MISPLACEMENT OF LUMBAR PEDICULAR SCREWS THAT PRODUCE POSTOPERATIVE MECHANICAL RADICULALGIA WITH NORMAL INTRAOPERATIVE NEUROPHYSIOLOGICAL STIMULATION : THE VALUE OF NEUROPHYSIOLOGICAL STIMULATION AT THE PEDICULAR MID-TRACK

Objective: To carry out a neurophysical evaluation that can identify these cases during surgery, and that prevents misplacement of pedicular screws. Methods: A total of 6739 screws were placed in 293 patients submitted to scoliosis correction via the posterior route with pedicular screws using the freehand technique. Of this total, eight patients (2.7%), with a mean age of 24 years, developed postoperative radiculopathy. Lumbar CT scans showed 10 misplaced lumbar screws (2L1-3L2-4L3-1L4), with minimal protrusion of the lower part of the screw. EMG thresholds of the screw and track were evaluated. Results: In the initial surgery, no anomalies were detected on palpation of the screw track or in the radioscopic control; neither were any neurophysiological alterations detected in the neurophysiological stimulation with t-EMG. All the patients had radicular pain in the standing and seated positions, which disappeared when lying on the bed. The screws were surgically removed at an average of 37 days after surgery (range: 4-182). In this surgical procedure, the neurophysiological monitoring was repeated, again showing normal thresholds (>11mA). After screw removal, stimulation of the probe within the track showed very low thresholds (range: 3.9-10.7 mA) at mid pedicular track. After a mean follow-up time of 4.4 years (Range: 2.6-6.8), five patients reported occasional radicular discomfort, and had minimal motor deficit in the affected limb. Conclusions: A type of misplacement of lumbar pedicle screws is presented that produces radicular pain in the standing and seated positions, and that may not be detected by conventional monitoring. Neurophysiological stimulation of the mid pedicular track, after removal of the screw, produces low stimulation thresholds. Systematic stimulation of the track prior to insertion of the lumbar pedicular screw is recommended.


INTRODUCTION
Surgical treatment is an accepted method for correcting and preventing the progression of deformity of the thoracolumbar spine.However, the possibility of radiculomedullary neurological complication during surgery is a major cause of concern for surgeons.
Neurological complications can occur at the moment of insertion of the screws or sublaminar wires/hooks, in the process of neural release or bending of the spine, or at the moment of reduction of the deformity.
The occurrence of pedicle screw misplacement in surgery for deformity of the thoracolumbar spine ranges from 6% to 30%, 1 usually without any neurological consequences.
In larger case series, an incidence of radicular complications during surgical treatment of spinal deformity of between 0.41% and 2.24% has been reported, despite the existence of correct neurophysiological monitoring and other controls. 2he rate of radicular complications caused by misplacement of pedicle screws that escaped detection by conventional neurophysiological and imaging systems is reported to be 3.38%. 3.4 Multiple intraoperative controls are recommended, such as the technique of pedicle insertion and intraoperative freehand palpation, neurophysiological control with somatosensory evoked potentials (SSEP) and motor evoked potentials (PEM) and monitoring of screws (t-EMG), and finally, intraoperative radioscopic control 5 in patients undergoing surgery for deformity of the thoracolumbar spine.
In relation to the pedicle screw stimulus-evoked thresholds (t-EMG) introduced by Calancie et al. 6 in 1991, there is some discrepancy between authors when it comes to determining the pathological values that suggest perforation of the medial cortex of the lumbar pedicle and close contact with the nerve root.Glassmann et al. 7 suggest lower thresholds of 10 mA as highly suggestive of cortical perforation.Toleikis et al. 8 suggest a threshold of <5 mA, and Raynor et al. 5 present a stimulus-evoked threshold of 8 mA as unacceptable, due to the possibility of medial perforation of the lumbar pedicle.
According to the scientific evidence and our surgical experience, there are a number of patients with postoperative radicular complications that go undetected despite multiple controls.
The objective of this work is to analyze a group of patients with postoperative pain and/or motor deficit of the lower limbs following surgical correction of spinal deformity, whose symptoms occur exclusively when in the standing and/or seated positions, seeking to improve the neurophysiological detection of misplaced lumbar screws that go undetected in the usual intraoperative controls (freehand pedicle insertion technique, t-EMG and intraoperative imaging control).

