Intraoperative neuromonitoring in spine deformity surgery: modalities, advantages, limitations, medicolegal issues - surgeons' views.

In spine deformity surgery, iatrogenic neurologic injuries might occur due to the mechanical force applied to the spinal cord from implants, instruments, and bony structures, or due to ischemic changes from vessel ligation during exposure and cord distraction/compression during corrective manoeuvres.Prompt reaction within the reversible phase (reducing of compressive/distractive forces) usually restores functionality of the spinal cord, but if those forces continue to persist, a permanent neurological deficit might be expected.With monitoring of sensory pathways (dorsal column-medial lemniscus) by somatosensory-evoked potentials (SSEPs), such events are detected with a sensitivity of up to 92%, and a specificity of up to 100%.The monitoring of motor pathways by transcranial electric motor-evoked potentials (TceMEPs) has a sensitivity and a specificity of up to 100%, but it requires avoidance of halogenated anaesthetics and neuromuscular blockades.Different modalities of intraoperative neuromonitoring (IONM: SSEP, TceMEP, or combined) can be performed by the neurophysiologist, the technician or the surgeon. Combined SSEP/TceMEP performed by the neurophysiologist in the operating room is the preferable method of IONM, but it might be impractical or unaffordable in many institutions. Still, many spine deformity surgeries worldwide are performed without any type of IONM. Medicolegal aspects of IONM are different worldwide and in many cases some vagueness remains.The type of IONM that a spinal surgeon employs should be reliable, affordable, practical, and recognized by the medicolegal guidelines. Cite this article: EFORT Open Rev 2020;5:9-16. DOI: 10.1302/2058-5241.5.180032.

Introduction

Expectations of patients undergoing correction surgery for spinal deformity have been shifting toward the ideal spine shape. In most cases, surgeons can meet those expectations with new surgical techniques, improved spinal implants, perioperative care, safer anaesthesia, invention of cell saver, etc. Unfortunately, with more extensive surgeries, the intraoperative risk of spinal cord injury increases. This neurologic injury may occur due to direct mechanical force applied to the spinal cord (implants, instruments, bony structures), or due to indirect ischemic changes (vessel ligation during exposure, cord distraction/compression during corrective manoeuvres). Quick identification of impending spinal cord injury opens the window of opportunity to the surgeon and anaesthesiologist to act before the injury becomes irreversible. Therefore, the monitoring of gross motor and sensory function in spine deformity surgery is extremely important, especially when correcting spine deformities in the paediatric population presenting with additional neurological risk: neglected rigid kyphoscoliosis, congenital spine anomalies, revision spine surgeries, marfanoid patients with lordoscoliosis, etc.[1] Surgeons should understand the rationale and the clinical basis for intraoperative neuromonitoring (IONM), interpret the monitoring alerts, and utilize them for a better surgical outcome.[2] In addition to a surgeon, the neuromonitoring team is composed of clinical neurophysiologist, anaesthesiologist, and monitoring technician.[3]

The simplest and the oldest way of detecting gross motor function deficit is the Stagnara wake-up test. It is based on intraoperative reduction of anaesthesia and asking the patient to move his/her limbs. This test is considered to be almost 100% accurate, while the limitations are mental disability, young age and pre-existing motor deficit. Its drawbacks are the risk of self-extubation, loss of intravenous lines, dispositioning on the operating table, air embolism, as well as potential psychological stress. Nowadays this test may be used only when neurophysiological recordings are unavailable or fail to return after excluding all technical, anaesthetic and surgical factors and when there is a problem of obtaining an IONM signal, as with patients suffering from thoracic myelopathy.[4]

The most commonly utilized IONM techniques include, but are not limited to: somatosensory-evoked potentials (SSEPs), transcranial electric motor-evoked potentials (TceMEPs) and electromyography (EMG).[5] The aim of this review article is to present the clinical importance of IONM, its different technical modalities, its advantages, limitations, and legal issues.

