Failure of Intraoperative Monitoring to Detect Postoperative Neurologic Deficits: A 25-year Experience in 12,375 Spinal Surgeries.

STUDY DESIGN: Retrospective. OBJECTIVE: The purpose was to categorize and evaluate intraoperative monitoring (IOM) failure to detect neurologic deficits occurring during spinal surgery. SUMMARY OF BACKGROUND DATA: The efficacy of spinal cord/nerve root monitoring regarding undetected neurologic deficits is examined in a large, single institution series involving all levels of the spinal column and all spinal surgical procedures. METHODS: Multimodality IOM included somatosensory-evoked potentials (SSEPs), descending neurogenic-evoked potentials (DNEPs), transcranial motor-evoked potentials (MEPs), dermatomal somatosensory-evoked potentials (DSEPs), and spontaneous and triggered electromyography (spEMG, trgEMG). We reviewed 12,375 patients who underwent surgery for spinal pathology from 1985 to 2010. There were 7178 females (59.3%) and 5197 males (40.7%); 9633 (77.8%) primary surgeries and 2742 (22.2%) revisions. Procedures by spinal level were cervical 29.7% (3671), thoracic/thoracolumbar 45.4% (5624), and lumbosacral 24.9% (3080). Age at surgery was > 18 years - 72.7% (8993) and < 18 years - 27.3% (3382). RESULTS: Forty-five of the 12,375 patients (0.36%) had false negative outcomes. False negative results by modality were as follows: spEMG (n = 22, 48.8%), trgEMG (n = 8, 17.7%), DSEP (n = 4, 8.8%), DNEP (n = 4, 8.8%), SSEP (n = 3, 6.6%), DSEP/spEMG (n = 3, 6.6%), and trgEMG/spEMG (n = 1, 2.2%). Thirty-seven patients had immediate postoperative deficits unidentified by IOM; 30 patients (81%) involved nerve root monitoring, four patients had spinal cord deficits, and three patients had peripheral sensory deficits. Eight patients had permanent neurologic deficits, six (0.048%) were nerve root and two (0.016%) were spinal cord in nature. CONCLUSION: Despite correct application and usage, IOM data failed to identify 45 (0.36%) patients with false negative outcomes out of 12,375 surgical patients. Eight patients (0.064%) of these 45 patients had permanent neurologic deficits, six patients had nerve root deficits in nature and two patients had spinal cord deficits. Although admittedly small, this represents the risk of undetected neurologic deficits even when properly using IOM. Deficits are at a higher risk to remain unresolved when not detected by IOM. LEVEL OF EVIDENCE: 4.

Spinal Cord Monitoring Data in Pediatric Spinal Deformity Patients With Spinal Cord Pathology.

STUDY DESIGN: Retrospective. OBJECTIVES: The purpose of this study is to review the efficacy of monitoring data and outcomes in pediatric patients with spinal cord pathology. SUMMARY OF BACKGROUND DATA: The incidence of spinal cord pathology in pediatric patients with scoliosis has been reported between 3% and 20%. Previous studies demonstrated that intraoperative spinal cord monitoring (IOM) during scoliosis surgery can be reliable despite underlying pathology. METHODS: A single-center retrospective review of 119 spinal surgery procedures in 82 patients with spinal cord pathology was performed. Diagnoses included Arnold-Chiari malformation, syringomyelia, myelomeningocele, spinal cord tumor, tethered cord, and diastematomyelia. Baseline neurologic function and history of prior neurosurgical intervention were identified. Outcome measures included ability to obtain reliable monitoring data during surgery and presence of postoperative neurologic deficits. Results were compared for 82 patients with adolescent idiopathic scoliosis (AIS). RESULTS: Usable IOM data were obtained in 82% of cases (97/119). Twenty-two cases (18%) had no lower extremity data. Patients with Arnold-Chiari malformation or syringomyelia pathologies, in isolation or together, had a significantly higher rate of reliable data compared to other pathologies (p < .0001). Among study group cases with usable data, there were 1 false negative (1%) and 4 true positive (4%) outcomes. There were no permanent neurologic deficits. The spinal cord pathology group demonstrated 80% sensitivity and 92% specificity. CONCLUSIONS: Spinal cord monitoring is a valuable tool in pediatric patients with spinal cord pathology undergoing spinal deformity surgeries. When obtained, data allow to detect changes in spinal cord function. Patients with a diagnosis of Arnold-Chiari or syringomyelia have monitoring data similar to those patients with AIS. Patients with other spinal cord pathologies have less reliable data, and surgeons should have a lower threshold for performing wake-up tests to assess spinal cord function intraoperatively.

