Anesthetic considerations in acute spinal cord trauma.

Patients with actual or potential spinal cord injury (SCI) are frequently seen at adult trauma centers, and a large number of these patients require operative intervention. All polytrauma patients should be assumed to have an SCI until proven otherwise. Pre-hospital providers should take adequate measures to immobilize the spine for all trauma patients at the site of the accident. Stabilization of the spine facilitates the treatment of other major injuries both in and outside the hospital. The presiding goal of perioperative management is to prevent iatrogenic deterioration of existing injury and limit the development of secondary injury whilst providing overall organ support, which may be adversely affected by the injury. This review article explores the anesthetic implications of the patient with acute SCI. A comprehensive literature search of Medline, Embase, Cochrane database of systematic reviews, conference proceedings and internet sites for relevant literature was performed. Reference lists of relevant published articles were also examined. Searches were carried out in October 2010 and there were no restrictions by study design or country of origin. Publication date of included studies was limited to 1990-2010.

INTRODUCTION

Approximately 12,000 new cases of nonfatal spinal cord injury (SCI) occur per year in the US alone.[1] Most of the spinal column injuries (55%) involve the cervical spinal column and 15% involve the thoracolumbar junction.[2] Injuries are generated from traumatic and non-traumatic causes. Almost half of the SCIs are caused by motor vehicle collisions, followed by falls, acts of violence (primarily gunshot wounds) and sporting accidents [Table 1]. Majority (80%) of the reported SCIs occur among males with an average age of 40.2 years. Most SCIs are associated with an injury to the vertebral column as well as a co-existing traumatic injury to other regions. This review article explores the up-to-date anesthetic management strategies of patients with acute SCI. A comprehensive literature search of Medline, Embase, and Cochrane database of systematic reviews, conference proceedings and internet sites for relevant literature was performed. Reference lists of relevant published articles were also examined. Searches were carried out in October 2010 and there were no restrictions by study design or country of origin. Publication date of the included studies was limited to 1990-2010.

Table 1

Etiology of spinal cord injuries

CLASSIFICATION

SCI develops in two stages – initial or primary injury, and later, secondary injury.[3-5] Primary injury occurs at the time of the traumatic insult. Sources of mechanical force may include bony fragments (e.g. vertebral body), joint dislocation (e.g. facet joints, intervertebral joints) or arthropathy (spondylosis, spondylolisthesis), ligamentous tears or herniation of intervertebral discs.[6] Secondary injury begins within minutes following the initial injury and can evolve over several hours. Mechanisms include ischemia, hypoxia, inflammation, excitotoxicity, lipid peroxidation and apoptosis of neurons. This underlying pathological process results in further cord edema and it reaches its maximum between 4 and 6 days after the injury.[78]

During the secondary survey, patient's neurological status should be documented to guide further management. The American Spinal Injury Association (ASIA) score is the grading scale for essential elements of neurological assessment for all patients with a spinal injury. These essential elements consist of bilateral strength assessment of 10 muscle groups and pin-prick discrimination assessment of 28 specific sensory locations. ASIA grade A refers to complete loss of motor and sensory function, whilst ASIA grade E refers to intact motor and sensory function. Grades B, C and D refer to progressively less severe involvement of motor and sensory pathways.[9]

AIRWAY MANAGEMENT

Airway management of the patient with known or suspected traumatic SCI can be categorized either as an emergent management of respiratory failure or airway protection as part of Advanced Trauma Life Support (ATLS) algorithm, or the perioperative management of the patient with known cervical injury for his/her spinal or other associated injuries.

SCI occurs in up to 2–5% of all major trauma cases and at least 14% of these cases have the potential to be unstable.[10] Techniques to minimize cervical movement should be employed during emergent management of the airway while considerations should be given to other potential injuries. Cervical SCI occurs in up to 10% of head injured patients.[11] As such, a high index of suspicion should be employed during airway management of the traumatically injured patient, especially with head injury.