Somatosensory-evoked potentials

The monitoring of somatosensory spinal pathways (dorsal column–medial lemniscus) is based on sub/cortical responses to continuous electrical stimulation of peripheral nerves (e.g. tibial, peroneal, ulnar/median nerve). Although sensory deficit is less debilitating than motor deficit, the monitoring of sensory pathways gives an insight into the function of motor pathways too, since the ischemic or mechanic injury usually affects both pathways. This, the most common type of IONM, which is easy to implement, and has no contraindications. It may be sufficient for the posterior spine approach in spine deformity surgery as it has the range of sensitivity of 25–92%, and specificity of 96–100%.[6]

Latency increase of more than 10% and reduction of amplitude more than 50% in comparison to the baseline signals are highly suggestive of neurologic impairment.[7] Halogenated or nitrous-oxide-based agents influence SSEP amplitude and latency, but SSEPs are generally more resistant to anaesthetics than MEPs.[2] Sevoflurane and desflurane may be used as long as the minimal alveolar concentration (MAC) of the inhaled agents is kept stable below 0.7%. The controlled intravenous application of sedatives and analgesics (midazolam, fentanyl/ketamine) along with N2O also allows recording of stable SSEP parameters.[8,9] Muscle relaxants do not affect SSEPs, in fact they might enhance the SSEP signal by decreasing electric noise by eliminating muscle artifacts (Fig. 1).

Fig. 1

Normal and altered somatosensory-evoked potentials (SSEPs): SSEP loss on left lower limb, concomitant to transcranial electric motor-evoked potential (TceMEP) loss (see Fig. 2.); blue line – baseline SSEP, purple line – SSEP during osteotomy; upper lines represent cortical, lower lines cervical responses; latency 27% (29.5 to 40.3 msec) of cervical response after stimulation of posterior tibial nerve; reduction of amplitude in left lower limb 76% (0.46 to 0.11) – red circles; rest of figure represents normal SSEPs.

Fig. 2

Normal and altered motor-evoked potentials (MEPs): upper lines represent response of tibialis anterior, lower lines of abductor hallucis muscles; no response in both left lower limb muscles on the left side, normal response on the right side.

SSEP recording requires signal averaging, which results in a time lag before data interpretation (1–20 min). Therefore, an injury can be irreversible before it is even detected.[2] Further limitations of SSEPs are the monitoring of patients with pre-existing neurologic deficit (myelopathy, spinal cord tumour, peripheral neuropathy), or situations with the intraoperative isolated motor pathway or nerve root injury, which can be detected only by MEP or EMG type of recording.

Motor-evoked potentials

MEPs directly monitor the function of the motor pathways in the anterior/central portions of the spinal cord and the nerve roots. These are very sensitive indicators of the corticospinal tract injury, and have proven to be very sensitive indicators of the spinal cord ischemia during spinal deformity correction.[6] Spinal cord motor function monitoring is based on muscle responses to intermittent stimulation of the motor cortex in real time.

Transcranial stimulation as single high voltage or multiple small stimuli of the motor cortex can be magnetic, and, more commonly in the operating room, electric (TceMEP). Recording sites are at the end muscle, preferably muscles rich in corticospinal tract innervations such as distal limb muscles. Electromyography signals – compound motor action potential (CMAP) – are typically acquired through needle electrodes inserted bilaterally into the upper limbs as control to determine systemic, anaesthesia and positioning-related changes[10] (abductor digiti minimi, the first dorsal interosseous or abductor pollicis brevis), and lower limbs (vastus lateralis, anterior tibialis, gastrocnemius medialis, abductor hallucis). TceMEP signals should have an amplitude ⩾ 50 μV to be considered as a ‘monitorable’. A higher number of channels on the lower extremities provide a more reliable picture in the case of developing changes that do not necessarily affect all recoding channels to the same degree. Unilateral or bilateral, reproducible decrease in TceMEP amplitude from the baseline greater than 60% should be considered as a ‘significant warning criteria’ in spine deformity surgery.[11] Since MEPs are much more unstable than SSEPs, other authors rather employ an interval of 50–80% amplitude reduction instead of a ‘cut off’ value (Fig. 2).[12]

MEP has high sensitivity (75–100%) and specificity (84%–100%)[6] but it requires avoidance of halogenated anaesthetics (halothane, sevoflurane, isoflurane, etc.), and neuromuscular blockades (vecuronium, rocuronium etc.). Ketamine-based anaesthesia allows for appropriate MEP recordings, but total intravenous anaesthesia (TIVA) with propofol is preferred.[13]

D-wave is subtype of MEP where the recording site is on the spinal cord. It is used together with other modalities of IONM in various spinal procedures,[14] mostly in spinal cord tumour procedures,[15] and in neurophysiological assessment of injured spinal cords.[16] In scoliosis surgery its usage is not the primary method of choice due to low sensitivity (27%),[17] absence of D-wave at the baseline, undeveloped neural system in children, and long-term injury of the spinal cord.

