critical care and perioperative management in …...review article critical care and perioperative...

Review Article

Critical Care and Perioperative Management in Traumatic SpinalCord Injury

*Robert D. Stevens, *†Anish Bhardwaj, *†Jeffrey R. Kirsch, and *†Marek A. Mirski

From the Neurosciences Critical Care Division, Departments of *Anesthesiology/Critical Care Medicine and †Neurology, JohnsHopkins University School of Medicine, Baltimore, Maryland

Summary: Traumatic spinal cord injury is frequently associated with brain injury andwith alterations in respiratory and cardiovascular function that require critical care man-agement. Complications include respiratory failure, atelectasis, pneumonia, neurogenicshock, autonomic dysreflexia, venous thromboembolism, and sepsis. While complica-tions may be managed with supportive care, the goal of ameliorating neurologic outcomehas proved elusive. Methylprednisolone, when instituted <8 hours after traumatic spinalcord injury, was associated in two clinical trials with statistically significant improve-ments in motor scores at 6 months and 1 year; however, critical reappraisal of these dataraises questions about their validity and clinical relevance. Until more evidence ofclinically effective therapies is available, acute management must be driven by patho-physiologic principles, with emphasis on interventions that attenuate secondary neuro-logic injury; these include the rational use of immobilization, cautious airway manage-ment, and promotion of cord perfusion and oxygenation with the appropriate level ofhemodynamic and respiratory support. Clinical trials of pharmacologic neuroprotectionhave yielded disappointing results, but the ongoing elucidation of spinal cord repair andregenerative mechanisms suggests new therapeutic prospects. Key Words: spinal cordinjury, critical care, anesthesia, perioperative management

EPIDEMIOLOGY

Acute spinal cord injury (SCI) is most commonly trau-matic but may also result from degenerative spine disease,ischemia, demyelination, inflammation, and rapidly ex-panding neoplastic, hemorrhagic, or pyogenic masses.1

This review is centered on conditions that are encounteredin a critical care or perioperative setting, in particular,traumatic SCI. The acute ischemic myelopathy that may

be observed as a complication of surgery of the descend-ing aorta is reviewed elsewhere.2,3

In the United States, traumatic SCI with or withoutbony injury has an annual incidence of 28 to 55 per mil-lion, with an ave rage of 10,000 new cases a year and aprevalence of 200,000. 4 The actual incidence may beeven greater because reported statistics do not include in-dividuals who die before admission to the hospital andwho account for 16% to 30% of cases.5 The average ageat the time of injury is 32 years and the male/female ratiois 4:1. More than half (55%) of traumatic SCI involves thecervical cord. The most common causes of SCI are trafficaccidents (motor vehicle, bicycle, pedestrian) (40%–50%), assault (10%–25%), falls (20%), work-related inju-ries (10%–25%), and sports/recreation-related injuries(10%–25%).3,5 In the United States, the total annual costsof treating patients with SCI is estimated at four to nine

Address correspondence and reprint requests to Robert Stevens, MD,Neurosciences Critical Care Division, Meyer 8-140, Johns Hopkins Hos-pital, 600 N. Wolfe St., Baltimore, MD 21287 (e-mail: [email protected]).

This work was supported in part by the U.S. Public Health ServiceNational Institute of Health grant NS 20020 (Dr. Kirsch). Dr. Bhardwajis an Established Investigator Awardee of the American Heart Associa-tion.

Accepted for publication on April 28, 2003.

Journal of Neurosurgical AnesthesiologyVol. 15, No. 3, pp. 215–229© 2003 Lippincott Williams & Wilkins, Inc., Philadelphia

215

billion dollars. 4 In traumatic cervical SCI, 3-month mor-tality is 20% to 21%, and independent predictors of mor-tality are level of cord injury, Glasgow Coma Scale, age,and respiratory failure.6,7 The most consistent predictor oflong-term outcome is the severity of neurologic injury,which may be characterized by the level and completenessof sensorimotor loss.8–11 Principal causes of death are re-spiratory disorders, cardiovascular disorders, pulmonaryembolism, infections, and suicide.11,12 Survival of patientswith traumatic SCI has improved over the past de-cades.13,14 In a cohort of patients with SCI followed over40 years, life expectancy of complete tetraplegics was70% of that expected in comparable noninjured subjects,whereas the life expectancy was 84% in complete paraple-gics and 92% in patients with incomplete SCI.12 Betterprognosis is thought to reflect changes in prehospital care,rapid triage to facilities with SCI expertise, and advancesin medical, surgical, and rehabilitative care; however, spe-cific factors linked to improved outcome have not beenclearly identified.

ANATOMY

The adult spinal cord15 extends from the medulla ob-longata to the L1–L2 vertebral level. It is encased in thespine whose individual segments are linked by facet jointsand intervertebral discs and bound by an extensive liga-mentous apparatus. The cervical spine and thoracolumbarjunction have the greatest mobility and are most vulner-able to injury, whereas the incidence of fractures is muchlower at the thoracic level. The spinal cord is enveloped inthree layers of meninges; however, unlike the brain, thedura at the cord level is delimited by a densely vascularand relatively compliant epidural space into which patho-logic processes such as abscesses, hematoma, and metas-tases may spread by a path of least resistance. Transversesections of the cord reveal a characteristic H-shaped massof gray matter enfolded in a mantle of white matter madeof ascending and descending tracts. Of the many whitematter tracts, only three can be readily assessed by physi-cal examination: 1) the posterior columns conveying finetouch, vibration, and position sense, 2) the posterolateralcorticospinal tract transmitting motor commands, and 3)the anterolateral spinothalamic tract containing fibers forpain, temperature, light touch, and pressure sensation.White matter columns are organized somatotopically, withmedial cervical fibers and lumbosacral elements placedlaterally. Spinal cord blood supply is provided by oneanterior and two posterior spinal arteries, which are

branches of the vertebral arteries and are augmented byradicular branches of the thyrocervical, costocervical, in-tercostal, and lumbar vessels. The anterior spinal arterysupplies the ventral two thirds of the cord, the remainingdorsal region being supplied by the posterior arteries. Tis-sues at the boundary between anterior and posterior spinalarteries and between thoracic segmental arteries are at riskfor “watershed” type infarction. Venous blood from thecord is drained by internal and external vertebral plexusesthat ultimately run into the intracranial venous sinuses orinto the caval system. Spinal cord blood flow (SCBF) hasmany of the characteristics of cerebral blood flow, aver-aging 40 to 60 mL · 100 g−1 · min−1, autoregulating be-tween mean arterial pressures of 60 and 150 mm Hg andincreasing with hypercapnia and severe hypoxemia.16 Asin traumatic brain injury, SCI is associated with loss ofautoregulation.17

PATHOPHYSIOLOGY

Traumatic SCI is commonly viewed as a sequence ofprimary and secondary pathogenetic events.18 Primary in-jury is the damage incurred during the initial insult; it istypically maximal at onset and unlikely to be modified bytherapeutic intervention. Secondary injury unfolds duringthe hours to days following the insult; it refers to a cascadeof tissue injury, including 1) vascular compromise withimpaired vasomotor function, ischemia, hemorrhage, va-sospasm, thrombosis, and increased permeability,19 2) in-flammatory changes with release of chemokines, cyto-kines, and eicosanoids and expression of cell adhesionmolecules and leukocyte infiltration, and 3) cellular dys-function with adenosine triphosphate depletion, plasmamembrane failure, free radical generation, lipid peroxida-tion, excitatory amino acid release, cellular calcium over-load, and mitochondrial insufficiency. A hallmark of sec-ondary injury is spinal cord edema, which may presentclinically as neurologic deterioration and on magneticresonance imaging (MRI) as parenchymal signal abnor-malities; cord edema typically peaks 3 to 6 days afterinjury and subsides over a period of weeks. Beyond theseacute changes, SCI may continue to unfold weeks andmonths after injury, namely, through apoptotic cell death,glial scar formation, and the generation of cystic cavi-ties.20,21

The clinical significance of secondary cord injury is thatit is exacerbated by systemic variables such as hypo-tension, shock, decreased arterial oxygen content, cate-cholamine release, hypercoagulability, and hyperthermia.Aggressive prevention and/or correction of such ab-

Stevens et al216

Journal of Neurosurgical Anesthesiology, Vol. 15, No. 3, 2003

normalities are accordingly cardinal goals for the inten-sivist and anesthesiologist.

