The use of evoked EMG in detecting misplaced thoracolumbar pedicle screws

Study Design. Experimental study performed using an animal model. Objectives. To determine if EMG responses generated by the electrical stimulation of thoracolumbar pedicle screws could be used to predict the screw position. Summary of Background Data. Evoked EMG has been used successfully to predict pedicle screw position in the lumbar spine. No data have been published on its effectiveness in the thoracic spine. Methods. A total of 91 screws were inserted into the pedicles from T8 to L2 in six sheep. Monitoring electrodes were placed into transversus abdominus at three levels, the lower two intercostal spaces, and into psoas. A constant voltage stimulus was applied to a probe inserted into each pedicle, and then to each pedicle screw after it had replaced the probe. The threshold voltage required to evoke EMG activity in the relevant myotome was noted. After monitoring the position of each screw was determined by gross dissection. Results. EMG responses in abdominal and intercostal muscles were successfully evoked by thoracic pedicle screw stimulation. Of the 91 screws, 50 were within the pedicle and required an average voltage of 15.12 V to stimulate an EMG response, compared with the 41 misplaced screws that had an average voltage of 7.53 V (P < 0.0001). Using a threshold of 10 V the technique has a sensitivity of 94% and a specificity of 90%, Conclusion. Electrical stimulation of pedicle screws and EMG recording in abdominal and leg muscles in sheep provide a reliable indication of pedicle screw position. This technique can be directly applied to human thoracolumbar surgery, but differences in pedicle size would mean that new threshold voltage criteria would need to be established.