Electromyography

EMG is a type of ‘real-time’ neuromonitoring modality, mostly employed in detecting nerve root injuries during minimally invasive spine surgery, transpsoas approaches, screw placements or decompressions.[2] Free-running EMG modality does not require stimulation, and it can be recorded continuously from preselected muscle groups based on the nerve roots at risk. If the nerve root is irritated, continuous electrical activity in the myotome is noted with 100% sensitivity, but only 23.5% specificity. An abnormal free-running EMG response during the spine procedure may or may not be associated with a clinical deficit, while normal response is predictive of the preserved nerve root.[18] Extensive nerve root manipulation or impingement will elicit an increased activity, while an accidental cut of the nerve root will manifest as the absence of recoded electrical signal: ‘silence’.

Triggered EMG modality records the electrical activity of a certain muscle after stimulation (triggering) in the proximity of the corresponding nerve root. In application, bone has high impedance requiring a high threshold to stimulate the adjacent nerve. If the pedicle cortex is intact, high stimulation is necessary to evoke CMAP. When there is perforation of osseous pedicular wall, direct stimulation of a pedicular screw can activate the adjacent nerve root and evoke CMAP in the appropriate myotome at lower stimulus intensities than would be expected with the imperforated pedicle cortex.[19] A positive EMG response at or below a constant current of < 6–10 mA may be an indication of inspection, redirection, or removal of the instrument or implant.[20] Lall et al summarize data from few studies concluding that triggered EMG has a high rate of false-positive alarms, without clinical correlation.[6] Both trigger and free-running EMG requires total avoidance of neuromuscular blockade (Fig. 3).

Fig. 3

Free-running and triggered electromyography (EMG); two purple lines on the left side show free-running EMG recorded in the upper and lower parts of the rectus abdominis muscle; right side – adequate response in left inferior rectus abdominis muscle after triggering of T10 left pedicle, indicating that there is no breach of pedicle.

Multimodality intraoperative neuromonitoring

In general, if a posterior approach is being used, SSEPs may be sufficient, but anterior approaches most likely warrant transcranial MEPs due to the risk of anterior spinal artery syndrome. In cases in which nerve root deficits are of concern, spontaneous EMG and triggered EMG monitoring may be of value. In cases in which spinal cord deficits are most likely (spine deformity surgery, intradural tumours), multimodal IONM is highly recommended. After selection of the appropriate monitoring modality, the anaesthesiologist should adjust anaesthetic agents to allow the best obtainable IONM recordings.[21] If transcranial MEPs are being used, halogenated anaesthetics are contraindicated, and TIVA will optimize signal acquisition. MEPs and any form of EMG monitoring preclude the use of neuromuscular blockade.

Although several false-negative results have been reported with SSEP, just one article has reported false-negative MEPs during an operation.[22] When SSEPs and MEPs are combined in spine deformity surgery, the sensitivity to detection of permanent motor and sensory neurologic injury during spinal deformity surgery is 99.6–100.0%,[23] and the specificity is 84–100%.[6] But, false-negative results can still be found even with a combined IONM. Diab et al reported a group of 1301 adilescent idiopathic scoliosis (AIS) patients; all were monitored using both SSEPs and MEPs. One patient had no changes identified during the combined monitoring, but he still awoke with a spinal cord injury that resolved spontaneously within three months (Table 1).[24]

Table 1.

Advantages, limitations, alarm criteria, anaesthetic requirements, sensitivity and specificity of SSEP, MEP and EMG neuromonitoring

	Monitoring	Advantages	Limitations	Alarm criteria	Anaesthetic requirements	Sensitivity	Specificity
SSEP	Functional integrity of sensory pathways	Continuous monitoring: firm warning criteriaMore recording sites, more reliable result.Generally better resistance of anaesthetics than MEP	Signal averaging results in time delay: injury can be irreversible before detectionInfluence of halogenated and nitrous-oxide-based agents	Latency increase > 10%, signal decrease > 50%	Intravenous anaesthesia, eventually dexmedetomidine	25–92%	96–100%
MEP	Functional integrity of motor pathways	Real-time monitoring, instant, no averagingReliable even after posterior myelotomyMore sensitive to ischemiaLess sensitive to electrical noise	Requires total intravenous anaesthesia without neuromuscular blockadeMovementsContraindicated in seizures and pace makersWide warning criteria	Signal decrease > 50–75%	Total intravenous anaesthesia, no halogenated agents or neuromuscular blockade	75–100%	84–100%
EMG	Functional integrity of peripheral nerves	Constant feedback,Combined with SSEP improved specificity,Easy performance and interpretation	High rate of false positivesVery sensitive on temp. changes and cauteryRriggered EMG: only insight in pedicle integrity	No firm alarm criteria	Avoidance of neuromuscular blockade	Free-running: 100%Triggered: 99.5%	Free-running: about 23%Triggered: low
Multimodal IONM	All of the above	Both modalities together improve sensitivity and specificity: most reliable	Highly trained personnelTechnical requirementsCost	Avoidance of each of the above-mentioned limitations with adjusted anaesthetic protocols will provide sensitivity of 100% and specificity of 84–100%.

