversion 1.6°. For repaired perforations, stem length did not affect subsidence (p=0.96), but repair method did (p=0.03, both 2-way ANOVA). For further analysis, subsidence of the three repair methods (mesh, plate and strut graft with subsidence of 0.24, 0.47 and 0.19 mm, resp.) was compared with that of controls (0.52 mm). Subsidence differed significantly (ANOVA, p=0.02), and stems with strut graft subsided significantly less than those with plate or controls (SNK). Permanent retroversion was similar for each group.
Dicussion: Non-reinforced defects with a standard stem generated high defect strain amplitudes. A long stem bypassing the defect reduced these strains by 30%, and might suffice in case of otherwise strong cortex. In other cases, augmentation of the perforated diaphysis with either strut graft or plate needed to minimise fracture risk. Stem migration in reconstructed perforated diaphyses was always less than control cases, suggesting stem migration is no specific problem in reconstruction.
Correspondence should be addressed to Dr Carlos Wigderowitz, Honorary Secretary of BORS, Division of Surgery & Oncology, Section of Orthopaedic & Trauma Surgery, Ninewells Hospital & Medical School Tort Centre, Dundee, DD1 9SY.