DIAGNOSIS

Clinical AssessmentClinical assessment of SCI begins with an evaluation of

airway, breathing, and circulatory function. Physical ex-amination should be attentive to noncord injuries, whichare present in 20% to 60% of patients with SCI, mostsignificantly involving the head, chest, or abdomen.4

Traumatic brain injury occurs in 25% to 50% of patientswith SCI; conversely, 5% to 10% of head-injured patientshave an associated SCI.22,23

The physical examination performed in the first daysafter traumatic SCI is highly predictive of short- and long-term outcome.6,7,8–10 The spine should be inspected fordeformity and palpated to elicit tenderness or a “step-off,”indicating a widened interspinous space. Motor and sen-sory function and deep tendon reflexes should be docu-mented. Careful note should be made of changes in cog-nitive function, level of consciousness, and cranial nerveabnormalities that might signal concomitant brain orbrainstem injury. Early neurologic findings may be con-founded by spinal shock, characterized by flaccid areflexicparalysis and anesthesia to all modalities. Spinal shock ispresent in one half of patients with SCI, and its pathogen-esis is poorly understood; it has a variable time course butin most cases resolves in <24 hours. Although the termsare sometimes used interchangeably, it is preferable todistinguish spinal shock from neurogenic shock. Spinalshock refers to an acute, transient neurologic syndrome ofsensorimotor dysfunction that develops with SCI at anylevel, whereas neurogenic shock is a hemodynamic syn-drome associated with upper thoracic and cervical SCI andis characterized by bradycardia and decreased systemicvascular resistance. The two patterns may or may not oc-cur concurrently.

The neurologic level is defined as the most caudal seg-ment of the spinal cord with normal bilateral motor(strength >3/5) and sensory (light touch and pinprick)function. An injury is said to be complete when all sensoryand motor function below the lesion is abolished; in com-plete injury, abnormal motor or sensory findings subjacentto the neurologic level are referred to as the zone of partialpreservation. Any residual voluntary motor function orsensation that is not contiguous with the neurologic levelqualifies as incomplete injury. Standardized neurologicassessment tools have been developed for SCI, the mostwidely used of which is the American Spinal Injury As-

sociation (ASIA) classification, composed of a motorscore that grades strength in 10 muscle groups in the upper(C5–T1) and lower (L2–S1) extremities, and a sensoryscore based on the light touch and pinprick responses in 28dermatomes.24 A related tool, the ASIA impairment scale,combines severity of deficits with completeness of injury,stratifying patients into five classes that correlate with out-come (Table 1).24

Depending on anatomic localization, a number of clini-cally distinct spinal cord syndromes have been describedand are outlined in Table 2.

ImagingThe primary goal of imaging is to rapidly and accu-

rately identify injury of the spine that places neural tissueat risk. Imaging should be obtained in all trauma patientswith risk factors for spine or cord injury; these includepatients with 1) neck or back pain or tenderness, 2) sen-sory or motor deficits, 3) impaired level of consciousness,4) alcohol or drug intoxication, and 5) painful injuries thatmay distract the patient from spine injury. Greater than95% of patients with cord injury have concomitant spineinjury (fracture and/or dislocation).25 Patients with cord-induced deficits who do not have any spine abnormality orinjury detectable with plain radiographs or computed to-mography (CT) are referred to as SCI without radio-graphic abnormality (SCIWORA). Originally described inchildren, SCIWORA constitutes up to 15% of adults withSCI; however, the incidence may be lower if magneticresonance imaging (MRI) is used to diagnose occult in-tervertebral disc or ligamentous injury.26,27

TABLE 1. ASIA impairment scale

Grade DescriptionIncidence

(%) Outcome

A Complete motor andsensory loss

25 10–15% convert togrades B–D; 3% tograde D

B Incomplete sensoryloss, completemotor loss

15 54% convert to gradeC–D

C Incomplete motor andsensory loss; >50%of muscles <3/5

10 86% of grades C–Deventually regainambulating ability

D Incomplete motor andsensory loss; >50%of muscles �3/5

30

E Normal motor andsensory function

ASIA � American Spinal Injury Association.From references 8–10 and 24.

Critical Care of Spinal Cord Injury 217


RadiographsStandard cervical spine radiographs include the antero-

posterior, lateral, and odontoid views. Lateral films shouldbe assessed in a systematic fashion for 1) technical ad-equacy with particular attention to whether or not the en-tire cervical spine and the C7–T1 intervertebral space arevisualized and 2) for abnormalities in vertebral alignment,bony structure, intervertebral space, and soft tissue thick-ening.

Computed TomographyCT is indicated if the spine is not adequately visualized

or appears abnormal on plain radiographs, or if there is ahigh clinical suspicion of injury despite technically ad-equate films that show no injury. In a recent prospectivestudy, 67 of 70 patients with C1–3 injuries were correctlyidentified by CT with 2 mm cuts, versus 38 of 70 correctlyidentified by plain films.28 Helical CT allows sagittal andthree-dimensional reconstruction and may provide agreater degree of accuracy.29 However, these studies havebeen construed as flawed because they lack a gold stan-dard of clinically relevant spine injury against which theimaging techniques may be measured.30 Although the roleof CT as a single diagnostic modality is uncertain, thenegative predictive value of combined of three-dimensional view plain films and CT is >99%.28,31,32

Magnetic Resonance ImagingMRI is the modality of choice for characterizing acute

cord injury, but it may be less sensitive than CT or plainfilms in distinguishing bony abnormalities, in particular,those of the posterior portion of the spine.33 MRI detectsspinal cord edema and hemorrhage, ligamentous injury,and other nonosseous changes that can be missed withother techniques; however, the interpretation of many liga-mentous lesions identified with MRI is unclear, and con-cern exists that MRI may “overcall” significant injury.30

Clearing the Cervical SpineInitial management of trauma patients assumes that the

cervical spine is unstable until proven otherwise. Rapidlyestablishing stability, or “clearing,” of the cervical spine isa management priority and should be undertaken in a sys-tematic fashion. All patients with trauma should bescreened for clinical risk factors of cord or spine injury.Prospective studies have shown that when all five clinicalrisk factors (neck pain or tenderness, neurologic deficits,impaired level of consciousness, intoxication, or painfuldistracting injuries) are absent, cervical spine injury maybe ruled out with a high degree of confidence (negativepredictive value >99%),34–36 and recent neurosurgicalguidelines emphasize that radiographic assessment bewithheld in patients that meet these exclusionary crite-ria.37

Patients with clinical risk factors should be imaged withthree-view radiographs supplemented by CT if plain ra-diographs are abnormal or inadequate. Those with neckpain or tenderness and normal three-view series may beassessed with dynamic flexion–extension radiographs or

TABLE 2. Spinal cord syndromes

Syndrome Setting Clinical findings

Complete cord(cordtranssection)

Trauma, infarction,hemorrhage, discherniation,transverse myelitis,tumor, abscess

Loss of all motor andsensory function

Cord transsectionabove C3 results inapnea and deathunless promptresuscitation

Brown-Sequard(cordhemisection)

Penetrating trauma,multiple sclerosis,tumor, abscess

Ipsilateral loss ofproprioception andmotor function

Contralateral pain andtemperature loss

Suspended ipsilateralloss of all sensorymodalities

Central cord Neck hyperextension,syringomyelia,intramedullarytumor

Motor impairmentgreater in upperthan lowerextremity

Suspended sensoryloss incervicothoracicdermatomes

Anterior cord Hyperflexion, discprotrusion, anteriorspinal arteryocclusion

Pain and temperatureloss with sparing ofproprioception

Variable motorimpairment

Posterior cord Syphilis, vitamin B12deficiency,posterior spinalartery disease,trauma, multiplesclerosis

Diminishedproprioception andfine touch

Conus medullaris Tumor, trauma, discherniation,inflammation,infection

Extension tolumbosacral rootsmay produce bothupper and lowermotor neuron signs

Spastic paraparesis,sphincterdysfunction, lowersacral “saddle”sensory loss

*Not a myelopathy but a radiculopathy or neuropathy involving lum-bosacral nerves.