Note. SSEP, somatosensory-evoked potentials; MEP, motor-evoked potentials; EMG, electromyography; IONM, intraoperative neuromonitoring.

Vitale et al established a consensus-based ‘Checklist for the response to IONM changes’. After announcement to the OR personnel (stop the unnecessary conversation and traffic in the room, summon attending anaesthesiologist and senior neurophysiologist, anticipate the need for imaging – C/O arm, etc.) and checking of electrodes, neck and limb positions, the anaesthesiologist will optimize mean arterial pressure, haematocrit, oxygenation, and pH. The next step would be discussion with the anaesthesiologist about anaesthetic agents (neuromuscular blockade, MAC, etc.) and discussion with the whole team about the actions prior to the signal loss. Accordingly, reversal actions would be tackled (removing of traction or other correcting forces, removing rod/screws, examine spinal cord for compression – osteotomy site, intraoperative imaging – malpositioned screws). Continuation of SSEP and/or TceMEP signal degradation indicates a wake-up test, steroid protocol, and consultation with a college (completing or staged surgery).[11] The same algorithm is generally accepted.[25]

Postoperative neurological deficit after spine deformity surgery is also possible. Its incidence is 0.01%, mostly appearing during the first 24 hours. Forty-one per cent of affected patients experience complete neurologic recovery, 26% partial, and 33% no recovery.[26] None of the monitoring techniques are able to predict a delayed-onset paraplegia that appears after the surgery. Postoperative imaging studies should be conducted to discover the cause of the neurologic deficit (epidural haematoma, misplaced screw, spinal cord ischemia secondary to excessive tensioning, etc.) without any unnecessary delay. Early decompression may improve the neurologic outcome for the patient with a new onset of neurologic deficit in the acute postoperative period. Conversely, if no abnormality is identified on computed tomography (CT) or magnetic resonance imaging (MRI) scan, the patient should be closely observed with a supportive treatment.[19]

Who can carry out neuromonitoring?

Monitoring services during a surgery may be provided by a variety of personnel with different education, annual volume, work per case, and types of cases.[27] Technologists acquire the intraoperative data and relay the information to the surgeon, anaesthesiologist, neurologist, neurophysiologist, or another professional for interpretation. Interpretation may be carried out in the operating room or by remote internet consultation (telemedicine) on a continuous or intermittent basis, but no more than three cases can be followed simultaneously. Surgeons should understand the qualifications and roles of the personnel responsible for data acquisition and the interpretation of intraoperative neurophysiological data monitoring.[28]

Finally, an appropriate standard of care relating to IONM is difficult to devise because of national variance with regard to qualifications of neurophysiologic, technical and professional personnel, different levels of training and certification, and anaesthesia protocols[29] and IONM service availability. According to a web-based survey among surgeons in Canada on their attitude to the interpretation of IONM, most of the interpretation is performed by either technologists or by surgeons themselves. Most surgeons would prefer professional oversight by the neurologist or neurophysiologist at the doctoral level. There appears to be a shortage of qualified personnel and a lack of Canadian guidelines for the performance of this task.[30] A survey from 2011 of 117 French spinal surgeons showed that only 36% had neurophysiological monitoring available (public healthcare facilities, 42%; private facilities, 27%).[31] In the article by Siller et al from 2019, based on a survey in German-speaking countries (Germany, Austria, Switzerland), about 76% of neurosurgical and 15% of orthopaedic spine centres utilize IONM. The main modalities were MEP and SSEP, and the main indications were scoliosis and intradural spinal tumour surgeries. IONM utilization was low in spine surgeries for degenerative, traumatic, and extradural tumour diseases. A more frequent IONM use, however, was mainly limited due to the shortage of skilled staff and restricted reimbursement.[32]

Authors’ view

Apart from the patient’s safety, IONM encourages surgeons to achieve better curve correction, and reduces his/her stress. In the US and other developed countries, multimodal IONM in spine deformity surgery is a standard of care and obligatory, but worldwide many surgeons perform spine surgeries without any type of IONM.

