Spinal Cord Monitoring (Intraoperative Neuromonitoring)
Overview: why monitor & the tracts at stake
Intraoperative neuromonitoring exists to give a real-time warning of an evolving neurological injury, whether from distraction or correction, direct trauma, implant malposition or ischaemia, so that it can be reversed before it becomes permanent.
Different tracts, different modalities. The posterior (dorsal) columns carry sensation and are monitored by SSEP. The corticospinal tracts carry motor output and are monitored by MEP. The two lie in different parts of the cord with different blood supplies, so no single modality is sufficient.
What SSEP misses. SSEP monitors the dorsal (posterior) columns, while the corticospinal tracts and the cord's blood supply are anterior. It can therefore miss an isolated anterior-cord or motor injury. That was historically the cause of the false-negative "normal SSEP but woke up paraplegic", and relying on SSEP alone is the trap.
What MEP adds. MEP monitors the corticospinal (motor) tracts and the anterior cord: the function that matters most, and the territory at risk in anterior spinal artery injury. Combined with SSEP, it gives complete coverage.
The anterior spinal artery. A single anterior spinal artery supplies the anterior two-thirds of the cord: the corticospinal (motor) tracts, the spinothalamic (pain and temperature) tracts and the anterior horns. Paired posterior spinal arteries supply the posterior third, which holds the dorsal columns (proprioception and vibration).
The artery of Adamkiewicz. The dominant thoracolumbar feeder is the artery of Adamkiewicz (arteria radicularis magna). It usually arises on the left between about T9 and T12 and supplies the lumbar enlargement. The mid-thoracic cord, around T4–T9, is a poorly perfused watershed.
Why the supply matters for monitoring. An ischaemic insult to the anterior cord, from correction or distraction stretching the feeders or from compromise of the anterior spinal artery, knocks out MEP while sparing the dorsal columns, so SSEP stays normal. That "normal SSEP, lost MEP" pattern is exactly what MEP was added to catch. It is also why raising the MAP to perfuse the anterior cord is the first response.
Anterior cord syndrome. Bilateral loss of motor function and of pain and temperature below the level, with proprioception, vibration and light touch preserved. It is the worst-prognosis incomplete cord syndrome.
Concepts: the monitoring modalities
Somatosensory evoked potentials. SSEP stimulates a peripheral nerve, for example the posterior tibial or ulnar, and records the response at the cortex or scalp, and sometimes at spinal or peripheral sites. It monitors the dorsal column–medial lemniscus (sensory) pathway. The standard alarm criteria are about a 50% decrease in amplitude or a 10% increase in latency.
Strengths and limits of SSEP. It is robust, continuous and relatively anaesthesia-tolerant. Its signals are averaged, so slightly delayed, and it does not monitor the motor tracts.
Motor evoked potentials. Transcranial MEP applies stimulation over the motor cortex and records the compound muscle action potential (CMAP) in target muscles, monitoring the corticospinal (motor) tract. It is more immediate and detects the motor, anterior-cord injury that SSEP misses. The price is that it is highly sensitive to inhalational agents and neuromuscular blockade, so it mandates an anaesthetic technique that preserves it.
The MEP alarm. The signal to act is a significant amplitude loss or threshold rise. Many teams treat any reproducible loss, or a drop of more than 50%, as the alarm.
Relative contraindications. Transcranial MEP is powerful but not free of hazards. It is relatively contraindicated with:
- epilepsy or a seizure disorder
- cortical lesions or skull defects near the stimulating electrodes
- implanted intracranial electrodes, vascular clips or shunts
- a cardiac pacemaker or implantable defibrillator
- cochlear or other electronic implants
Complications. The commonest is a tongue or lip laceration, with dental or jaw injury, from the masseter and temporalis contraction the stimulus evokes, so a soft bite block is mandatory. The limb twitch can interrupt surgery at the critical moment. Rarely there are seizures, cardiac arrhythmias or scalp burns at the electrode sites.
Limitations. MEP is unreliable in young children, because the corticospinal tract is not fully myelinated in the first few years. A baseline is hard to obtain in a pre-existing motor deficit or established myelopathy. So confirm a reliable baseline at the start before trusting later changes.
Electromyography. EMG has two modes. Free-running EMG listens continuously for nerve-root irritation, which shows as bursts, trains or neurotonic discharges with traction or manipulation, and is non-specific. Triggered (stimulated) EMG tests a pedicle screw: the screw is stimulated and the threshold needed to evoke a muscle response is recorded.