Stevens et al218


MRI. Negative flexion–extension films or negative MRIwithin the first 48 hours of injury, together with normalplain radiographs and CT, are thought to be sufficient toclear the cervical spine. In comatose or intoxicated pa-tients with normal plain films and CT, the incidence ofspine injury is <1%; if suspected clinically, flexion–extension views obtained under fluoroscopic guidanceor MRI if performed <48 hours after injury may be use-ful.38–40 MRI is highly sensitive but lacks specificity, andfalsely positive studies may lead to undue prolongation ofimmobilization and its inherent risks.

CRITICAL CARE MANAGEMENT

Management of SCI is directed toward limiting second-ary injury and maximizing neurologic recovery. This maybe achieved surgically by relieving or preventing com-pression of neural structures and medically by optimizingcord perfusion and oxygenation and by disrupting cellinjury mechanisms.

Concept of Spinal StabilityStability is the capacity of the spine to withstand physi-

ologic loading and positioning without neurologic injury,deformity, or pain.41 Stability is predicted using clinicaland anatomic characteristics and is critical in decisionsregarding airway and surgical management. For the tho-racolumbar spine, a three-column concept based on CTimaging is widely used; damage to two or more columnsresults in instability.42 For the cervical spine, stability isscored according to a system that integrates static anddynamic clinical and radiographic criteria.43

ImmobilizationThe goal of immobilization is to prevent or limit sec-

ondary neurologic injury in the presence of an unstablespine. Because spinal injury can occur at several noncon-tiguous levels, immobilization of the entire spine is rec-ommended until injury is ruled out with appropriate physi-cal examination and imaging. Common measures includeimmobilization of the head between two sandbags, place-ment of a rigid cervical collar, transportation on a rigidspine board, and log rolling of the patient. The rationalefor immobilization seems unambiguous, and it is widelyviewed as a standard of care in patients at risk for spineinjury. Nonetheless, immobilization is not a benign inter-vention as it may be complicated by pain, pressure sores,and impaired chest wall mobility in up to 70% of patients;in addition, neck immobilization is thought to increase the

risk of airway compromise, difficult intubation, aspiration,and increased intracranial pressure.44,45

SurgeryGoals of surgical management in patients with trau-

matic spine and SCI are 1) to decompress neural tissue and2) to prevent cord injury by ensuring mechanical stabilityof the spine. Options include bed rest in traction, externalimmobilization, and open reduction with internal fixation.Many clinicians have abandoned traction because of thecomplications associated with prolonged bed rest.

Numerous aspects of surgical management of SCI aredebated.46 Two recent systematic reviews concluded that1) data are insufficient to generate specific treatment stan-dards or guidelines for surgical management, 2) in patientswith cord compression and stable neurologic findings,early (ie, <24 hours after injury) surgical intervention hasnot been shown to improve or worsen neurologic out-comes, when compared with delayed decompression, and3) although compelling data are lacking, urgent decom-pression is indicated in patients with irreducible bilateralfacet dislocation and incomplete tetraplegia, and in thepresence of a rapidly deteriorating neurologic deficits.47,48

AirwayIndications for endotracheal intubation in patients with

SCI are listed in Table 3. Intubation in the setting of anunstable spine has been linked to severe cord injury anddeath.49,50 Conversely, the imperative of maintainingspine immobilization can augment the difficulty of airwaymanagement. In trauma patients, intubating conditionsmay be aggravated by concomitant injuries of the face and

TABLE 3. Suggested indications for endotracheal intubationin spine and spinal cord injury

Airway compromiseComaEdemaRetropharyngeal hematomaElevated aspiration risk

Respiratory failureSignificant decline in FVC or FVC <15 mL/kgIncreased work of breathingPaO2 <60 mm Hg or significant decline despite supplemental

oxygenPaCO2 >60 mm Hg

Associated traumatic brain injuryGCS <8Intracranial hypertensionHerniation

FVC � forced vital capacity; GCS � Glasgow Coma Scale.



neck and obstruction of the airway by blood, vomitus,foreign bodies, edema, or retropharyngeal hematoma.

Studies in individuals without spine injury, in intactcadavers, and in cadavers with cervical spine injury indi-cate that varying degrees of displacement of cervical spineelements occur with nearly all airway interventions, in-cluding chin lift, jaw thrust, cricoid pressure, mask venti-lation, laryngeal mask airway placement, combi-tubeplacement, laryngoscopy, and orotracheal tube place-ment.51–54 However, it is unclear to what degree these datamay be extrapolated to patients with spine injury. Postu-lated mechanisms of peri-intubation cord injury includemechanical compression, vascular compromise, and ob-struction of cerebrospinal fluid; moreover, intubation isfrequently followed by events that may independently pre-cipitate or potentiate SCI, including surgical manipulation,changes in position, hypotension, and hypoxia. McLeodand Calder propose five clinical criteria that would sug-gest a causal link between laryngoscopy and neurologicinjury: 1) presence of myelopathy on recovery from se-dating agents or anesthesia, 2) a short interval betweenlaryngoscopy and return to consciousness, 3) autonomicinstability following laryngoscopy, 4) difficult laryngos-copy, and 5) disease of the craniocervical junction or in-stability below C3. Although intuitively satisfying, the va-lidity of these criteria is unknown.50

Airway management strategies that address the problemof spine instability include blind nasotracheal intubation,direct laryngoscopy with manual in line stabilization, andawake fiberoptic intubation. Manual in line stabilizationrequires a second operator to hold the head in a neutralposition while the trachea is cannulated; it has been shownto limit, but not abolish, cervical spine movement and isthought to be safer than axial traction.51 Fiberoptic intu-bation is an attractive option because immobilizing de-vices are not removed and displacement of the spine isreduced to a minimum; however, unwanted neck move-ment may occur during topical local anesthetic applicationor if topicalization of the airway is insufficient. Whetherspecific airway techniques impact differentially on neuro-logic risk is unknown. Guidelines for airway managementare given in Table 4.

Respiratory ManagementAbnormal respiratory function and pulmonary compli-

cations are a prominent concern in patients with cervicalSCI. Respiratory failure is an independent predictor of3-month mortality,6,7 and pulmonary complications arethe leading causes of death and morbidity in patients withSCI.11–13

Pathophysiologic AlterationsNeurogenic respiratory failure can result from lesions in

the pons, medulla, cord, peripheral nerves, and neuromus-cular junction. Breathing at rest is active in inspiration andpassive in expiration. The diaphragm (C3–C5) and inter-costals (T1–T11) are the principal inspiratory muscles,while the sternocleidomastoid, trapezius (both cranialnerve XI), and scalene (C3–C8) are accessory inspiratorymuscles. Active expiration is characteristic of respiratorydistress and is provided by the muscles of the anteriorabdominal wall: rectus (T8–L2), obliques (T7–L2), andtransversus (T7–L2). Effective coughing and clearance ofsecretions are also dependent on expiratory muscle func-tion.

Cervical cord injury is associated with significant alter-ations in ventilatory control, breathing patterns, respirato-ry mechanics, and bronchial reactivity.55 Respiratorymuscle weakness leads to alveolar hypoventilation andhypercapnic hypoxemic respiratory insufficiency. Recruit-ment of accessory muscles with inspiration results in ex-pansion of the upper rib cage while the flaccid diaphragmascends into the thoracic cavity, generating a paradoxicalinward movement of the lower rib cage and abdominalwall. All lung volumes except the residual volume aresignificantly reduced, producing a clinical pattern of rapidshallow breathing and a restrictive defect on pulmonaryfunction testing.56 Pulmonary and chest wall complianceis decreased, contributing to a greater work of breathing.57

Severity of ventilatory dysfunction after SCI correlateswith the level and the completeness of the lesion. In aprospective observational study, the mean duration of me-chanical ventilation was 65 days for patients with C1–C4lesions, 22 days in patients with C5–C8 injury, and 12days for patients with thoracic injury.58

TABLE 4. Suggested guidelines for airway management inpatients with spine or spinal cord injury

1. Goal is to ensure rapid control of the airway while incurringminimal neurologic risk.

2. Airway management plan should take into account patientcharacteristics and operator skill and experience.

3. Direct laryngoscopy and endotracheal intubation with MILS ispreferred when the airway must be secured emergently and/or inan uncooperative patient.