Significant change or loss of intraoperative monitoring data: a 25-year experience in 12,375 spinal surgeries.

STUDY DESIGN: Retrospective. OBJECTIVE: The purpose of this study was to report the spectrum of intraoperative events responsible for a loss or significant change in intraoperative monitoring (IOM) data. SUMMARY OF BACKGROUND DATA: The efficacy of spinal cord/nerve root monitoring is demonstrated in a large, single institution series of patients, involving all levels of the spinal column (occiput to sacrum) and all spinal surgical procedures. METHODS: Multimodality IOM included somatosensory-evoked potentials, descending neurogenic-evoked potentials, neurogenic motor-evoked potentials, and spontaneous and triggered electromyography. A total of 12,375 patients who underwent surgery for spinal pathology between January 1985 and December 2010 were reviewed. There were 59.3% female patients (7178) and 40.7% male patients (5197). Procedures by spinal level were as follows: cervical 29.7% (3671), thoracic/thoracolumbar 45.4% (5624), and lumbosacral 24.9% (3080). Age at the time of surgery was as follows: older than 18 years, 72.7% (242/8993) and younger than 18 years, 27.3% (144/3382). A total of 77.8% (9633) patients underwent primary surgical procedures and 22.2% (2742) patients underwent revision surgical procedures. RESULTS: A total of 406 instances of IOM data change/loss occurred in 386 of 12,375 (3.1%) patients. Causes for data degradation/loss included the following: instrumentation (n = 131), positioning (n = 85), correction (n = 56), systemic (n = 49), unknown (n = 24), and focal spinal cord compression (n = 15). Data loss/change was seen in revision (6.1%/167 patients) surgical procedures more commonly than in primary procedures (2.3%/219 patients; P < 0.0001). Data improvement was demonstrated by 88.7% (n = 360) after intervention versus 11.3% (n = 46) with no improvement in IOM data. One patient with improved data after intervention versus 14 with no improvement despite intervention had a permanent neurological deficit (P < 0.0001). CONCLUSION: IOM data identified 386 (3.1%) patients with loss/degradation of data in 12,375 spinal surgical procedures. Fortunately, in 93.3% of patients, intervention led to data recovery and no neurological deficits. Reduction from a potential (worst-case scenario) 3.1% (386) of patients with significant change/loss of IOM data to a permanent neurological deficit rate of 0.12% (15) patients was achieved (P < 0.0001), thus confirming efficacy of IOM.

Validity and reliability of intraoperative monitoring in pediatric spinal deformity surgery: a 23-year experience of 3436 surgical cases.

STUDY DESIGN: This was a 23-year retrospective study of 3436 consecutive pediatric orthopedic spinal surgery patients between 1995 and 2008. OBJECTIVE: To demonstrate the effectiveness of multimodality electrophysiologic monitoring in reducing the incidence of iatrogenic neurologic deficit in a pediatric spinal surgery population. SUMMARY OF BACKGROUND DATA: The elective nature of many pediatric spinal surgery procedures continues to drive the need for minimizing risk to each individual patient. Electrophysiologic monitoring has been proposed as an effective means of decreasing permanent neurologic injury in this population. METHODS: A total of 3436 consecutive monitored pediatric spinal procedures at a single institution between January 1985 and September 2008 were reviewed. Monitoring included somatosensory-evoked potentials, descending neurogenic-evoked potentials, transcranial electric motor-evoked potentials, and various nerve root monitoring techniques. Patients were divided into 10 diagnostic categories. True-positive and false-negative monitoring outcomes were analyzed for each category. Neurologic deficits were classified as transient or permanent. RESULTS: Seven of 10 diagnostic groups demonstrated true positive findings resulting in surgical intervention. Seventy-four (2.2%) potential neurologic deficits were identified in 3436 pediatric surgical cases. Seven patients (0.2%) had false-negative monitoring outcomes. These patients awoke with neurologic deficits undetected by neuromonitoring. Intervention reduced permanent neurologic deficits to 6 (0.17%) patients. Monitoring data were able to detect permanent neurologic status in 99.6% of this population. The ratio of intraoperative events to total monitored cases was 1 event every 42 surgical cases and 1 permanent neurologic deficit every 573 cases. CONCLUSION: The combined use of somatosensory-evoked potentials, transcranial electric motor-evoked potentials, descending neurogenic-evoked potentials, and electromyography monitoring allowed accurate detection of permanent neurologic status in 99.6% of 3436 patients and reduced the total number of permanent neurologic injuries to 6.

