ABSTRACT
STUDY DESIGN: Thoracic pedicle screw location and the current needed to stimulate adjacent neural tissue was evaluated using brief, high-frequency pulse trains and monitoring electromyography (EMG) from muscles in the lower limbs. OBJECTIVE: To establish a safe and reliable method for detecting medial malpositioning of pedicle screws placed in the thoracic spine during instrumentation and fusion. SUMMARY OF BACKGROUND DATA: Neurophysiologic studies for testing thoracic pedicle screw placement used single-pulse stimulation and monitored EMG from thoracic-innervated muscles. We propose that with this approach, stimulation fails to activate lower motor neurons innervated by spinal cord axons, such that medial malplacement of screws will go largely undetected. METHODS: EMG was monitored from multiple lower-limb muscles. Pedicle tracks were created free-hand, using a curved pedicle finder. A ball-tipped probe-insulated along its shaft-was used to palpate the walls of the pedicle tracks. During probing, constant-current, high-frequency 4-pulse stimulus trains were delivered through the ball tip, and the minimum current (i.e., threshold) needed to evoke EMG was determined for each pedicle track. The threshold current for stimulation through each screw was also determined. Postoperative serial computed tomography scans of all implanted thoracic and L1 screws were rated with respect to screw position and the pedicle wall. RESULTS: A total of 116 screws were implanted in 7 subjects. Two pedicle tracks were redirected during surgery because of particularly low thresholds to stimulation. Definite medial defects were found in 19 screws, 18 of which were detected by the experimental technique. For these screws, the average threshold to probe stimulation of their associated pedicle tracks was 7.9 +/- 4.6 mA, much lower than current thresholds for less medially placed pedicle tracks. Stimulation of these screws resulted in high thresholds (19.8 +/- 5.3 mA) when a response was evoked at all; stimulating 8 of these 19 medially malpositioned screws failed to elicit any lower-limb EMG at considerably higher (25 or 30 mA) stimulus intensities. CONCLUSION: This preliminary study supports the hypothesis that high-frequency stimulus pulse trains areeffective at detecting defects in the medial wall of pedicles in the thoracic spine during instrumentation, thereby improving on techniques using single-pulse stimulus protocols.