4. Awake fiberoptic-guided intubation via the nasal or oral route ispreferred when intubation is not urgent and patient cooperation isanticipated.

5. Hypotension should be avoided by careful titration of sedatingagents and use of vasopressors and fluids if needed.

6. Succinylcholine is contraindicated if SCI occurred >24 hoursbefore administration

MILS � manual in-line stabilization; SCI � spinal cord injury.

Stevens et al220


Complete injury above the C3 level entails apneic re-spiratory arrest and death unless immediate ventilatoryassistance is provided. Survivors have the highest mortal-ity when compared with other SCI injury levels, are atsignificant risk for pulmonary complications, and needlong-term ventilatory support or diaphragmatic pacing.

Injury at C3–C5 is associated with variable degrees ofrespiratory failure. Acutely these patients have markedlyreduced lung volumes requiring ventilatory support, but asignificant proportion are liberated from mechanical ven-tilation weeks to months after injury.59 Ventilatory func-tion has a characteristic time course, worsening during thefirst 2 to 5 days after injury and then progressively im-proving, although never returning to baseline. Recovery offunction reflects functional descent of the neurologic in-jury level with attenuation of cord edema, recruitment ofaccessory respiratory muscles, and retraining of decondi-tioned muscles. Progressive spasticity of chest wallmuscles may assist with weaning by increasing thoracicstability with better transduction of residual muscle forceinto lung volume.60

Injury below C5 is associated with lesser degrees ofventilatory impairment, but patients remain at risk for pul-monary complications. Effective coughing and the abilityto clear secretions are limited by paralysis of expiratorymuscles, increasing risk of atelectasis, and pneumonia. Inthis group and in thoracic spine or cord injury, respiratoryfailure commonly is the result of direct chest trauma(pneumothorax, hemothorax, flail chest, and pulmonarycontusions).

Atelectasis, pneumonia, aspiration pneumonitis, pulmo-nary edema, pleural effusions, and pulmonary embolismoccur in more than two thirds of patients with cervical andupper thoracic SCI. Pulmonary complications are pro-moted by, and may exacerbate, underlying neurogenicventilatory insufficiency, generating a spiral of deteriorat-ing respiratory function. Pulmonary edema may reflectmyocardial injury, excessive fluid administration, neuro-genic pulmonary edema, or acute respiratory distress syn-drome.

ManagementMechanical ventilation is beneficial in selected patients

with SCI by reversing hypoxemia and hypercapnia, de-creasing the work of breathing, reexpanding collapsedlung units, and facilitating pulmonary toilet. Rapid correc-tion of hypoxemia and respiratory acidosis may be criticalin limiting the extent of secondary cord injury.18 Althoughof unknown benefit in this patient population, potentially

useful interventions include noninvasive ventilation, ag-gressive pulmonary hygiene (suctioning, chest percussion,positional changes, deep breathing, incentive spirometry),and bronchodilator therapy.55

There are few data correlating a specific ventilator man-agement strategy with respiratory outcomes and success-ful weaning in SCI. Patients with complete cord injuries atthe C3 level or above require full ventilator support modessuch as controlled mandatory ventilation or assist-control.With incomplete injury and injury below C3, modes thatincorporate spontaneous breathing (eg, pressure supportventilation, synchronized intermittent mandatory ventila-tion) may limit respiratory muscle deconditioning and at-rophy and promote weaning. Many clinicians advocatelarger tidal volumes (10–15 mL per kg), intermittent sighs,or positive end-expiratory pressure to enhance recruitmentof atelectatic lung units. However, the use of larger tidalvolumes may be injurious, particularly in the setting ofacute respiratory distress syndrome.61 Early tracheostomyshould be considered when the need for ventilatory sup-port is anticipated to exceed 2 to 3 weeks. Tracheostomyhas been associated with enhanced subjective tolerance,decreased dead space ventilation, reduced airway resis-tance, and perhaps shorter ventilator weaning when com-pared with orotracheal intubation.62

Cardiovascular ManagementNeurogenic Shock

Cardiovascular instability is a frequent complication ofSCI, especially when the upper thoracic or cervical cord isinvolved. Sympathetic denervation results in arteriolar di-lation and pooling of blood in the venous compartment,while interruption of cardiac sympathetic innervation (T1–T4) promotes bradycardia and reduces myocardial con-tractility. Neurogenic shock is suggested by a pattern ofdecreased heart rate, blood pressure, and systemic vascu-lar resistance. Impaired systolic function may become ap-parent as congestive heart failure, particularly if fluids areadministered injudiciously. Experimental data indicatethat hypotension and shock are particularly deleterious tothe injured spinal cord, contributing to cord hypoperfusionand perpetuating secondary injury.18 Animal models alsoshow that cardiovascular depression may be preceded bya transient phase of severe hypertension thought to reflecta massive, simultaneous discharge of sympathetic neu-rons. This hyperadrenergic response may be a factor in thesubendocardial myocardial injury and neurogenic pulmo-nary edema, which have been observed after SCI andother forms of CNS injury.63



Hemodynamic SupportHemodynamic management in SCI is inferred from the

data on traumatic brain injury, which states that bloodpressure should be targeted to cerebral perfusion pressuregoals. In SCI, however, there is no clinically usefulmethod of assessing spinal cord perfusion pressure, and arelationship between hemodynamic management and neu-rologic outcome has yet to be demonstrated in clinicaltrials. Five retrospective series64–68 and one prospectiveuncontrolled study69 suggest that hemodynamic supportwith volume expansion and/or blood pressure augmenta-tion can be conducted safely and may be associated withbetter neurologic results after SCI. However, it is unclearhow these results are to be interpreted given the lack ofappropriate comparative groups and the presence of con-founding variables, including monitoring and manage-ment in an ICU. A reasonable strategy is one in which 1)hypotension is rigorously avoided and 2) cord perfusion isoptimized with volume expansion and blood pressure aug-mentation, so long as these interventions do not impactunfavorably on other organ systems.70

The clinician managing traumatic SCI should be atten-tive to all potential causes of hemodynamic instability,including neurogenic shock, bleeding, tension pneumo-thorax, myocardial injury, pericardial tamponade, and sep-sis. When the primary hemodynamic disturbance isdeemed to be neurogenic, a stepwise approach is useful,beginning with volume resuscitation and progressing topharmacologic support, ideally guided by invasive hemo-dynamic monitoring. To avoid congestive heart failure,fluids should be titrated with particular attention to incre-mental changes in cardiac output and filling pressures. Inthe presence of decreased systemic vascular resistance andadequate cardiac output and heart rate, a vasopressor suchas phenylephrine or norepinephrine may be used. If car-diac output and/or heart rate are decreased, an agent withinotropic properties such as dopamine is likely to be moreuseful.