Loss of spinal cord monitoring signals in children during thoracic kyphosis correction with spinal osteotomy: why does it occur and what should you do?

STUDY DESIGN: A retrospective review of pediatric kyphosis patients undergoing a spinal cord-level osteotomy for correction. OBJECTIVE: To evaluate the prevalence, etiology, timing, and intervention related to loss of spinal cord monitoring data during surgical correction of pediatric kyphosis in the spinal cord region. SUMMARY OF BACKGROUND DATA: Although much has been written regarding the risks inherent to scoliosis surgery, there is less literature available regarding the neurologic outcomes of pediatric kyphosis surgery. As more surgeons contemplate posterior-only kyphosis correction with spinal cord-level osteotomies, the importance of maintaining spinal cord neurologic function is paramount. METHODS: Forty-two patients with pediatric kyphosis undergoing a posterior-only spinal reconstruction with a spinal cord level osteotomy or posterior-based vertebral column resection performed were reviewed. Patients were categorized by diagnosis, type and incidence of osteotomies, and loss of neurogenic mixed-evoked potential (NMEP) data. Interventions required to regain data and postoperative neurologic outcomes were also reviewed. RESULTS: Of the 42 patients, 9 (21.4%) demonstrated a complete loss of NMEP data sometime during surgery while concomitant somatosensory sensory-evoked potentials (SSEP) remained within acceptable limits of baseline values. All 9 patients had intraoperative intervention including: blood pressure elevation (n = 1), release of corrective forces (n = 2), blood pressure elevation and correction release (n = 3), malalignment/subluxation adjustment (n = 1), further bony decompression (n = 1), or restoration of anterior column height via a titanium cage along with further posterior decompression (n = 1). In all cases, SSEPs were unchanged and NMEPs returned varying from 8 to 20 minutes after loss, with all patients having a normal wake-up test intraoperatively and a normal neurologic examination after surgery. CONCLUSION: Intraoperative multimodality monitoring with some form of motor tract assessment is a fundamental component of kyphosis correction surgery in the spinal cord region in order to create a safer, optimal environment and to minimize neurologic deficit. The surgeon must be able to trust the information monitoring provides and act on it accordingly.

Correlation between low triggered electromyographic thresholds and lumbar pedicle screw malposition: analysis of 4857 screws.

STUDY DESIGN: A retrospective analysis of 1078 spinal surgical procedures with lumbar pedicle screw placement at a single institution. OBJECTIVE: Based on previously established normative values, triggered electromyographic stimulation (TrgEMG) was re-examined to evaluate its efficacy in determining screw malposition. SUMMARY OF BACKGROUND DATA: Threshold values for confirmation of intraosseous placement of pedicle screws with EMG stimulation is controversial. METHODS: TrgEMG threshold values for 4857 pedicle screws placed from L2 to S1 from 1996 to 2005 were analyzed. An ascending method of constant current stimulation was applied to each pedicle screw to obtain a compound muscle action potential (CMAP) from lower extremity myotomes. Previously determined threshold value normative data from a published clinical series of 233 screws were as follows: 0 to 4 mA, high likelihood of pedicle wall breach; 4 to 8 mA, possible pedicle wall breach; >8 mA, no pedicle wall defect. RESULTS: A total of 7.74% (376 of 4857) of all screws tested had threshold values <8.0 mA. A total of 19.1% (72 of 376) of these were <4.0 mA: 54% (39 of 72) were repositioned (26) or removed (13) while the remaining 33 screws were left in place following repalpation. A total of 80.9% (304 of 376) had thresholds between 4 and 8 mA: 17.4% (53) were repositioned (38) or removed (15). Nine screws had thresholds of 8.0 mA, 17.4% for 4.0 to 8.0 mA, 54.2% for <4.0 mA, and 100% for <2.8 mA. At 2.8 mA, triggered EMG has a specificity of 100%, with sensitivity of 8.4%; at 4.0 mA, specificity of 99% and sensitivity of 36%; and at 8.0 mA, 94% specificity and 86% sensitivity. TrgEMG is an adjunct technique and should always be used in conjunction with palpation and radiography to optimize safe pedicle screw placement.

