Experimental scoliosis in the rat: II. Biomechanical analysis of the forces during harrington distraction

The purpose of this study was to characterize the forces resulting from Harrington distraction of the spine in an experimental model of scoliosis in the rat, in order to establish both the similarity of this model to human scoliosis and identify potential force parameters that may be useful for clinical decision-making. Harrington distrac-tion was performed in 36 rats that had been made scoliotic 9-12 weeks earlier by the method described in the previous paper. Distractions were carried out in discrete and timed steps until separation of the vertebral laminae (mechanical failure) occurred at the upper hook site. Distractive forces were monitored continously by a strain gauge mounted on the tension side of the upper arm of the outrigger. The resulting data were compared among the various curvature groups. The relationship between the length of distraction and the maximum force produced was similar for all animals regardless of curvature. This relationship was quadratic and was characterized by an inflection point where forces increased rapidly with each distraction. The amount of distraction necessary to reach both the inflection and failure points differed only for curves above 100°. The amount of force required to reach failure was lower for curves above 75°. Curves above 50° had a lower percent correction at the inflection point. Bending and tensile forces were calculated by vector analysis. Axial load efficiencies were greater for curves above 50°, as evidenced by increased bending forces in these animals. The viscoelasticity of the spine decreased after inflection in all animals. These results indicate that the forces during Harrington distraction in this model are qualitatively and quantitatively similar to those observed during the correction of human scoliosis. The data predict that, as in humans, the risk of distraction trauma will be greatest for curves above 50°. Several parameters studied may serve clinically to predict the maximum safe correction and the advent of mechanical failure. These include values that are constant for all curvatures, such as the length and force of distraction at inflection and viscoelasticity, and values that are lower for curves above 50°, such as bending forces.