METHODS
We studied 294 patients who underwent surgery for spinal deformity of multiple etiologies between 2004 and 2010, in three hospitals.All the surgical operations were performed by a primary surgeon with wide experience in surgical correction of deformities, assisted by several orthopedic surgeons with experience in spinal surgery, who were actively involved in the insertion of the pedicle screws, particularly on the right side of the patient.
A surgical technique by the posterior approach was used, essentially only with screws.
A total of 6,765 pedicle screws were inserted in these patients: 1,550 lumbar, 5,088 thoracic, and 127 iliac screws.
The screws were inserted freehand in most cases.All patients underwent intraoperative neurophysiological monitoring with potentials (SSEP and MEP) and monitoring of the screws (t-EMG).All patients underwent an intraoperative imaging exam at the end of the surgery, to confirm the correct positioning of the screws.
The neurophysiological study was carried out with a specific intraoperative monitoring Keypoint machine (Alpine Skovlunde-Denmark).The pedicles were stimulated with a pedicle screw Probe (Viasys Healthcare, Madison WI), used as a cathode that sits on the screw head, and a monopolar needle of 12 mm in length (Neuroline Subdermal, Ambu), which was inserted into the paraspinal musculature, acting as an anode.
A constant stimulus current was applied with a frequency of 1 Hz and duration of 0.1 ms, with intensity increasing from 0.2 mA to 50 mA until a repeatable compound action potential was obtained in the muscle of the corresponding myotome.The responses were recorded using filters at between 20Hz and 5 KHz, with amplifier sensitivity 50 μV per division and 10 ms of stimulation pulse.A pair of stainless steel monopolar needle recording electrodes, 12 mm in length (Neuroline Subdermal, Ambu), were inserted into the muscles of the corresponding myotomes.
The acceptable thresholds, according to the recommendations of Glassman et al., 7 were threshold >15 mA: 98% confidence for the intrapedicular position of the screw.Threshold 10-15 mA: 87% confidence for the intrapedicular position of the screw.Threshold <10 mA: perforation of the cortex in 90% of cases.In the screws with thresholds below 15 mA, the track was reviewed in all cases and the screw was definitively inserted, according to the palpation assessment, particularly if the threshold was between 10-15 mA and it was considered a fundamental screw.
Statistical analysis was performed using the software program SPSS v 1800 (IBM Corp, Armonk, NY) and the chi-square test of independence for the study of the association between "surgeon's experience" and "outcome of the implant".
Of the 294 patients who underwent surgery, 8 (2.7%) patients with a mean age of 24 years (11-39) developed postoperative radicular pain/motor deficit.An axial computed tomography (CT) was requested in these patients, which showed that 10 (0.67%) malpositioned lumbar screws (2L1-3L2-4L3-1L4), all of them with minimal prominence in the lower region of the pedicle.(Figure 1) The malpositioned screws were removed, and the EMG thresholds of the screw and pedicle track were assessed during the surgery.

RESULTS
The eight patients with postoperative radiculopathy presented palpation of the track, t-EMG and intraoperative radioscopy without pathological changes during the first surgery.Postoperative CT showed, in all cases (10 screws) misplacement of the lumbar screw, with invasion <2 mm in the lower region of the pedicle.The screws were removed without replacement at 37 days, on average (4-182).
Prior to removal of the screw, a new stimulation was performed on the malpositioned screw, which once again showed normal values (>11 mA).After withdrawal of the screw, stimulation of the pedicle track was performed, showing very low values in its mid portion (3.9 -10.7 mA) and normal values (>11 mA) at the bottom of the pedicular track (Figures 2 and 3).
After a mean follow-up of 4.4 years (2-6.8years), two patients reported moderate pain in the affected root, one reported mild motor deficit of the quadriceps (L3D), one patient had paresthesia, and four patients were asymptomatic.
The malpositioned screws mainly occurred on the right side of the patient (9/1), a fact that presents a statistically significant relationship, with p<0.05 (chi-square).

Table 1 .
Data on patients with malpositioned screws.