Neuromonitoring with pulse-train stimulation for implantation of thoracic pedicle screws: a blinded and randomized clinical study. Part 2. The role of feedback

Abstract

Object The authors have reported in Part 1 of this study on a novel neuromonitoring test for the prevention of medial malpositioning of thoracic pedicle screws. In the present paper they examine the impact of providing the results of the test as intraoperative feedback to the surgical team. Methods This is the second part of a 2-part report of a prospective, blinded and randomized neuromonitoring study designed to lower the incidence of medially malpositioned thoracic pedicle screws. Details of the neuromonitoring technique and data supporting the alarm criteria used are contained in the companion article (Part 1). For the majority of pedicle screw placements, intraoperative test results were withheld from the study team (that is, the team members were blinded to the test results). However, for a limited number of pedicle sites the authors provided one of 2 forms of testing feedback to the surgical team: 1) “break the blind” feedback, if testing suggested that screw placement would result in direct contact between screw and the dura mater; and 2) “planned” feedback, beginning during the later stages of the study and provided for 50% of pedicle sites. Feedback gave the surgeon the opportunity to adjust the trajectory that the screw would ultimately take within the pedicle. The final screw position relative to the pedicle's medial wall for all sites in which feedback was withheld from the surgical team was compared with the screw position for those sites in which either form of feedback (“break the blind” or “planned”) was provided to and acted upon by the surgical team. Results Of the 820 pedicle tracks tested among the 71 surgical cases included in this study, a total of 684 were operated upon without any form of feedback. Planned feedback was provided for an additional 107 pedicle tracks, of which 15 triggered an intraoperative alarm (evoked electromyogram response in leg muscles to stimulus intensity ≤ 10 mA) leading to a warning to the surgical team of a medially biased pedicle track. Finally, the blind was broken 29 times, in each case when testing revealed a particularly low threshold (≤ 4 mA) for evoked responses in leg muscles when stimulating along the pedicle track with the ball-tipped probe. As detailed in the companion paper to this one, there were 32 screws with threads lying at least 2 mm medial to the pedicle wall. In all 32 instances (100%), either these screws were in the “no feedback” category (n = 29) or they were in a feedback category but the surgeon elected to not revise the pedicle-track trajectory. Two patients returned to the operating room for revision of screw placements because the screws were encroaching upon the central canal; the pedicle tracks for these screws had been in the “no feedback” category. Conclusions This is the first blinded and randomized study to prove that implementing a novel neuromonitoring strategy during placement of thoracic pedicle screws can significantly reduce the incidence of clinically relevant thoracic pedicle screw medial malpositioning.