Spinal cord bypass surgery with intercostal and spinal accessory nerves: an anatomical feasibility study in human cadavers.

Object: Despite extensive study, no meaningful progress has been made in encouraging healing and recovery across the site of spinal cord injury (SCI) in humans. Spinal cord bypass surgery is an unconventional strategy in which intact peripheral nerves rostral to the level of injury are transferred into the spinal cord below the injury. This report details the feasibility of using spinal accessory nerves to bypass cervical SCI and intercostal nerves to bypass thoracolumbar SCI in human cadavers.

Methods: Twenty-three human cadavers underwent cervical and/or lumbar laminectomy and dural opening to expose the cervical cord and/or conus medullaris. Spinal accessory nerves were harvested from the Erb point to the origin of the nerve's first major branch into the trapezius. Intercostal nerves from the T6-12 levels were dissected from the lateral border of paraspinal muscles to the posterior axillary line. The distal ends of dissected nerves were then transferred medially and sequentially inserted 4 mm deep into the ipsilateral cervical cord (spinal accessory nerve) or conus medullaris (intercostals). The length of each transferred nerve was measured, and representative distal and proximal cross-sections were preserved for axonal counting.

Results: Spinal accessory nerves were consistently of sufficient length to be transferred to caudal cervical spinal cord levels (C4-8). Similarly, intercostal nerves (from T-7 to T-12) were of sufficient length to be transferred in a tension-free manner to the conus medullaris. Spinal accessory data revealed an average harvested nerve length of 15.85 cm with the average length needed to reach C4-8 of 4.7, 5.9, 6.5, 7.1, and 7.8 cm. The average length of available intercostal nerve from each thoracic level compared with the average length required to reach the conus medullaris in a tension-free manner was determined to be as follows (available, required in cm): T-7 (18.0, 14.5), T-8 (18.7, 11.7), T-9 (18.8, 9.0), T-10 (19.6, 7.0), T-11 (18.8, 4.6), and T-12 (15.8, 1.5). The number of myelinated axons present on cross-sectional analysis predictably decreased along both spinal accessory and intercostal nerves as they coursed distally.

Conclusions: Both spinal accessory and intercostal nerves, accessible from a posterior approach in the prone position, can be successfully harvested and transferred to their respective targets in the cervical spinal cord and conus medullaris. As expected, the number of axons available to grow into the spinal cord diminishes distally along each nerve. To maximize axon "bandwidth" in nerve bypass procedures, the most proximal section of the nerve that can be transferred in a tension-free manner to a spinal level caudal to the level of injury should be implanted. This study supports the feasibility of SAN and intercostal nerve transfer as a means of treating SCI and may assist in the preoperative selection of candidates for future human clinical trials of cervical and thoracolumbar SCI bypass surgery.