Spinal immobilization should be employed in all trauma victims. Although there is little evidence to support its practice, missing a diagnosis of a spinal injury may have dire consequences for the trauma victim. Immobilization of patients with SCI during the prehospital setting should include neck immobilization with a cervical collar, lateral supports and straps, and spinal hardboard. Time spent on the hardboard should be minimized and patients should be transferred off the hardboard on admission to a facility as soon as it is feasible. If there is a delayed transfer to another institution, the patient should be temporarily taken off the hardboard. Padded boards or inflatable bean bag boards should be utilized to reduce pressure on the occiput and sacrum.[12] Cervical collar, lateral supports and straps are essential components of C-spine immobilization and must be applied until it can be cleared by an appropriate senior clinician as soon as it is possible. Collar use in isolation could lead to the possibility of cervical spine misalignment. Semirigid collars should be preferred to reduce collar-associated injuries.

Current evidence suggests that in subjects without a cervical injury, direct laryngoscopy causes extension of the cervical spine, mostly at the atlanto-occipital junction, and to a lesser extent at the C1 to C2 joint.[1314] The subaxial cervical segments from C4 through C7 are minimally displaced but it generates flexion at the cervico-thoracic junction.[15] During laryngoscopy, pressure exerted by the blade on airway structures is consequently transmitted to the spinal cord.[16] Instability of the occiput–atlas–axis complex may lead to anterior movement of the atlas during direct laryngoscopy, thereby reducing the space available for the spinal cord. Disruption involving subaxial cervical vertebrae may act as a second focus for extension during laryngoscopy, in addition to extension at the atlanto-occipital junction.

The urgent nature of airway interventions usually requires direct or indirect laryngoscopy with manual in-line stabilization (MILS). The goal of MILS is to apply sufficient opposite forces to the head and neck to limit the movement during airway intervention. MILS is recommended by current ATLS guidelines as a standard for airway intervention in patients with known or suspected cervical injury.[17] MILS reduces cervical movement better than a rigid collar during laryngoscopy. Also, an improved view is obtained at laryngoscopy with MILS compared to a hard collar, owing to better mouth opening.[1819] However, a randomized controlled trial in normal subjects without cervical injury showed MILS to increase the tracheal intubation failure rate at 30 seconds and worsened laryngeal visualization during direct laryngoscopy.[20] Another study conducted in normal subjects looked at the pressures transmitted to airway tissue with and without MILS. Investigators detected significantly increased pressures when MILS was applied.[21] Nevertheless, when MILS is utilized, incidence of neurological impairment due to endotracheal intubation has been reported to be extremely rare.[22] Care should also be taken with mask ventilation. The absolute minimum of jaw thrust and chin lift should be used to maintain the patient's airway.[2324]

The use of cricoid pressure in this situation remains somewhat controversial despite a cadaver model of upper cervical injury showing that cricoid pressure did not result in significant movement.[25]

Perioperative intubation of the patient presenting for operative fixation of his/her spine or associated injuries affords time to plan for successful airway intervention. The patient may present in cervical traction or a halo, providing a physical obstacle impeding access to the airway. Patients with pre-existing cervical spinal pathology including spondylosis, rheumatoid arthritis, Klippel-Feil, ankylosing spondylitis, tumor, pre-existing cervical instrumentation and upper (vs. lower) cervical disease may increase the difficulty of airway interventions.[26] Several varieties of intubation techniques exist, but no one technique has been proved to be superior to others. Awake intubation has not been shown to be superior to asleep intubation.[27-29] Awake intubation enables a neurological exam to be performed after intubation and positioning; however, it requires a cooperative patient. Patient without an existing neurological impairment and acceptable radiological findings can be managed with direct laryngoscopy. If the airway is potentially difficult and the patient has an existing neurological deficit and actual C-spine instability, an alternative should be considered. A reinforced endotracheal tube reduces the risk of kinking during patient positioning and also prevents tracheal compression during retraction employed in anterior cervical procedures.