Reading the threshold. A low threshold suggests a medial cortical breach, with the screw near a nerve root; a high threshold is reassuring that the screw is contained. Reading it the wrong way round is one of the traps that cost marks. Threshold cut-offs vary, and EMG, like MEP, needs no muscle relaxant.
The Stagnara wake-up test. This is the historical confirmatory test. Anaesthesia is lightened intraoperatively and the patient is asked to move the feet on command, giving a gross confirmation of motor function. It is used when monitoring alarms, to confirm a true deficit, or when monitoring is unavailable.
Limits of the wake-up test. It tests only gross movement at a single time point. It interrupts surgery, with risks of recall, air embolism and self-extubation or dislodgement.
- Monitors
- Posterior columns (sensory)
- Alarm / signal
- ~50% amplitude drop or ~10% latency increase
- Key caveat
- Misses isolated motor/anterior-cord injury; averaged (delayed)
- Monitors
- Corticospinal (motor) / anterior cord
- Alarm / signal
- Significant amplitude loss / threshold rise
- Key caveat
- Very sensitive to volatile agents & paralysis (need TIVA, no relaxant)
- Monitors
- Nerve-root irritation
- Alarm / signal
- Bursts / neurotonic discharges
- Key caveat
- Non-specific; needs no muscle relaxant
- Monitors
- Pedicle screw position
- Alarm / signal
- LOW stimulation threshold = medial breach
- Key caveat
- Threshold cut-offs vary; needs no relaxant
- Monitors
- Gross voluntary motor function
- Alarm / signal
- Patient cannot move feet on command
- Key caveat
- Single time point; interrupts surgery; risks (recall, extubation, embolism)
Clinical relevance: anaesthesia & the response to a signal change
The anaesthetic must preserve the signals. For MEP and EMG to work, the anaesthetic must preserve neuromuscular transmission and cortical excitability. That means total intravenous anaesthesia (TIVA), propofol-based, avoiding or minimising volatile (inhalational) agents, which suppress MEP, and no long-acting muscle relaxant once motor monitoring is running. Get this wrong and a drug effect will be misread as an injury.
Physiology. Keep physiology stable, with an adequate mean arterial pressure, normothermia and haemoglobin. Hypotension, hypothermia and anaemia themselves reduce the signals and threaten cord perfusion.
A real change is an emergency. A genuine, persistent change calls for a structured response. The first physiological move is to raise the MAP for cord perfusion and correct anaemia and hypothermia, before assuming a surgical cause. Then release the correction or distraction and remove recent implants rather than pushing on.
How firm the sequence is. No evidence-based treatment algorithm exists for a signal change. The checklist below is consensus and physiological reasoning, as the Evidence Base sets out.
- 1Confirm it is realCheck for technical/equipment causes and verify with the neurophysiologist; note whether it is SSEP, MEP, EMG or all — and whether unilateral or bilateral.
- 2Anaesthetic & physiologicalRaise the mean arterial pressure (improve cord perfusion); correct hypotension, anaemia, hypothermia and hypocapnia; confirm the patient is not inadvertently paralysed or too deeply anaesthetised.
- 3SurgicalRelease/reduce the correction or distraction, remove recently placed instrumentation or grafts, decompress and inspect the cord, then reassess the signals.
- 4Confirm & treatConsider a Stagnara wake-up test; if a true deficit persists, maintain a high MAP, consider steroids per local protocol, and plan imaging/management. Act quickly — the goal is to reverse before it becomes permanent.
Viva practice
Exam viva
Practise clinical reasoning and management decisions out loud
“During scoliosis correction, what neuromonitoring modalities would you use, what does each monitor, and why is multimodal monitoring preferred?”
“Midway through the correction the MEPs are lost. What anaesthetic conditions are required for monitoring, and how do you respond to this signal change?”