Some IONM devices have ‘surgeon-directed MEP mode’, where a surgeon can easily set up the system, and monitor the signals during the surgery.[33] Interpretation of MEPs that represent a gross motor function of the lower limbs, in relation to the current action of the surgeon or anaesthesiologist, is not a problem for a high-volume spine surgeon. There are some reports about IONMs monitored by surgeons,[30,31,33] although there was some criticism, especially of the statement ‘Whatever the monitoring technique used, it remains preferable to the absence of monitoring’.[34]

We have done our best to discover legal aspects of IONM in different countries, but there is no common approach. Medicolegal aspects of IONM are different worldwide and in many cases some vagueness remains. Finally, the surgeon and/or hospital responsible for the IONM services must be sure that each provider has adequate training and experience, consistent with the national legislation.

Discussion

A fear of complications, especially neurological ones, is experienced by both patients and physicians. Prior to surgery, the complication rate can be diminished by optimizing factors such as general condition, anaemia, pulmonary function, etc. During surgery, proper monitoring of cardiopulmonary, urinary and spinal cord functions can additionally decrease the overall complication rate. Neuromonitoring recognizes functional changes in the spinal cord, usually in the reversible phase, when prompt reaction prevents a neurological complication. If distractive/compressive forces persist, neurologic deficit may be expected.[35-37]

In the pre-IONM era, the rate of permanent paraplegia was 0.4% according to the Scoliosis Research Society’s database from 1975,[38] while in the IONM era the reported incidence of major neurologic injury is from 0.4% to 1.9%. This can be explained by different methodologies of the mentioned studies and more aggressive surgeries nowadays. In the article by Schwartz et al from 2007, there was no permanent neurological deficit in a group of 1121 MEP/SSEP-monitored scoliosis surgeries, although the relevant signal changes were noted in 3.4% surgeries.[39] IONM significantly decreases risk in spinal surgeries, and today it is the standard of care in the US and many other developed countries. Multimodal IONM is accurate and comfortable for spine surgeons, but it may be unaffordable to many of them. The monitoring of MEP alone is a safe, simple and feasible modality in spine deformity surgeries, but multimodal IONM is a superior method.[40] If the choice is to be made between MEP and SSEP, the former is a better option.[41] The surgeon-directed MEP type of IONM allows the surgeon to perform it alone with minimal engagement during the surgery. Everything else requires the surgeon’s ability to interpret the amplitudes and correlate them with the situation in the operative field and anaesthesia parameters.[33] It should not be the first solution for serious spinal surgeries since experienced neuromonitoring teams had fewer than one-half as many neurologic deficits per 100 cases compared to teams with relatively little monitoring experience.[27] However, surgeon-directed MEP is a better option than no IONM, disregarding its limitations and consequences.

The average four-hour spine case SSEPs costs US$942, and Tce-MEPs US$1115, and US$1423 in combination.[42] Except the costs and duration of anaesthesia due to MEP/SSEP setting, IONM has no adverse effects. Large retrospective reviews have revealed the incidence of tongue lacerations to be about 0.2%,[43] and seizures 0.03% related to IONM.[44] On the other hand, IONM was significantly associated with increased home discharge and lower risk of neurologic complications, while hospital charges and length of hospitalization were not affected by IONM.[45]

IONM is also utilized for other spine surgery indications: tethered cord, injured spinal cord, intramedullary tumours, extramedullary tumours, minimally-invasive surgery, cervical myelopathy, as well in other fields of surgery: cranial, vascular, cardiothoracic, etc. It is helpful in preventing perioperative peripheral nerve injury (PPNI) due to excess mechanical pressure and torsion of the limbs and neck. Prone patient positioning is designed to achieve optimal exposure and operative conditions; however, it might be potentially harmful due to over-prolonged contact with the operating table frames.[10,46,47] Most commonly observed are positioning brachial plexopathies with a risk of 6.2% in Labrom et al’s study,[48] and 3.6% in Schwartz et al’s study.[49] Repositioning the arm(s) or shoulder(s) resulted in nearly immediate improvement of amplitude of SSEP, resulting in postoperative neurologically intact patients.[49]

Finally, combined SSEP/MEP modality performed until the wound closure[50] has the highest reliability in spine deformity surgeries,[6,11] and should be the IONM option that surgeons insist on. Perioperatively, close follow-up, especially in the first 48 hours, and prompt action if needed is also mandatory.[26] Maintaining a high quality of practice with correct application, and an appropriately trained team is critical.[51]

Conclusion

Patient selection, preoperative planning, surgical technique and diligent IONM with frequent perioperative neurologic examination are the key points in prevention of neurodeficit in spine deformity surgeries. The type of IONM that the spinal surgeon employs should be reliable, affordable, practical, and recognized under the relevant national law. Combined SSEP/MEP performed by the neurophysiologist is the preferable method of IONM in surgeries of paediatric spine deformities.