Venous ThromboembolismIndividuals with acute SCI have the highest risk of ve-

nous thromboembolic disease among all hospital admis-sions.71,72 The incidence of deep venous thrombosis(DVT) in patients with SCI not receiving prophylaxis is39% to 100%, compared with 9% to 32% in untreatedmedical and surgical ICU patients.72 In a large prospectiveevaluation of trauma patients, SCI was the strongest inde-pendent predictor of DVT, with an odds ratio of 8.59.73 Arecent meta-analysis indicated that patients with SCI havea more than threefold risk of sustaining a DVT compared

with trauma patients without cord injury.74 Among pa-tients with SCI, DVT risk is higher with complete versusincomplete lesions, with thoracic versus cervical level in-jury, and during the first 3 months after injury.72 Theincidence of pulmonary embolism after SCI and in theabsence of prophylaxis is less well documented but isestimated at 4% to 10%.72,75 Pulmonary embolism is oneof the three most common causes of death after SCI.11,12

There is as yet no well-powered, adequately controlledrandomized trial of DVT prophylaxis in SCI. Small-scaleinvestigations have focused on a variety of regimens in-cluding low-dose unfractionated heparin, adjusted doseunfractionated heparin, low molecular weight heparin,warfarin, aspirin/dipyridamole, pneumatic compressiondevices, graduated compression stockings, electricalstimulation, rotating beds, and vena cava filters. Thesestudies are heterogeneous in design, type of prophylaxis,timing of intervention, and tests used to diagnose DVT.Recently, much of these data were analyzed in two sys-tematic reviews71,74 and in two evidence-based consensusconferences72,76 Table 6 which recommend that patientswith SCI receive prophylaxis with either low molecularweight heparin or low-dose unfractionated heparin com-bined with a nonpharmacologic device (pneumatic com-pression, elastic stockings). Although there is agreementthat prophylaxis is necessary, uncertainty persists as towhich is the best prophylactic regimen, whether low mo-lecular weight heparin combined with nonpharmacologicmeans is superior to either alone, how soon after SCI itmay be instituted, how long it should be continued, andthe utility of vena cava filters in this high-risk population.

Pharmacologic Cord ProtectionElucidation of the cellular and molecular underpinnings

of SCI has spurred investigation of a number of pharma-cologic strategies targeted at secondary injury mecha-nisms. Agents that have been tested in human clinicaltrials include corticosteroids, tirilazad mesylate, naloxone,and GM-1 ganglioside.77

Animal models indicate that corticosteroids attenuateinflammatory changes, edema, lipid peroxidation, excito-toxicity, and cytoskeletal degradation associated withSCI. Based on results of several randomized controlledtrials78,79,117 and follow-up studies,80,81,118 methylpred-nisolone therapy has been introduced into widespreadclinical practice and is regarded as a standard of care in themanagement of traumatic acute SCI. However, close criti-cal scrutiny has revealed flaws in the design, reporting,and interpretation of these trials,82–84 and a more recentrandomized trial indicated no difference in neurologic out-

Stevens et al222


come between methylprednisolone- and placebo-treatedpatients85 (Table 5). In a consensus conference, theAmerican Association of Neurologic Surgeons and theCongress of Neurologic Surgeons stated: “Treatment withmethylprednisolone for either 24 or 48 hours is recom-mended as an option in the treatment of patients with acutespinal cord injuries that should be undertaken only withthe knowledge that the evidence suggesting harmful sideeffects is more consistent than any suggestion of clinicalbenefit.”84

GM-1 ganglioside, a glycolipid found in abundance in

the outer leaflet of neuronal and glial plasma membranes,has both neuroprotective and neuroregenerative proper-ties77. Its role in the treatment of acute SCI has beeninvestigated in two randomized controlled trials.86,87 Inthe first of these, significantly better motor scores andFrankel scales were observed at 1 year in patients receiv-ing GM-1 ganglioside compared with placebo.86 In a fol-low-up multicentric trial, neurologic function as measuredby functional and motor scores at 26 weeks was not sig-nificantly different between groups GM-1 ganglioside andplacebo groups.87 Interestingly, patients in the control

TABLE 5. Randomized controlled trials of corticosteroid therapy in acute spinal cord injury

Trial N Group Follow-up Efficacy Comment

NASCIS I(1984)117,118

330 1. MPSS 100 mg/d for 10 d2. MPSS 1000 mg/d for 10 d

6 wk, 6 mo, 12 mo No difference inneurologic outcomebetween groups

Significantly more woundinfections in higherdose group

No control group

NASCIS II(1990)78,80

487 1. MPSS 30 mg/kg bolus then5.4 mg � kg−1 � h−1

for 23 h2. Naloxone 5.4 mg/kg bolus, then

4.5 mg � kg−1 � h−1 for 23 h3. Placebo

6 wk, 6 mo, 12 mo All patients combined, nodifference inneurologic outcomebetween groups

MPSS given �8 afterinjury associated withstatistically significantimprovement in motorscores at 6 month and1 yr

MPSS given >8 h afterinjury associated withworse motor scoresthan placebo

Positive effect basedon a post hocanalysis of a

subgroup of patientsUnclear clinical

significance ofmarginal motorscore changes

Questionable statisticalmethodology

NASCIS III(1997)79,81

499 1. MPSS 30 mg/kg bolus then5.4 mg � kg−1 � h−1 for 24 h

2. MPSs 30 mg/kg bolus then5.4 mg � kg−1 � h−1 for 48 h

3. MPSS 30 mg/kg bolus thentirilazad 2.5 mg/kgevery 6 h for 48 h

6 wk, 6 mo, 12 mo No significant differencein neurologic outcomebetween groups at anytime point

48 h MPSS initiated 3–8h after injuryassociated withstatistically significantimprovement in motorscores at 6 weeks and6 months, cf 24 hMPSS

Significantly higher rateof sepsis andpneumonia in 48 hMPSS group

No placebo groupPositive effect based

on a post hocanalysis ofsubgroups

Equivalent FIM in allgroups

Questionable statisticalmethodology

Pointillart et al.(2000)85

106 1. MPSS 30 mg/kg bolus then5.4 mg � kg−1 � h−1 for 23 h

2. Nimodipine 0.03 mg/kg then0.015 mg � kg−1 � h−1 for 7 d

3. MPSS and nimodipine4. No medication

12 mo No significant differencein neurologic outcomebetween groups

Trend toward greaterincidence of infectiouscomplications inpatients receivingMPSS

Inadequate statisticalpower

NASCIS � National Acute Spinal Cord Injury Studies; MPSS � methylpresnisolone; FIM � Functional Independence Measure scale.



group, all of whom received methylprednisolone in dosesequivalent to NASCIS II, did not reproduce the modestneurologic gain over time that was seen in the priorstudy.79

Other Critical Care Issues (Table 7)Infectious complications are a leading cause of death

and morbidity after SCI, most frequently involving thelungs and urinary tract.11,12 Among trauma patients, thepresence of SCI is a strong independent predictor of in-fection.88 There is evidence that individuals with SCI havedysfunctional immune responses, possibly in proportion tothe severity of the neurologic deficit.89 Factors that mayincrease the risk of pneumonia are aspiration, prolongedtracheal cannulation, reduced ability to clear secretions,and atelectasis. Urinary tract infections are promoted byincontinence, elevated intravesical pressure, reflux,stones, and neurologic obstruction. High doses of cortico-steroids are associated with an increased rate of infectiouscomplications.78,79,85 When an infectious source is not ap-parent, consideration should be given to acute abdominalprocesses that may be clinically occult in SCI, namely,pancreatitis, cholecystitis, diverticulitis, bowel perfora-tion, or bowel ischemia.

Acute SCI increases the risk of stress ulceration andupper gastrointestinal bleeding, in particular, when asso-ciated with mechanical ventilation and high-dose cortico-steroids. There is no specific data on stress ulcer prophy-laxis in SCI; however, it is suggested that all patients withacute SCI receive an H2 antagonist or sucralfate, as theefficacy of these agents has been demonstrated in othercritically ill patient populations.105 Acute SCI may also beassociated with gastric distension and ileus, contributing

to the risk of aspiration and further compromising respi-ratory function; nasogastric or orogastric suctioning andthe use of prokinetic agents such as metoclopramide orerythromycin may be useful in this setting.

Psychiatric complications are common after SCI andinclude depression (30%–40% of individuals), anxietydisorders (20%–25%), and substance-related disorders(40%–50%).91–93 Risk of suicide is increased two to sixtimes when compared with the general population.93 In theintensive care unit, every effort should be made to pro-mote appropriate relief from pain and anxiety. The poten-tial role of other interventions such as psychiatric coun-seling in the acute setting has not been studied.

