Since Dawson demonstrated it in humans in 1946-1947 monitoring SSEPs during spinal surgery has become the norm rather than an exceptional available tool. It has spread from high risk deformity correction surgeries, such as scoliosis, to more mundane and typical procedures like fusions for trauma or degenerative disease.
Neuromonitoring during spinal surgery has become a huge industry and while population sized reports show a 50% decrease in neurological complications since monitoring came into wide use not all monitoring appears created equal and prospective studies trying to tease out the benefit of costly neuromonitoring have been inconclusive at best.
One of the newer additions to neuromonitoring is evoked EMG. And here is where the paper by Montes, et al out of the Hospital Ramon y Cajal in Madrid. In evoked EMG you stimulate something, usually the pedicle screw in spinal fusions, and see if the stimulation is conducted to neural tissue. The lower the level of stimulation required supposedly the closer the screw is to damaging a neural element (the spinal cord or a nerve, etc). One of the primary fears when placing these pedicle screws is placing the screw so that it is out of bone and in the spinal canal. It is a long held belief with EMG in neuromonitoring that stimulating the screw might give you evidence if you’ve broken through the bone medially. If the stimulation of the screw evoked a response at a low enough threshold then you removed the screw and replaced.
In an animal model the pediatric orthopedic practice in Spain wanted to see if there was a relationship between the stimulation needed to evoke a motor response in each screw and the integrity of the medial wall of the pedicle.
Montes, et al used a porcine model. They placed a total of 18 viable screws in the thoracic columns of anesthetized pigs. The placed screws were measured and found 8mm from the spinal cord. With leads in the intercostal muscles the screws were stimulated in this position, 8mm from the cord with the medial pedicle bone intact, and the threshold at which stimulation of the muscle was achieved was recorded. The team then placed different materials in the canal, between the medial border of the pedicle (and the screw) and the dural sac. This organic material included bone, fat and muscle. With these interspersed tissues the EMG thresholds were again recorded. The team then removed the screws, took off some medial pedicle bone creating a “breach” and then replaced the screw so that it was 2mm from the spinal cord. EMG thresholds recorded the team then repeated the thresholds with fat, bone and muscle between the screw and the dural sac.
The only association with the threshold for stimulation was the distance from the spinal cord.
The tissues between the screw made no difference. So whether the medial pedicle wall was intact or not had no effect at what stimulation there was a response in the muscle.
It is true that other in vitro studies have shown that stimulation relates to the impedence from the screw to the neural elements (ie the tissues between the screw and the spinal cord) and the authors provide no answer for the discrepancy.
Taken alone these findings argue against somewhat the usefulness of evoked EMG in spinal fusion surgery in some ways. EMG cannot really give us a high sensitivty for breach of the pedicle cortex.
All EMG leaves us with is to have a very high tolerance for borderline thresholds on stimulation. The screws that should be replaced are those that have very, very low thresholds on EMG implying that they are touching or nearly touching neural elements. That in and of itself may still be of some use as its probably that proximity, more than the violation of the bone, that most puts patient at risk for injury (ie, even if there is no intervening tissue between the screw and the cord the fact that there’s some distance is safe in and of itself).