Biomechanical Testing
After each tenodesis, the humerus was amputated 5 inches distal to the fixation site. All extraneous soft tissue was dissected away, leaving the distal aspect of the biceps tendon as a free graft. Each proximal humerus–biceps tendon construct was then mounted on a materials testing machine. A custom-designed soft-tissue clamp was used to secure the distal aspect of the biceps tendon to the test actuator and load cell (Figure 3A). A custom-designed jig was used to stabilize the proximal humerus to the platform of the materials testing machine (Figure 3B). The specimens were mounted so that the line of pull throughout the testing protocol was applied parallel to the long axis of the humerus, thereby approximating the in vivo biceps force vector (Figure 3C). Digital cameras recorded each test for analysis of the mechanism of failure for each specimen. Marker dots were drawn on each tendon to assess tendon stretch before construct failure.
The tendons were preloaded to 10 N and then cycled at 0 to 50 N for 100 cycles at 1 Hz. After cyclic loading, axial load to failure was performed at a rate of 1.0 mm per second until a peak load was observed and subsequent loading led to tendon elongation with no further increase in load. Displacement and force applied during cyclic loading and load to failure were recorded. Stiffness was then calculated as the slope of the linear portion of the force-displacement curve using the least mean squares approach. Mechanism of failure was documented for each specimen.
Statistical Analysis
Analyzing our preliminary data with G*Power, we determined that a total sample size of 8 would be required (effect size [Cohen dz] was 1, α error probability was .05, power was .8). We hypothesized that the ultimate strength and stiffness of one group would be less than 1 SD above those of the other group. Paired t test with significance set at P < .05 was used to compare the techniques.
Results
Both repair constructs exhibited standard load-displacement curves, with a linear increase in load with displacement until the point of failure, at which time further displacement occurred with no discernible increase in load (Figure 4). Ultimate load to failure is determined as the highest point of the curve, and stiffness is calculated as the slope of the load-displacement curve.
Mean axial displacement parallel to the shaft of the humerus with cyclic loading was 7.1 mm with the modified PITT technique and 7.9 mm with the interference screw technique. There was no macroscopically visible high-grade tearing or slippage of the biceps tendon in any specimen with cyclic loading for either repair construct.
Mean (SD) ultimate load to failure was significantly (P = .003) higher with the modified PITT technique, 157 (41) N, than with the interference screw technique, 107 (29) N; actual effect size (Cohen dz) was 1.19, difference of means was 50 N, and pooled SD was 42 N. The interference screw technique yielded significantly (P = .010) more mean (SD) stiffness, 38.7 (14.7) N/mm, than the modified PITT technique, 15.8 (9.1) N/mm; actual effect size (Cohen dz) was 1.37, difference of means was 22.9 N/mm, and pooled SD was 16.7 N/mm.
In the interference screw technique, the mode of failure was consistent. Of the 8 specimens, 7 failed at the screw–tendon interface at the distal aspect of the tunnel; in the eighth specimen, the entire tendon pulled out from under the screw construct. In the modified PITT technique, there was more variability in failure: tendon slipped through suture (4 specimens), tendon/suture construct as unit pulled through transverse ligament/rotator interval tissue (3), and suture failure (1).