PERIOPERATIVE MANAGEMENT

Anesthesia for acute SCI recapitulates many of the con-cerns relevant to the critical care setting (Table 7). Preop-

TABLE 6. Antithrombotic prophylaxis in SCI

Possibly effective prophylaxisLow molecular weight heparinLow-dose unfractionated heparin combined with

nonpharmacologic devices*Adjusted dose unfractionated heparin

Ineffective prophylaxisLow-dose unfractionated heparinWarfarin

Interventions of unknown efficacyPneumatic compression devicesElastic stockings aloneVena cava filtersLow molecular weight heparin combined with nonpharmacologic

devices

*Nonpharmacologic devices include pneumatic compression devices,elastic stockings, and electrical stimulation.

Based on references 71, 72, and 76.

TABLE 7. Critical care issues in spinal cord injury

System Problem Management

Neurologic Secondary injury ImmobilizationSurgical

decompression?Adequate perfusion

and oxygenationCorticosteroids?

Cardiovascular Neurogenic shock Invasive monitoringVolume resuscitationVasopressor agentsInotropic agents

Autonomicdysreflexia

Removal of stimulusVasodilators

Hemostasis Deep venousthrombosis

Pulmonary embolism

LMWH prophylaxisTherapeutic heparinVena cava filter?

Respiratory Ventilatory failure Endotrachealintubation

Mechanical ventilationTracheostomy

Pneumonia Antimicrobial therapyAtelectasis Incentive spirometry,

PEEPWeaning

Gastrointestinal Stress ulcer H2-blockerprophylaxis

IleusOccult peritonitis Surgery,

antimicrobialsUrinary Urinary tract infection AntimicrobialsSkin Decubitus ulcers Prevention protocols

Wound carePlastic and recon-

structive surgeryPsychiatric Anxiety Sedation, pain control

Depression Counseling?Suicide

LMWH � low molecular weight heparin; PPI � proton-pump in-hibitor.

Stevens et al224


erative assessment should be attentive to associated inju-ries, airway characteristics, oxygenation, ventilatorycapacity, hemodynamic variables, and neurologic exami-nation. An anesthetic plan should be devised with particu-lar attention to the technique for securing the airway, an-ticipated need for hemodynamic support, likelihood ofextubation, and indication for postprocedure ICU admis-sion. Patients with acute SCI undergoing surgery will al-most universally require general anesthesia. In selectedpatients with chronic SCI undergoing surgery below thelevel of the lesion, some authors have advocated alterna-tive strategies including monitored anesthesia care, spinal,or epidural anesthesia.94,95 In this patient group, benefitsof regional anesthesia need to be carefully weighedagainst risks, which include difficulty in assessing thelevel of anesthetic block, hypotension, and the potentialfor neurotoxicity in the setting of injured neural tissue.96

PositioningPositioning of the patient to achieve adequate exposure

and reduction of the spine must be accomplished withclose attention to complications, which include worseningof neurologic injury; abdominal compression with venousengorgement and increased intraoperative bleeding; im-paired chest wall compliance; ischemic optic neuropathy;and peripheral nerve injury. In patients with SCI withimpaired sympathetic outflow to the vasculature or heart,changes in position may have significant hemodynamicconsequences; sudden adoption of the head-up positionmay lead to severe hypotension through venous pooling ofblood, whereas the head-down position may precipitatecardiac failure.

Electrophysiologic MonitoringDuring spine surgery, distraction, instrumentation, or

direct compression of cord vascular structures may sig-nificantly affect SCBF and promote cord infarction if un-corrected.97,98 Spine surgery conducted in the prone posi-tion increases intraabdominal and vertebral venouspressure, possibly promoting venous cord infarction.99 So-matosensory evoked potentials (SSEPs) and motor evokedpotentials (MEPs) may assist in the detection of intraop-erative spinal cord dysfunction, prompting corrective in-terventions before irreversible neural damage has oc-curred. While numerous reports attest to the helpfulness ofthese techniques, their impact on postoperative outcomeshas not been tested in a prospective randomized fash-ion.100,101 SSEPs assess the ascending sensory tracts lo-cated in the posterior columns, whereas MEPs evaluatethe descending anterior and posterolateral corticospinal

tracts.100,101 During surgery of the spine, SSEP identifiesnew-onset cord dysfunction with a sensitivity of >95%,but the specificity of the monitor is disappointing, with ahigh false-positive rate reflecting multiple confoundingvariables including anesthetic agents, hypothermia, or hy-potension.102 Limited evidence indicates that MEPs maybe as sensitive as SSEPs during spine surgery; however,the clinical validity of MEPs, alone or in combination withSSEPs, needs further clarification.101

Hemodynamic ManagementHypotension is associated with neurologic deterioration

in patients with SCI and must be circumvented by judi-cious administration of anesthetic agents, fluid resuscita-tion, and vasopressor and/or inotropic support whenneeded. Placement of a central venous catheter and anarterial line may assist with intravascular volume andblood pressure management. Pulmonary artery catheter-ization or transesophageal echocardiography should beconsidered in the presence of severe hemodynamic com-promise. Using pulmonary artery catheters, Mackenzie etal. demonstrated diminished cardiac output responses tofluid loading in a subset of acutely tetraplegic patientsundergoing surgery, indicating impaired left ventricularfunction.103

Autonomic DysreflexiaAutonomic dysreflexia is a cluster of symptoms char-

acterized by paroxysmal hypertension, headache, and bra-dycardia in response to a stimulus originating below thelevel of the SCI.104 It appears in up to 90% of patientswith lesions above T6, and its severity is greater the higherthe level of SCI. Onset of symptoms may occur weeks toyears after the initial injury. Common precipitating eventsinclude distension of hollow viscera (bladder, bowel,uterus, gallbladder), cutaneous stimulation, and surgicalprocedures, frequently involving pelvic organs or thelower extremities.94,104 Below the level of the lesion,clinical signs include cutaneous vasoconstriction, pilo-erection, and bladder spasm, while above the lesion levelpatients present with cutaneous flushing, sweating, nasaland conjunctival congestion, blurred vision, and nausea.Untreated, autonomic dysreflexia can lead to encephalop-athy, stroke, seizures, intracerebral or retinal hemorrhage,myocardial infarction, congestive heart failure, arrhyth-mias, and death. Suggested pathophysiological mecha-nisms include loss of supraspinal inhibitory control of seg-mental sympathetic neurons, adrenergic receptor up-regulation, and abnormal synaptic connections due topostinjury sprouting.105 Management priority is removal



of the precipitating stimulus. If the crisis occurs intraop-eratively, deepening the level of anesthesia may be help-ful. Hypertension should be controlled with vasodilatingagents such as nitroprusside, hydralazine, phentolamine,or nicardipine.

Anesthetic AgentsAnesthetic agents exert profound effects both on neu-

ronal function and blood flow in the spinal cord. Mini-mum alveolar concentration of volatile anesthetics is un-affected by decerebration or cervical cord transsection,suggesting that unresponsiveness to noxious stimuli dur-ing general anesthesia is mediated by anesthetic effects atthe cord level.106,107 SCBF and cerebral blood flow aresimilarly influenced by anesthetic agents, increasing withpotent inhalational agents, decreasing with thiopental.SCBF increases observed with isoflurane are concomitantwith decreases in spinal cord neuronal activity, indicatingflow-metabolism uncoupling as is observed in thebrain.108 The clinical significance of these results is un-known, although a neuroprotective effect on the spinalcord has been suggested for thiopental.109

There are no data to support the use of one anestheticagent over another in patients with SCI. Succinylcholine iscontraindicated because of the potential for lethal hyper-kalemic responses related to potassium efflux from extra-junctional acetylcholine receptors.110 The risk of this com-plication is apparent within 24 hours after SCI and may berelevant for up to 18 months.

REGENERATION

Treatment of acute SCI is currently centered on miti-gating secondary injury by the institution of a limitednumber of physiologic interventions whose ability tomodify neurologic outcomes is unproven. As with strokeand traumatic brain injury, clinical trials of neuroprotec-tive agents have consistently failed to demonstrate a fa-vorable effect on neurologic function after SCI.

