MECHANICAL TESTING
Mechanical testing was performed on a biaxial 370.02 Bionix Testing System (MTS Systems Corp.). The femur was rigidly and horizontally secured to a custom-built test frame, and the lower leg was left free to move. The quadriceps tendon was secured in a freeze clamp and was attached to the MTS actuator for loading via a pulley system such that raising the actuator was translated into a simulated quadriceps extensor force.
A differential variable reluctance transducer (DVRT) (Lord MicroStrain) was placed across the osteotomy site to measure the distraction, or gap, across the fracture line. The minimum load to full extension for each specimen was then determined under a slow, controlled increase in load until the leg was in a fully extended position. Any distraction across the fracture line during the initial loading phase was determined by using digital calipers. The specimen was then subjected to a preconditioning phase with 10 cycles from 0 N to full extension under the previously determined load, which was applied at the rate of 5 N/s. Meanwhile, displacement across the fracture site was recorded via the DVRT. Following the preconditioning phase, each specimen was then tested to failure in displacement control at the rate of 1.5 mm/s. Failure was defined as implant failure (screw pullout) or DVRT gapping across the osteotomy site >3 mm.10,14
Outcome measures included stiffness (N/mm), which was calculated as the slope of the linear change in load from full extension to failure vs DVRT displacement during the final loading phase; failure load (N); gapping (mm) across the osteotomy site at each cycle during the preconditioning phase; and failure mode (pullout vs >3.0 mm gap).
STATISTICAL ANALYSIS
An a priori power analysis revealed that 13 knees per group would be required to obtain an α of 0.05 and a power of 0.80. This calculation was based on a 20% difference in fracture displacement calculated by using the standard deviation and mean previously reported for cannulated screws with nonabsorbable sutures.14
Means and standard deviations for all dependent outcome measures were computed and compared across the independent measure of fixation type (BF vs SF) through repeated measures Analysis of variance (ANOVA-GLM, SAS 9.3, SAS Institute, Inc.) after controlling for bone mineral density (BMD), gender, and age. Multivariate repeated-measures ANOVA with Tukey's studentized range was applied to cyclic gap data. The mode of failure was compared across fixation type (BF vs SF) for matched data using McNemar’s test. Intracorrelations were computed and examined over all data and separately on the basis of screw fixation type (BF vs SF). All tests were considered statistically significant when P < .05.
Continue to: Results...