The utility of somatosensory evoked potential monitoring during cervical spine surgery: how often does it prompt intervention and affect outcome?

STUDY DESIGN: Retrospective review of the results of somatosensory evoked potentials (SSEP) performed in cervical spine surgery. PURPOSE: To evaluate the utility of spinal cord monitoring during cervical spine surgery in a single surgeon's practice, based on how often it prompted an intraoperative intervention. OVERVIEW OF LITERATURE: Intraoperative monitoring during cervical spine surgery is not a universally accepted standard of care. This is due in part to the paucity of literature regarding the impact of monitoring on patient management or outcome. METHODS: SSEP for tibial, median, and ulnar nerves were monitored in 809 consecutive cervical spine operations performed by a single surgeon. The average patient age was 52 years (range, 2 to 88 years), with 472 males and 339 females. Cases were screened for significant degradation or loss of SSEP data. Specific attention was paid to 1) what interventions were performed in response to the SSEP degradation with subsequent improvement, and 2) whether SSEP changes corresponded with postoperative neurological deficits. RESULTS: Seventeen of 809 patients (2.1%) had SSEP degradation that met warning criteria and therefore prompted intervention. Release of shoulder tape (8) or traction (4) most often resulted in SSEP improvement. Failure of SSEP data to return to within acceptable limits of baseline was associated with neurological deficit (p=0.04). Two patients awoke with new postoperative neurological deficits, which resolved in 6 hours and 2 months respectively. Patients with ossification of the posterior longitudinal ligament (OPLL) were at seven-fold greater risk of intraoperative SSEP degradation. CONCLUSIONS: SSEP monitoring in this surgical population proved sensitive to perioperative factors which may increase the risk of postoperative neurologic deficit, and probably prevented neurological deficits in 15 of 809 patients (1.9%). Improvement in data following intervention appears to correlate well with unchanged neurologic status. Experience with intraoperative monitoring in this patient series has led to incorporation of these techniques as a standard of care in cervical spine surgeries performed by this surgeon.

Intraoperative electrophysiologic monitoring: considerations for complex spinal surgery.

Intraoperative neurophysiologic monitoring techniques have evolved as the complexity of spinal surgery has increased and the limitations of individual modalities have become apparent. Current monitoring strategies include a combination of techniques directed toward detecting changes in sensory, motor, and nerve root function. Close coordination and communication between the monitoring personnel, surgeon, and anesthesiologist is essential to effective intraoperative monitoring.

Increased risk of postoperative neurologic deficit for spinal surgery patients with unobtainable intraoperative evoked potential data.

STUDY DESIGN: This was a retrospective study of 4,310 patients undergoing spinal surgery between 1994 and 2003. OBJECTIVES: To examine the incidence and potential causality of unobtainable somatosensory evoked potential (SSEP) and neurogenic mixed evoked potential (NMEP) data for a population of spinal surgery patients. SUMMARY OF BACKGROUND DATA: Patients with absent or unobtainable evoked potential data may increase the risk of undetected neurologic injury. To date, a comprehensive review of this patient population has not been reported. METHODS: A total of 4,310 consecutive orthopedic spinal surgeries at one institution from January 1994 through December 2003 were reviewed. Cases lacking sufficient monitoring data, despite functional neural integrity (ambulators, intact sensation), were identified. Diagnoses were divided into six general categories. The association between absent evoked potential data and associated neurologic and/or medical pathology was evaluated. RESULTS: A total of 59 of 4,310 cases (1.37%) had absent SSEP and/or NMEP intraoperative data despite functional neural integrity (44 ambulators/15 nonambulators)" 5.08% of study patients awoke with increased neurologic deficit (3 of 59), 2 global deficits, and 1 nerve root deficit. The incidence of postoperative neurologic deficit in the entire surgical population was 0.77% (33 of 4,310), 8 global (0.19%), and 25 nerve root deficits (0.058%). A Fisher's exact test demonstrated a statistically significant difference between the incidence in these two populations (P = 0.0121) and the incidence of global paraplegic deficits (P = 0.0075). CONCLUSION: Patients with unobtainable data pose a much higher risk (P = 0.0121) for postoperative neurologic deficits. Multiple Stagnara wake-up tests are strongly recommended when evoked potential data cannot be obtained.

