OBJECTIVETo evaluate the optimal insertion position of the Coflex lumbar interspinous dynamic stabilization device.

METHODSSix fresh adult human cadaveric lumbar spine specimens (L1-L5) were mounted in a materials testing machine by embedding to clamps with L1 and L5 vertebrae. L3-4 motion segment of each specimen was operated by selective decompression and Coflex interspinous device insertion. The L3 and L4 vertebrae was inserted one needle attached with four marker points respectively, which were used to record the range of motion (ROM). Each lumbar spine specimen was tested according to the loading sequence at 5 groups: intact (keeping lumbar ligamenta and facet joints intact) group, partial destabilized (resection of L3-4 interspinous ligamenta, ligamentum flavum, facet capsule, and bilateral resection 50% of L3 inferior facets) group, 10 mm insertion (distance between apex of U-shaped Coflex and dural sac was 10 mm)group, 5 mm insertion (distance was 5 mm)group, and 0 mm insertion (distance was 0 mm)group. Each lumbar spine specimen was tested repeatedly 3 times according to a loading sequence consisting of flexion, extension, left/right lateral bending, left/right axial rotation, loaded with pure moments of 8 N·m, and was recorded the ROM of operative segment at the third time. ROM of 5 groups in 6 directions respectively were analyzed with one-way ANOVA test and multiple comparisons were based on LSD method.

RESULTSThe means ROM of 5 groups were not all equal in flexion, extension, left/right lateral bending, left/right axial rotation (F = 8.472, 18.301, 7.700, 12.473, 16.809, 6.624; all P < 0.01). The 10 mm insertion group had significant high ROM in 6 directions than the intact group (t = 3.80, 3.82, 4.49, 5.60, 4.96, 2.98, all P < 0.01), but it was no difference comparing with the partial destabilized group (P > 0.05). The ROM of the 5 mm and 0 mm insertion group were no significant differences comparing with the intact group in flexion, extension, left/right axial rotation (P > 0.05), but it were significant differences comparing with the partial destabilized group in the same directions (5 mm insertion group: t = 3.19, 6.34, 5.26, 3.43, all P < 0.01; 0 mm insertion group: t = 4.21, 6.68, 5.81, 3.72, all P < 0.01). There were significant differences in the ROM of left/right lateral bending between the 5mm/0mm insertion groups and the intact group (5 mm insertion group: t = 3.71 and 5.22, all P < 0.01; 0 mm insertion group: t = 3.44 and 4.95, all P < 0.01), but there were no differences comparing with the partial destabilized group in the same directions (P > 0.05).

CONCLUSIONSThe insertion of Coflex interspinous dynamic stabilization device can maintain the stability of a partially destabilized specimen back to an intact one in flexion, extension and axial rotation when distance between apex of U-shaped Coflex and dural sac was ≤ 5 mm, but can't return the stability in lateral bending. The Coflex can't return the stability of a partially destabilized specimen back to an intact one in 6 directions when distance between apex of U-shaped Coflex and dural sac was ≥ 10 mm.