Exam cheat sheet
Modalities & what they monitor
- SSEP = posterior columns (sensory); alarm ~50% amplitude drop or ~10% latency rise
- MEP = corticospinal (motor)/anterior cord; immediate; anaesthesia-sensitive
- EMG: free-run (root irritation) + triggered (low threshold = medial pedicle-screw breach)
Why multimodal
- SSEP misses motor/anterior-cord injury; MEP misses sensory
- Multimodal (SSEP + MEP ± EMG) is now standard
- Risk factors for signal change: preop deficit, severe kyphosis, large curve, cord shortening (VCR)
Anaesthesia
- TIVA; avoid/minimise volatile agents (suppress MEP)
- No long-acting muscle relaxant once MEP/EMG running
- Maintain MAP, normothermia, Hb (perfusion)
Response to a signal change
- Confirm real (technical/neurophysiologist); check anaesthesia/paralysis
- Raise MAP; correct hypotension/anaemia/hypothermia
- Release correction/distraction, remove recent instrumentation; Stagnara wake-up test; treat if true deficit
Evidence Base
The use of IONM in scoliosis/deformity surgery has risen steadily, and multimodal monitoring (e.g. combined SSEP + EMG, increasingly with MEP) is used more often than single-modality monitoring, because the modalities are complementary. Risk factors for intraoperative signal changes include a pre-operative neurological deficit, severe kyphosis, a large curve magnitude, and significant cord shortening (e.g. during vertebral column resection). Definitive randomised proof that monitoring reduces deficits is hard to obtain (the highest-risk cases are the ones monitored), but multimodal IONM allows potential injuries to be localised and managed in real time and is considered a valuable safety tool in deformity surgery.
Somatosensory evoked potential spinal cord monitoring reduces neurologic deficits after scoliosis surgery: results of a large multicenter survey
- Landmark Scoliosis Research Society multicentre survey of SSEP spinal-cord monitoring during scoliosis surgery.
- Experienced SSEP monitoring teams had FEWER THAN HALF as many neurologic deficits per 100 cases as inexperienced teams; false-negative monitoring (a definite deficit despite stable SSEPs) occurred in only 0.063% of patients.
- Surgeon experience and use of the wake-up test were also independently associated with fewer deficits.
Neuromonitoring in spinal deformity surgery: a multimodality approach
- Multimodal intraoperative monitoring (SSEP, MEP, EMG) localises and manages potential neurologic injuries in real time and is a valuable tool for the safety of spinal deformity surgery.
- Risk factors for intraoperative signal changes include preoperative neurologic deficit, severe kyphosis, increased curve magnitude and significant cord shortening (e.g. vertebral column resection).
- Although no evidence-based treatment algorithm exists for signal changes, structured strategies help prevent and address neurologic alarms.
Demographic trends in the use of intraoperative neuromonitoring for scoliosis surgery in the United States
- Use of intraoperative neuromonitoring for scoliosis surgery rose from 27% (2005) to 46.9% (2011); multimodal monitoring was used more commonly than unimodal (64.6% vs 35.4%).
- The most commonly used combination was SSEP + EMG, and the least used was MEP alone.
- Observed neurological-injury rates were similar with and without monitoring (1.8% vs 2.0%), but the authors caution this likely reflects that higher-risk surgeries preferentially use monitoring.
The foundational efficacy data (experienced SSEP teams under half the deficit rate; false-negatives only 0.063%) come from the SRS multicentre survey, Nuwer et al. 1995 (DOI); the multimodal approach, risk factors for signal changes and the real-time-management rationale from Laratta et al. 2017 (DOI); and the usage trends/modality data from Ajiboye et al. 2018 (DOI). The specific SSEP alarm criteria (50% amplitude / 10% latency), the SSEP-vs-MEP tract distinction, triggered-EMG pedicle-screw testing, the TIVA/no-relaxant requirement and the Stagnara wake-up test are standard, well-established neuromonitoring teaching.
Now the uncomfortable part, because an examiner may press on it. Ajiboye's database study is cited above for usage trends, but its headline result is a null: neurological injury occurred in 1.8% of scoliosis cases monitored and 2.0% not monitored (P = 0.561) - no significant difference. That is not a reason to abandon monitoring, and the authors say so themselves: inherently riskier operations are the ones that get monitored, so confounding by indication works against IONM in any observational comparison, and the true effect may be larger than the raw numbers suggest.
The two datasets are also answering different questions. Nuwer compared experienced monitoring teams with inexperienced ones - that is a study of how well monitoring is done, not of whether to do it - while Ajiboye compared monitored with unmonitored cases in administrative data with no adjustment for case complexity. Neither is a randomised comparison, and there will not be one, because withholding monitoring from a high-risk deformity correction is not something an ethics committee will approve. Say that plainly rather than overclaiming: the case for IONM rests on physiological rationale, a very low false-negative rate in experienced hands (0.063%), and the ability to act on a change in real time - not on trial evidence of reduced deficits.
Note too that Laratta's review states directly that no evidence-based treatment algorithm exists for responding to a signal change. The checklist above is consensus and physiological reasoning, and should be presented as such.