The significance of anode location for stimulus-evoked electromyography during iliosacral screw placement.

OBJECTIVES: To determine the effect of anode location on the current threshold required to provoke an electromyograph response during stimulus-evoked electromyography for iliosacral screw placement. DESIGN: Prospective cohort. SETTING: Level I trauma center. PATIENTS: Nineteen consecutive patients with 23 unstable posterior pelvic ring injuries treated with iliosacral screws. INTERVENTION: Iliosacral screws were inserted percutaneously over guidewires. Twenty-seven screws were inserted, all into the first sacral vertebrae. The guidewire was used as the cathode for constant-current, stimulus-evoked electromyography for all data collection. Stimulus-evoked electromyographs were obtained with the guidewire at four different stations: at the sacroiliac joint (station I), at the first sacral neuroforamen (station II), in the body of the sacrum (station III), and when the iliosacral screw was in final position over the guidewire (station IV). MAIN OUTCOME MEASURE: Stimulus-evoked electromyographs were obtained with the anode at four different locations for each of the implant stations. Location A had the anode adjacent to the percutaneous insertion site of the guidewire, location B at the ipsilateral anterior superior iliac spine, location C at the midline, and location D at the contralateral anterior superior iliac spine. RESULTS: Moving the anode from midline (location C) toward the entry point of the guidewire increased the current threshold required to provoke an EMG response as much as 67.1% (p < 0.05). Moving the anode from midline to the contralateral anterior superior iliac spine decreased thresholds as much as 3.4% (p > 0.05). In one case, anode placement close to the guidewire insertion site (locations A and B) failed to identify a potentially dangerous implant because current thresholds were >8 mA. With the anode at the midline, current thresholds were <8 mA, indicating unsafe guidewire position leading to redirection of the guidewire. CONCLUSION: The physical location of the anode during stimulus-evoked electromyography monitoring for iliosacral screw placement significantly changes the current thresholds required to provoke an electromyograph response. Current thresholds required to stimulate nerves increase as the anode is moved toward the stimulating electrode. Anode placement ipsilateral to the stimulating electrode may provide a false indication of safe guidewire placement. We recommend anode location at or beyond the midline for stimulus-evoked electromyography monitoring during iliosacral screw placement.

Can triggered electromyograph thresholds predict safe thoracic pedicle screw placement?

STUDY DESIGN: A prospective clinical study of thoracic pedicle screws monitored with triggered electromyographic testing. OBJECTIVE: To evaluate the sensitivity of recording rectus abdominis triggered electromyographs to assess thoracic screw placement. SUMMARY OF BACKGROUND DATA: Triggered electromyographic testing from lower extremity myotomes has identified medially placed lumbar pedicle screws. Higher thresholds indicate intraosseous placement because of increased resistance to current flow. Lower thresholds correspond to compromised pedicles with potential for nerve impingement. No clinical study has correlated an identical technique with rectus muscle recordings, which are innervated from T6 to T12. METHODS: A total of 677 thoracic screws were placed in 92 consecutive patients. Screws placed from T6 and T12 were evaluated using an ascending method of stimulation until a compound muscle action potential was obtained from the rectus abdominis. Threshold values were compared both in absolute terms and also in relation to other intrapatient values. RESULTS: Screws were separated into three groups: Group A (n = 650 screws) had thresholds >6.0 mA and intraosseus placement. Group B (n = 21) had thresholds <6.0 mA but an intact medial pedicle border on reexamination and radiographic confirmation. Group C (n = 6) had thresholds <6.0 mA and medial wall perforations confirmed by tactile and/or visual inspection. Thus, 3.9% (27 of 677) of all screws had thresholds <6.0 mA. Only 22% (6 of 27) had medial perforation. Group B screws averaged a 54% decrease from the mean as compared with a 69% decrease for Group C screws (P = 0.0160). There were no postoperative neurologic deficits or radicular chest wall complaints. CONCLUSION: To assess thoracic pedicle screw placement, triggered electromyographic thresholds <6.0 mA, coupled with values 60-65% decreased from the mean of all other thresholds in a given patient, should alert the surgeon to suspect a medial pedicle wall breach.