Impact of caudal segment bone fusion on rod load at the cranial segment in a two-level spinal interbody fusion...

One-dimensional implant load measurement has been validated as a method for assessing spinal fusion progression in preclinical and in vitro studies. However, multilevel fusion procedures, which are more susceptible to non-union and implant failure than single-level fusions, pose additional challenges. Specifically, the need for implantable sensors at each segment is questioned, especially in space-constrained regions such as L5-S1, where non-union and implant failure are more common. This study explores whether a single sensor at the cranial level can effectively monitor fusion progression of a caudal segment in a two-segment posterior instrumentation following TLIF surgery. It is hypothesized that adjacent level fusion has no influence on the rod load above the separating pedicle screw.

Seven human cadaveric spines (L3-S1), stabilized with posterior instrumentation following TLIF surgery at levels L4-L5 and L5-S1 were used. Bone fusion at level L4-L5 was simulated during the instrumentation. Flexibility testing and implant load measurements were conducted to compare segmental range of motion (ROM) and rod strain measured at L4-L5 between 2 states- without and with simulated bone fusion at level L5-S1. For strain measurement, an implantable sensor was used as well as one strain gauge (SG) aligned with the rod axis and one SG aligned at a 45° angle to the rod axis.

Axial strain measurements taken by the implantable sensor and the axial SG at L4-L5 during flexion-extension (FE), lateral bending (LB), and axial rotation (AR) were unaffected by L5-S1 fusion (p ≥ 0.172). However, torsional strain at L4-L5 decreased during AR following L5-S1 fusion (p = 0.018). A strong correlation was found between the relative decrease in ROM and torsional strain during AR motion (r = 0.828, p = 0.022).

In this biomechanical study the rod load measurements taken by an implantable strain sensors were unaffected by adjacent level fusion. This implies that the decrease in implant load measured by this sensor, only reflects biomechanical changes at the segment containing the sensor. The strong correlation between the reduction in AR-ROM and AR-torsional strain demonstrated the potential of an alternative assessment method of multilevel fusion. However, the interpretation of the sensor signal would be challenging, as it cannot be differentiated between the contributions of each segment to the implant load changes. Consequently, this precludes the assessment of potential different fusion stages of the individual segments.