Stimulus-evoked electromyography testing of percutaneous pedicle screws for the detection of pedicle breaches: a clinical study of 409 screws in 93 patients.

OBJECT

Percutaneous pedicle screws have recently become popularized for lumbar spinal fixation. However, successful anatomical hardware placement is highly dependent on intraoperative imaging. In traditional open surgery, stimulus-evoked electromyography (EMG) responses can be useful for detecting pedicle screw breaches. The use of insulated sleeves for percutaneous screws has allowed for EMG testing in minimally invasive surgery; however, no reports on the reliability of this testing modality have been published.

METHODS

A total of 409 lumbar percutaneous pedicle screws were placed in 93 patients. Levels of instrumentation included L-1 (in 12 patients), L-2 (in 34), L-3 (in 44), L-4 (in 120), L-5 (in 142), and S-1 (in 57 patients). Intraoperative EMG stimulation thresholds were obtained using insulating sleeves over a metallic tap prior to final screw placement. Data were compared with postoperative fine-cut CT scans to assess pedicle screw placement. Data were collected prospectively and analyzed retrospectively.

RESULTS

There were 5 pedicle breaches (3 medial and 2 lateral; 3 Grade 1 and 2 Grade 2 breaches) visualized on postoperative CT scans (1.2%). Two of these breaches were symptomatic. In 2 instances, intraoperative thresholds were the sole basis for screw trajectory readjustment, which resulted in proper placement on postoperative imaging. Thirty-five screw trajectories were associated with a threshold of less than 12 mA. However, all breaches were associated with thresholds of greater than 12 mA. Using thresholds below 12 mA as the indicator of a screw breach, this resulted in a sensitivity of 0.0, specificity of 90.3, positive predictive value of 0.0, and negative predictive value of 0.98. Utilizing a threshold of any decreased stimulus (< 20 mA) would have detected 60% of breaches, with a mean threshold of 16.25 mA.

CONCLUSIONS

While these data are limited by the low number of radiographic breaches, it appears that tap stimulation with an insulating sleeve may not be reliable for detecting low-grade radiographically breached pedicles using typical stimulation thresholds (< 12 mA). Imaging-based modalities remain more reliable for assessing percutaneous pedicle screw trajectories until more robust and sensitive electrophysiological testing methods can be devised.