In recent years a growing body of work has challengedthe classic paradigm that neuronal repair and cellular re-newal cannot occur in the central nervous system. In theinjured spinal cord, for instance, axonal regrowth may becoaxed by selectively targeting inhibitory signals ex-pressed on myelin sheaths and in the glial scar.111 Inter-ventions for SCI that are being investigated in animalmodels include 1) promoting axonal regrowth with neu-rotrophic factors (BDNF, GDNF, NT3, nerve growth fac-tor)112; 2) blocking myelin-associated inhibition (anti-Nogo antibody, anti-MAG antibody, Nogo peptide)113; 3)

removing the inhibitory extracellular matrix componentsof scar tissue (chondroitinase ABC)114; 4) bridging spinalcord lesions with cellular scaffolds that support axonalgrowth (olfactory-ensheathing cells, Schwann cells)115;and 5) transplantation of precursor cells to reconstitute losttissue.116 Therapeutic implementation of such strategies inhumans may fulfill the goal of restoring function in thispredominantly young patient population.

Acknowledgments: The authors thank the nurses and staff ofthe Neurosciences Critical Care Unit of Johns Hopkins Univer-sity School of Medicine for providing excellent care to this chal-lenging group of patients.

REFERENCES

1. Ghezzi A, Baldini SM, Zaffaroni M. Differential diagnosis of acutemyelopathies. Neurol Sci. 2001;22(suppl 2):60–4.

2. Lintott P, Hafez HM, Stansby G. Spinal cord complications ofthoracoabdominal aneurysm surgery. Br J Surg. 1998;85:5–15.

3. Cheung AT, Weiss SJ, McGarvey ML, et al. Interventions forreversing delayed-onset postoperative paraplegia after thoracic aor-tic reconstruction. Ann Thorac Surg. 2002;74:413–9.

4. Sekhon LHS, Fehlings MG. Epidemiology, demographics, andpathophysiology of acute spinal cord injury. Spine. 2001;26(suppl24):2–12.

5. Surkin J, Gilbert BJ, Harkey HL, et al. Spinal cord injury in Mis-sissippi: findings and evaluation, 1992–1994. Spine. 2000;25:716–21.

6. Claxton AR, Wong DT, Chung F, et al. Predictors of hospitalmortality and mechanical ventilation in patients with cervical spi-nal cord injury. Can J Anaesth. 1998;45:144–9.

7. Daverat P, Gagnon M, Dartigues JF, et al. Initial factors predictingsurvival in patients with a spinal cord injury. J Neurol NeurosurgPsychiatry. 1989;52:403–6.

8. Marino RJ, Ditunno JF Jr, Donovan WH, et al. Neurologic recov-ery after traumatic spinal cord injury: data from the model spinalcord injury systems. Arch Phys Med Rehabil. 1999;80:1391–6.

9. Waters RL, Adkins RH, Yakura JS, et al. Motor and sensory re-covery following incomplete tetraplegia. Arch Phys Med Rehabil.1994;75:306–11.

10. Katoh S, el Masry WS. Motor recovery of patients presenting withmotor paralysis and sensory sparing following cervical spinal cordinjuries. Paraplegia. 1995;33:506–9.

11. DeVivo MJ, Krause JS, Lammertse DP. Recent trends in mortalityand causes of death among persons with spinal cord injury. ArchPhys Med Rehabil. 1999;80:1411–9.

12. Yeo JD, Walsh J, Rutkowski S, et al. Mortality following spinalcord injury. Spinal Cord. 1998;36:329–36.

13. Frankel HL, Coll JR, Charlifue SW, et al. Long term survival inspinal cord injury: a fifty year investigation. Spinal Cord. 1998;36:266–74.

14. Frankel HL, Coll JR, Charlifue SW, et al. Long term survival inspinal cord injury: a fifty year investigation. Spinal Cord. 1998;36:266–74.

15. Schoenen J. Clinical anatomy of the spinal cord. Neurol Clin. 1991;3:503–32.

16. Hickey R, Albin MS, Bunegin L, et al. Autoregulation of spinalcord blood flow: is the cord a microcosm of the brain? Stroke.1986;17:1183–9.

17. Guha A, Tator CH, Rochon J. Spinal cord blood flow and systemicblood pressure after experimental spinal cord injury in rats. Stroke.

Stevens et al226


1989;372–7.18. Dumont RJ, Okonkwo DO. Verma S, et al. Acute Spinal Cord

Injury. Part I: Pathophysiologic Mechanisms. Clinical Neurophar-macology. 2001;24:254–64.

19. Tator CH, Koyanagi I. Vascular mechanism in pathophysiology ofhuman spinal cord injury. J Neurosurg. 1997;86:483–92.

20. Lu J, Ashwell KW, Waite P. Advances in secondary spinal cordinjury: role of apoptosis. Spine. 2000;25:1859–66.

21. Edgar R, Quail P. Progressive post-traumatic cystic and non-cysticmyelopathy. Br J Neurosurg. 1994;8:7–22.

22. Holly LT, Kelly DF, Counelis GJ, et al. Cervical spine traumaassociated with moderate and severe head injury: incidence, riskfactors, and injury characteristics. J Neurosurg. 2002;96(suppl3):285–91.

23. Michael DB, Guyot DR, Darmody WR. Coincidence of head andcervical spine injury. J Neurotrauma. 1989;6:177–89.

24. International standards for neurologic classification of spinal cordinjury. Chicago: American Spinal Cord Injury Association, 2000.

25. Demetriades D, Charalambides K, Chahwan S, et al. Nonskeletalcervical spine injuries: epidemiology and diagnostic pitfalls. JTrauma. 2000;48:724–7.

26. Gupta SK, Rajeev K, Khosla VK, et al. Spinal cord injury withoutradiographic abnormality in adults. Spinal Cord. 1999;37:726–9.

27. Hendey GW, Wolfson AB, Mower WR, et al. Spinal cord injurywithout radiographic abnormality: results of the National Emer-gency X-Radiography Utilization Study in blunt cervical trauma. JTrauma. 2002;53:1–4.

28. Schenarts PJ, Diaz J, Kaiser C, et al. Prospective comparison ofadmission computed tomographic scan and plain films of the uppercervical spine in trauma patients with altered mental status. JTrauma. 2001;51:663–8.

29. Berne JD, Velmahos GC, El-Tawil Q, et al. Value of completecervical helical computed tomographic scanning in identifying cer-vical spine injury in the unevaluable blunt trauma patient withmultiple injuries: a prospective study. J Trauma. 1999;47:896–902.

30. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Radiographic assess-ment of the cervical spine in symptomatic trauma patients. Neuro-surgery. 2002;50(suppl):36–43.

31. Borock EC, Gabram SG, Jacobs LM, et al. A prospective analysisof a two-year experience using computed tomography as an adjunctfor cervical spine clearance. J Trauma. 1991;31:1001–6.

32. Brohi K, Wilson-Macdonald J. Evaluation of unstable cervicalspine injury: a 6-year experience. J Trauma. 2000;49:76–80.

33. Klein GR, Vaccaro AR, Albert TJ, et al. Efficacy of magneticresonance imaging in the evaluation of posterior cervical spinefractures. Spine. 1999;24:771–4.

34. Hoffman JR, Mower WR, Wolfson AB, et al. Validity of a set ofclinical criteria to rule out injury to the cervical spine in patientswith blunt trauma: National Emergency X-Radiography Utiliza-tion. Study Group. N Engl J Med. 2000;343:94–9.

35. Gonzales RP, Fried PO, Bukhalo M, et al. Role of clinical exami-nation in screening for blunt cervical spine injury. J Am Coll Surg.1999189:152–7.

36. Roth BJ, Martin RR, Foley K, et al. Roentgenographic evaluationof the cervical spine: a selective approach. Arch Surg. 1994;129:643–5.

37. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Radiographic assess-ment of the cervical spine in aymptomatic trauma patients. Neu-rosurgery. 2002;50(suppl):30–5.