Rigid indirect videolaryngoscopy has become an alternative to conventional direct laryngoscopy with the The GlideScope®(GS) (Verathon Medical, Bothell, WA, USA) and the Airtraq®(AT) (Prodol Ltd., Vizcaya, Spain) being the two popular devices amongst others. Numerous studies currently exist but it is difficult to draw conclusions at present due to heterogeneity of the studied population. Studies in normal subjects with either MILS or hard collar in situ to simulate cervical immobilization showed that the AT performed better than direct laryngoscopy, but GS compared to direct laryngoscopy yielded conflicting results.[3031] GS prolonged the intubation time for experienced laryngoscopists.[3233] Studies looking at cervical motion suggested less movement occurred with AT,[3435] while GS produced equivocal results.[3336]

Supraglottic devices are part of the failed or difficult intubation algorithm. In the “can’t intubate, can’t ventilate” scenario, early consideration should be given to the surgical airway or cricothyroidotomy.[37] These techniques may still produce critical movement of the cervical spine, but this should not prevent their use for life-saving procedures.

The gum elastic bougie is a useful adjunct during direct laryngoscopy. It allows accepting higher-grade laryngoscopy views of the vocal cords, thereby limiting the forces transmitted to the cervical spine.

The decision to extubate immediately postoperative is influenced by many factors. These include the extent of surgery, surgical complications (e.g. recurrent laryngeal nerve injury), duration of the procedure, prone positioning, the extent of blood loss and subsequent fluid resuscitation, and ease of intubation. The presence of a cuff leak demonstrated on either inspiration or expiration in the spontaneously breathing patient has not consistently been shown to predict subsequent airway obstruction from edema. Extubating with an airway exchange catheter in situ will facilitate emergent re-intubation in the event of an obstruction from airway edema or hematoma. Good clinical judgment is necessary, and if there is concern, the patient should be extubated at a later time.

Succinylcholine should be avoided between 3 days and 9 months following SCI due to the risk of succinylcholine induced hyperkalemia caused by denervation hypersensitivity.[38] Rocuronium should be considered as an alternative.

BLOOD PRESSURE MANAGEMENT

Traumatic SCI is frequently complicated by systemic hypotension and reduced spinal cord perfusion pressure (SCPP).[39] This in turn contributes to the worsening of secondary neurologic injury and should be avoided. Both the anterior and posterior spinal arteries arise from the vertebral arteries in the neck and descend from the base of the skull to supply spinal cord blood flow. There are also various radicular branches off the thoracic and abdominal aorta to provide segmental contributions. SCPP is determined by the difference in mean arterial pressure (MAP) and cerebrospinal fluid pressure (CSFP) (SCPP = MAP – CSFP).

Spinal cord perfusion is autoregulated over a wide range of systemic blood pressures in the same fashion as cerebral blood flow.[40] Systemic hypotension most commonly results either from hemorrhage due to associated traumatic injuries (chest, intra-abdominal, retroperitoneal, pelvic or long bone fractures) or neurogenic shock, or a combination of both. Neurogenic shock refers to the hypotension and inadequate tissue perfusion as a result of vasodilatation and loss of central supraspinal sympathetic control.[41] It is more common after cervical SCI and is usually associated with bradycardia from unopposed vagal tone. Spinal cord blood flow has been shown to be adversely affected following traumatic SCI, and an increase in blood pressure leads to significant improvement in axonal function both in the motor and somatosensory tracts of the cord.[4243] Similarly, several clinical studies involving aggressive hemodynamic goal-directed management in patients with SCI have suggested improved outcome.[44]

