This study examines the biomechanical performance of five types of fixation techniques in a model of pathological fracture of the diaphyseal humerus.

In forty synthetic humeri, a hemi cylindrical defect centered in the middle third of the diaphysis was created. A transverse fracture was created through the centre of each defect. The bones were randomly assigned to five groups. Group A was fixed with standard ten hole DCP plates centered over the defect with five screws inserted on either end. In group B, the screw holes were injected with bone cement and then the screws and plate were reapplied while the cement was still soft. The defect was also filled with cement. Group C was fixed by injecting the cement into the entire intramedullary canal. The fracture was then reduced and the screws and plate were applied once the cement had hardened. In group D, the specimens were fixed with locked antegrade IM nail with one proximal and one distal interlocking screw. Group E was same as D except that the defect was filled with cement. Each specimen was tested in external rotation to failure by fracture.

There was no significant difference in torsional stiffness between groups B, C, and E (P> 0.16), whereas there were differences between all other groups using pairwise comparisons(p< 0.001). Groups B, C, and E were of highest stiffness followed by A and then D. Group C had the highest torque to failure, followed by groups A/B and then D/E. Total cumulative energy to failure for group C was statistically greater than each of B, D, and E (p< 0.005), but not different from A, though it approached significance (p=0.057).

This study demonstrates that, in a model of a fracture through a hemicylindrical defect in the middiaphysis of the humerus, fixation with a broad ten-hole dynamic compression plate after filling the entire medullary canal with cement is associated with the highest torque to failure and energy to failure with torsional forces. This fixation technique may best accomplish the clinical goal of maximal initial stability.