Wednesday, September 26, 2018 7:35 AM–9:00 AM ePosters: P11. The regional effects of lumbosacral anterior column support on rod and screw strain in a pedicle subtraction osteotomy model: an in silica investigation

Abstract

BACKGROUND CONTEXT Pedicle subtraction osteotomy (PSO) for the treatment of adult spinal deformity is associated with a 20%–30% revision rate, often as the result of rod fracture at the PSO level or lumbosacral junction. Recent clinical literature reports decreased rates of instrumentation failure when multirod reconstruction or an anterior lumbar interbody fusion (ALIF) device is used at the L5–S1 disc space. While multirod techniques are shown to reduce rod strain at the PSO, it is presently unclear whether anterior column support (ACS) prevents disc collapse or reduces rod strain as a flexion blocker. Such information is necessary for determining appropriate reconstruction following PSO. PURPOSE The present study investigated whether ACS reduces rod strain, thus explaining the reduced rates of rod fracture seen clinically following PSO. STUDY DESIGN/SETTING Finite element analysis. METHODS A lumbosacral finite element model (T12-S1) was developed and validated with cadaveric range of motion data. Vertebral segments were modeled as three-dimensional solid elements. Intervertebral discs (nucleus and annulus fibrosis) were structured as hyperelastic materials. The intact model underwent PSO at L3, and titanium pedicle screws and rods from T12-S1. Simulations assessed (1) the addition of a short accessory rod affixed to the left primary rod at the PSO level (three-rod) or bilaterally (four-rod), and (2) ACS at L5-S1. A pure 10 Nm moment was applied at T12 to simulate flexion-extension. ACS was simulated by changing intervertebral disc properties to match cancellous bone. Strain was normalized to the two-rod construct (100%), averaged between the left and right primary rods, and evaluated along the primary rods at the PSO, L5-S1, and S1 screws. RESULTS Both the three- and four-rod constructs reduced primary rod strain at the PSO and L5-S1 relative to two-rod, irrespective of ACS. At the PSO, three- and four-rod constructs reduced primary rod strain to 65% and 49%, respectively; the addition of ACS had a negligible effect (three-rod, 66%; four-rod, 50%). At the lumbosacral junction, three- and four-rod constructs reduced primary rod strain to 82% and 85%, respectively; the addition of ACS dramatically reduced rod strain to 31% and 36%. Lastly, three- and four-rod constructs observed sacral screw strain of 102% and 94%, respectively; the addition of ACS reduced rod strain considerably to 34% and 34%. CONCLUSIONS While both multirod reconstruction techniques reduced motion-induced strain across the primary rods at the PSO, the four-rod construct reduced strain an additional 16% in comparison to the three-rod configuration. Lumbosacral ACS dramatically reduced L5-S1 rod strain between an additional 49% and 52% in comparison to multirod constructs without ALIF; sacral screw strain was reduced an additional 60%–68%. Nevertheless, reduction of mechanical demand of the rods and screws appears to be limited to the lumbosacral junction, with minimum influence of motion-induced rod strain at the PSO. These results suggest that the mechanical benefit of ACS may be limited to the localized level of insertion, and that reduction of rod fractures at the PSO observed clinically may be due to maintenance of disc height and subsequent sagittal balance.