Currently, there is little evidence regarding targeted blood pressures and the duration of support required to improve outcome in SCI.[45] American Association of Neurological Surgeons (AANS) published recommendations for the hemodynamic goals for a SCI patient. These include maintaining MAP to 85–90 mmHg and avoiding SBP less than 90 mmHg (Class 3 evidence) for over 5–7 days.[39] Injuries at cervical and upper thoracic levels down to T6 warrant an agent with inotropic, chronotropic as well as vasoconstrictive properties. Agents such as dopamine, norepinephrine, or epinephrine fulfil these requirements with their α1- and β1-agonist properties.[46] Phenylephrine preferentially works as an α1-receptor agonist with minimal β1effects. It can be used to counteract the peripheral vasodilation associated with lower thoracic and lumbar cord injuries. Caution is warranted with its use due to the potential for developing reflex bradycardia. Dobutamine exerts its effect prominently as an inotropic agent and its use in SCI is limited because of its effect on vasodilation and possible reflex bradycardia.[4647] Persistent bradycardia may be seen in high cervical (C1 through C5) lesions in the first 2 weeks after traumatic SCI and requires the use of anticholinergic agents or application of pacemakers.[4849] These parameters should be maintained throughout the perioperative period and will require judicious use of intravenous fluids, vasopressors and inotropes. A retrospective review on anterior cervical fusion cases found that intraoperative deterioration of evoked potential monitoring was associated with hypotension in 1% of cases.[50] In their prospective randomized trial, Kwon et al. reported that intrathecal pressure may become elevated in the postoperative period of SCI which may lead to concomitant reduction in spinal cord perfusion. These findings raise the question of appropriateness of intrathecal pressure monitoring as in traumatic brain injury patients.[51] Consideration should be given during airway management of the patients with neurogenic shock. Stimulation of airway tissues may result in profound bradycardia, hypotension and cardiac arrest.[52]

FLUID MANAGEMENT

Perioperative fluid management in patients with SCI requiring operative intervention can be challenging. Procedures involving multiple levels on the thoraco-lumbar spine may incur large blood loss and necessitate subsequent need for transfusion of blood products. In elective spinal surgery, pre-operative hemoglobin level less than 12 g/dl, age greater than 50 years and procedures requiring transpedicular osteotomy are identified to predict blood transfusion requirements.[53] Also, patients requiring spinal surgeries with instrumentation have greater blood loss potential than those requiring non-instrumented spine surgeries.[54]

Major blood loss may lead to blood, platelet, and factor transfusions. There are several known risks of blood component transfusion, including potential transfusion reactions and alloimmunization as well as infectious risks such as hepatitis, human immunodeficiency virus, cytomegalovirus, transfusion-associated bacterial sepsis and increased risk of postoperative infections. Additionally, the costs of blood replacement must be considered.

Several strategies have been used to minimize intraoperative blood loss. Use of the Jackson table, where the abdomen hangs free from compression, reduces the vena cava pressure which in turn lessens epidural venous bleeding when compared to positioning prone on the Wilson frame.[55] There are currently no comparative studies existing to evaluate the effectiveness of normovolemic hemodilution or hypotensive anesthesia.[56] Furthermore, hypotensive anesthesia should be avoided in patients with SCI, as it may exacerbate the secondary SCI. Antifibrinolytic agents have been shown to decrease intraoperative and total perioperative blood loss. A randomized study in patients undergoing antero-posterior spinal fusion comparing aminocaproic acid versus aprotonin versus control showed an absolute decrease in both total perioperative blood loss and transfusion requirements compared to control. For the aminocaproic acid group, however, these did not achieve significance.[57] As of 2008, aprotonin has been withdrawn from clinical use. A randomized trial of tranexamic acid versus placebo in patients undergoing posterior thoracic or lumbar instrumented fusions showed significantly less perioperative blood loss compared to placebo; however, there was no difference in the amount of blood products transfused between the two groups.[58] There was no increase in thromboembolic complications.

A small randomized dose escalation trial using recombinant factor VIIa (rFV11a) in multiple level posterior spinal fusion showed an absolute decrease in intraoperative blood loss but no significant decrease in transfusion requirements for the rFVIIa groups at any dose studied.[59] One thromboembolic event leading to death was reported in the rFVIIa group.

