Session VII Pelvis and Acetabulum

Intraoperative Stimulus-Evoked Emg Monitoring for Placement of Iliosacral Screws: An Animal Model and Initial Clinical Results

Berton R. Moed, MD, Mark Anders, MD, B. K. Ahmad, MD, Joseph Craig, MD, Gary P. Jacobson, PhD

Henry Ford Hospital, Detroit, MI

Purpose: To develop and then evaluate the efficacy of stimulus-evoked EMG monitoring - a technique never previously applied to pelvic fracture surgery - as a method for minimizing the risk of nerve root injury during the placement of iliosacral screws for acute pelvic fracture fixation.

Materials and Methods: First a canine animal model was designed to evaluate the feasibility of stimulus-evoked EMG monitoring of the lumbosacral nerve roots. Under general anesthesia 2.5mm diameter K-wires were percutaneously inserted using C-arm fluoroscopy image-intensifier control into the S1 body in 3 large adult mongrel dogs. Stimulus-evoked EMG monitoring was utilized with the K-wire acting as the live electrode. A searching current of twenty milliamps was initially applied to the K-wire with monitoring electrodes placed in the gastrocnemius muscle. Four K-wires were inserted across each of five of the sacroiliac joints into the S1 body with an actual attempt made to compromise the S1 canal and the S1 nerve root. Minimum milliamp thresholds required to evoke an EMG response in the gastrocnemius muscle were recorded for each K-wire. The dogs were subsequently euthanized and actual K-wire location was first estimated by CT scan and then directly documented by dissection and direct visualization of K-wires and nerve roots.

Results: Evaluation of the 20 percutaneously inserted iliosacral K-wires revealed that 4 scored direct hits on the nerve root. All 4 had a stimulus threshold of less than 4 milliamps. K-wires within 2mm of the nerve root had thresholds of 4 to 6 milliamps. K-wires located approximately 5mm from the nerve root required a stimulus of greater than 6 but less than 8 milliamps. Those at greater distance from the nerve root required more than 8 milliamps of stimulus to evoke an EMG response.

As a result of the success of this canine feasibility study, the stimulus-evoked EMG monitoring method was applied for a subsequent consecutive clinical series of 18 patients, all with Type C unstable injuries to the pelvis. Thirty-five iliosacral screws were inserted in these 18 patients with 20 unstable posterior injuries (2 bilateral). A threshold of 8 milliamps was selected as a safe zone. Based on this criteria 3 of the 35 screws were redirected as a result of stimulus-evoked EMG monitoring. After the insertion of the screws using the stimulus-evoked EMG monitoring method, SSEP's were also obtained to document the status of the nerve roots prior to awakening the patient from anesthesia. There were no iatrogenic nerve injuries in this series.

Discussion: The incidence of iliosacral screw malposition and resultant nerve root injury during the reduction and fixation of a posterior pelvic ring injury has been reported to be as high as 10%. Currently available monitoring systems, either SSEP or spontaneous EMG, though useful, record abnormal signals only after injury, transient or permanent, has occurred. Intraoperative stimulus-evoked EMG monitoring has been successfully used during facial nerve surgery and has recently been applied for use during transpedicular lumbosacral spine instrumentation. Stimulus-evoked EMG monitoring if technically possible, could provide an early warning of impending nerve root injury allowing the surgeon to alter the direction of the wire, drill bit or screw insertion and avoid nerve root injury.

Conclusion: On the basis of the results of our animal and clinical studies, it appears that stimulus-evoked EMG monitoring provides reliable data indicating the proximity of the iliosacral implants to the lumbosacral nerve roots. This inexpensive, rapid and easily implemented method of intraoperative monitoring may significantly decrease the risk of iatrogenic nerve root injury.