Studies reporting the effectiveness of cell saver in reducing the need for homologous transfusion have shown variable results. Three retrospective cohort studies have looked at this issue in spine surgery.[60-62] One study reported no decrease in the need for homologous transfusion, and in fact found a significant increase in the use of homologous blood. Sources of bias identified in this study include less meticulous hemostasis in the cell saver group, and patients having procedures with less anticipated blood loss might have been selected to not have cell saver used for their case. The other two studies revealed less need for homologous transfusion associated with the use of cell saver. A recent systematic review concluded that there is little in the literature to support the cost-effective use of cell saver in routine elective spine surgery.[56]

Thoraco-lumbar spinal surgery frequently involves large blood loss and subsequent fluid administration. Excessive fluid administration in the prone patient is associated with significant edema (including airway edema), cardiac failure, electrolyte abnormalities, coagulopathy and prolonged duration of postoperative intensive care unit stay.[63] Optimal fluid therapy in SCI patients remains unknown. Hypotonic crystalloids such as D5W and 0.45% NS, however, may exacerbate cord swelling and should be avoided. Albumin use is relatively contraindicated following recommendations from SAFE-TBI study of increased mortality in patients with traumatic brain injury.[6465] Goal-directed treatment using cardiac output monitoring devices might improve intraoperative fluid administration and possibly reduce the morbidity associated with excessive fluid administration.

Coagulopathy may happen during spine surgery following massive blood transfusion. Significant blood loss should be anticipated when multilevel instrumentation is performed. Transfusion rate is reported to be 50-80% without the preventative strategies. Homeostasis assays should be used to evaluate the ongoing blood loss during surgery.

Results obtained from standard coagulation testing are too slow to be used in actively bleeding surgical patients. Viscoelastic point-of-care coagulation assays have been reported to be useful in penetrating trauma by providing rapid turnaround time and overall assessment of the hemostatic system including coagulation factors, platelets, and fibrinolysis.[66] The drawbacks of point-of-care testing are lower accuracy, errors in predicting need for transfusion, lack of whole blood controls, performance by non-laboratory-trained staff and cost.[67-69]

It is important that in severe ongoing hemorrhage, blood assays should produce results rapidly and accurately. Emergency hemorrhage panel (EHP) was developed recently to produce accurate results and a short turnaround time. Prothrombin time (PT), fibrinogen, platelet (PLT) count, and hematocrit (Hct) have been identified as pivotal factors to guide the transfusion decisions included in EHP assessment. This approach will appropriately guide the transfusion strategies and prevent unnecessary donor product exposures.[70]

EVOKED POTENTIAL MONITORING

Since the Stagnara wake-up test provides only a single time point of assessment of neurological function,[71] modern intraoperative assessment techniques have been developed to provide continuous monitoring of evoked potentials (sensory and motor) and spontaneous electromyography (EMG). When used collectively, these are referred to as multimodal intraoperative monitoring (MIOM).

Somatosensory evoked potentials (SSEP) are elicited by stimulation over peripheral nerves and gauging the response at some point along the sensory pathway, usually at the somatosensory cortex. Motor evoked potential (MEP) monitoring involves transcortical stimulation over the motor cortex and recording the muscle response in the respective muscle groups. Spontaneous EMG activity is recorded by electrode placement in the muscle innervated by the nerve to be monitored. It is particularly useful in monitoring the mechanical irritation of nerve roots.

The proposed benefit of evoked potential monitoring is to identify the deterioration of the spinal cord function, thus giving the opportunity to correct offending factors. Such factors include patient position (e.g. neck position, shoulder position), hypotension, hypothermia, and factors related to the surgical intervention. Identification and correction of these factors may prevent the deterioration of the spinal cord function from becoming irreversible. In elective spinal surgery without EP monitoring, iatrogenic neurological injuries have been estimated to be 0.46% for anterior cervical discectomy, 0.23–3.2% with scoliosis correction, and between 23.8 and 65.4% with intramedullary spinal cord tumor resection.[72] A recent systematic review indicated that although there is a high level of evidence that MIOM is sensitive and specific for detecting intraoperative neurologic injury during spine surgery, there is a low level of evidence that MIOM reduces the rate of new or worsening perioperative neurological deficits and there is very low evidence that an intraoperative response to a neuromonitoring alert reduces the rate of perioperative neurologic deterioration.[73]

The major implication of electrophysiologic monitoring on anesthetic technique is the requirement of total intravenous anesthesia for MEP monitoring. Volatile anesthetic may be used when SSEPs are being monitored, provided their dosing does not exceed 1 MAC. Volatile anesthetics may also be used for spontaneous EMG recording, provided muscle relaxants are avoided. Volatile anesthetics and nitrous oxide are best avoided and a total intravenous technique without muscle relaxation is used when MEP monitoring is performed.[74] Volatiles cause a dose-dependent reduction in MEP signal amplitude, commencing at low concentrations. Opioids do not impact evoked potential monitoring. Ketamine has been shown to enhance evoked potential monitoring.[75] Dexmedetomidine has been used as a supplement to TIVA, allowing reduction of propofol dose, with no evidence of detriment to evoked potential monitoring.[7677] A stable anesthesia without significant changes in blood pressure or dosing of anesthetic agents needs to be provided so that changes in evoked responses may be attributed solely to surgical technique.

CORTICOSTEROIDS

Methylprednisolone (MP) has been used as a treatment option in acute, non-penetrating SCI.[22] MP stabilizes cell membranes through inhibition of lipid peroxidation, which reduces ischemia and necrosis.[2324] In addition, when used in high doses, MP, with its anti-inflammatory properties, decreases the release of interleukins, prostaglandins, and thromboxanes. All of these help to increase the perfusion to injured areas of the spinal cord, decrease edema, improve impulse generation, protect the blood–spinal cord barrier, and have a positive effect on electrolyte concentrations.[525]

Three multicenter, double-blind, randomized clinical trials were conducted to investigate the efficacy of MP in acute SCI: The National Acute Spinal Cord Injury Studies (NASCIS) I, II, and III published in 1984, 1990, and 1997, respectively.

A bolus dose of 30 mg/kg over 1 hour, followed by an infusion of 5.4 mg/kg/hour continued for either 23 or 47 hours is used.[7879] This dosing, if commenced between 3 and 8 hours following injury, was shown in a subsequent meta-analysis to be associated with greater motor, but not functional recovery, compared to other treatments.[80] Subsequent reviews on this topic have revealed deficiencies in trials on which these conclusions have been based.[81] These studies were underpowered and lacked placebo arms to detect treatment effects and also revealed an increased risk of serious side effects including wound infections, steroid myelopathy and gastrointestinal hemorrhage with high dose MP.

Despite the modest benefits of corticosteroids, a 2006 survey of 305 neurosurgeons in the US revealed that 91% of clinicians used them for treatment of nonpenetrating traumatic SCI within 8 hours.[82] Currently, the guidelines of the American Association of Neurological Surgeons/CNS Joint Section on Disorders of the Spine and Peripheral Nerves Guidelines committee recommend the use of MP only as a treatment option (while considering the risks and benefits associated with glucocorticoid use) and not as the standard of care.[23]

CONCLUSION

The most important anesthetic management principles in the treatment of spine trauma are high index of suspicion for early detection and prevention of secondary injury through adequate oxygenation, blood pressure support through volume replacement (and, if necessary, cardiovascular support), and immobilization. Intrathecal pressure monitoring may be beneficial to improve SCPPs in critical care areas. It is important to employ strategies to minimize intraoperative blood loss. Homeostasis assays with rapid turnaround time should be used to evaluate the ongoing blood loss during surgery. Multimodality neuromonitoring is sensitive and specific enough for detecting intraoperative neurologic injury during spine surgery and choice of anesthetic has an impact on its quality. Corticosteroids may be used only after careful consideration of associated risk and benefits.