38. Lewis LM, Docherty M, Ruoff BE, et al. Flexion-extension viewsin the evaluation of cervical spine injuries. Ann Emerg Med. 1991;20:117–21.

39. DÆAlise MD. Benzel EC, Hart BL: Magnetic resonance imagingevaluation of the cervical spine in the comatose or obtunded trauma

patient. J Neurosurg. 1999;91(suppl 1):54–9.40. Davis JW, Parks SN, Detlefs CL, et al. Clearing the cervical spine

in obtunded patients: the use of dynamic fluoroscopy. J Trauma.1995;39:435–8.

41. White AA, Panjabi MM. The role of stabilization in the treatmentof cervical spine injuries. Spine. 1984;9:512–22.

42. Denis F. The three column spine and its significance in the classi-fication of acute thoracolumbar spinal injuries. Spine. 1983;8:817–31.

43. Panjabi MM, Thibodeau LL, Crisco JJ, et al. What constitutesspinal instability? Clin Neurosurg. 1988;34:313–39.

44. Kwan I, Bunn F, Roberts I. WHO Pre-Hospital Trauma Care Steer-ing Committee. Spinal immobilization for trauma patients (Co-chrane Review). In: The Cochrane Library, Issue 2, 2002.

45. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Cervical spine immo-bilization before admission to the hospital. Neurosurgery.2002;50(suppl):7–17.

46. Amar AP, Levy ML. Surgical controversies in the management ofspinal cord injury. J Am Coll Surg. 1999;188:550–66.

47. Fehlings MG, Sekhon LH, Tator C. The role and timing of decom-pression in acute spinal cord injury: what do we know? Whatshould we do? Spine. 2001;26(suppl 24):101–10

48. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Treatment of subaxialcervical spine injuries. Neurosurgery. 2002;50(suppl):156–65.

49. Muckart DJJ, Bhagwanjee S, van der Merwe R. Spinal cord injuryas a result of endotracheal intubation in patients with undiagnosedcervical spine fractures. Anesthesiology.

50. McLeod ADM, Calder I. Spinal cord injury and direct laryngos-copy—the legend lives on. Br J Anaesth. 2000;705–8.

51. Lennarson PJ, Smith DW, Sawin PD, et al. Cervical spinal motionduring intubation: efficacy of stabilization maneuvers in the settingof complete segmental instability. J Neurosurg. 2001;94(suppl2):265–70.

52. Brimacombe J, Keller C, Kunzel KH, et al. Cervical spine motionduring airway management: a cinefluoroscopic study of the poste-riorly destabilized third cervical vertebrae in human cadavers.Anesth Analg. 2000;91:1274–8.

53. Gerling MC, Davis DP, Hamilton RS, et al. Effects of cervicalspine immobilization technique and laryngoscope blade selectionon an unstable cervical spine in a cadaver model of intubation. AnnEmerg Med. 2000;36:293–300.

54. Lennarson PJ, Smith D, Todd MM, et al. Segmental cervical spinemotion during orotracheal intubation of the intact and injured spinewith and without external stabilization. J Neurosurg. 2000;92(suppl 2):201–6.

55. Slack RS, Shucart W. Respiratory dysfunction associated with trau-matic injury to the central nervous system. Clin Chest Med. 1994;15:739–49.

56. McMichan JC, Michel L, Westbrook PR. Pulmonary dysfunctionfollowing traumatic quadriplegia. JAMA. 1980;243:528.

57. Scanlon PD, Loring SH, Pichurko BM, et al. Respiratory mechan-ics in acute quadriplegia: lung and chest wall compliance and di-mensional changes during respiratory maneuvers. Am Rev RespirDis. 1989;139:615–20.

58. Jackson AB, Groomes TE. Incidence of respiratory complicationsfollowing spinal cord injury. Arch Phys Med Rehabil. 1994;75:270–5.

59. Wicks AB, Menter RR. Long term outlook in quadriplegic patientswith initial ventilator dependency. Chest. 1986;90:406.

60. Haas F, Axen K, Pineda H, et al. Temporal pulmonary functionchanges in cervical cord injury. Arch Phys Med Rehabil. 1985;66:139–44.

61. Acute Respiratory Distress Syndrome Network. Ventilation withlower tidal volumes as compared with traditional tidal volumes for



acute lung injury and the acute respiratory distress syndrome. NEngl J Med. 2000;342:1301–8.

62. Heffner JE. The role of tracheotomy in weaning. Chest. 2001;120(suppl):477–81.

63. Colice GL, Matthay MA, Bass E, et al. Neurogenic pulmonaryedema. Am Rev Respir Dis. 1984;130:941–8.

64. Levi L, Wolf A, Belzberg H. Hemodynamic parameters in patientswith acute cervical cord trauma. Neurosurgery. 1993;33:1007–17.

65. Levi L, Wolf A, Rigamonti D, et al. Anterior decompression incervical spine trauma. Neurosurgery. 1991;29:216–22.

66. Tator CH, Rowed DW, Schwartz MI, et al. Management of acutespinal cord injuries. Can J Surg. 1984;27:289–93.

67. Wolf A, Levi L, Mirvis S, et al. Operative management of bilateralfacet dislocation. J Neurosurg. 1991;75:883–90.

68. Zach GA, Seiler W, Dollfus P. Treatment results of spinal cordinjuries in the Swiss Paraplegic Centre of Basel. Paraplegia. 1976;14:58–65.

69. Vale FL, Burns J, Jackson AB, et al. Combined medical and sur-gical treatment after acute spinal cord injury. J Neurosurg. 1997;87:239–46.

70. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Blood pressure man-agement after acute spinal cord injury. Neurosurgery. 2002;50(suppl):58–62.

71. Attia J, Ray JG, Cook DJ, et al. Deep vein thrombosis and itsprevention in critically ill adults. Arch Intern Med. 2001;161:1268–79.

72. Geerts WH, Heit JA, Clagett GP, et al. Prevention of venous throm-boembolism. Chest. 2001;119(suppl):132–75.

73. Geerts WH, Code KI, Jay RM, et al. A prospective study of venousthromboembolism after major trauma. N Engl J Med. 1994;331:1601–6.

74. Velmahos GC, Kern J, Chan LS, et al. Prevention of venous throm-boembolism after injury: an evidence-based report: II. Analysis ofrisk factors and evaluation of the role of vena caval filters. JTrauma. 2000;49:140–4.

75. Green D. Prevention of thromboembolism after spinal cord injury.Semin Thromb Hemost. 1991;17:347–50.

76. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Deep venous thrombo-sis and thromboembolism in patients with cervical spinal cord in-juries. Neurosurgery. 2002;50(suppl):73–80.

77. Dumont RJ, Verma S, Okonkwo DO, et al. Acute spinal cordinjury: II. Contemporary pharmacotherapy. Clin Neuropharmacol.2001;24:265–79.

78. Bracken MB, Shepard MJ, Collins WF, et al. A randomized, con-trolled trial of methylprednisolone or naloxone in the treatment ofacute spinal-cord injury: results of the second national acute spinalcord injury study. N Engl J Med. 1990;322:1405–11.

79. Bracken MB, Shepard MJ, Holford TR, et al. Administration ofmethylprednisolone for 24 or 48 hours or tirilazad mesylate for 48hours in the treatment of acute spinal cord injury. JAMA. 1997;277:1597–604.

80. Bracken MB, Shepard MJ. Collins, et al. Methylprednisolone ornaloxone treatment after acute spinal cord injury: 1-year follow-updata. Results of the second National Acute Spinal Cord InjuryStudy. J Neurosurg. 1992;76:23–31.

81. Bracken MB, Shepard MJ, Holford TR, et al. Methylprednisoloneor tirilazad mesylate administration after acute spinal cord injury:1-year follow up. Results of the third National Acute Spinal CordInjury randomized controlled trial. J Neurosurg. 1998;89:699–706.

82. Hurlbert RJ. The role of steroids in acute spinal cord injury: anevidence-based analysis. Spine. 2001;26(suppl):39–46.

83. Short DJ, El Masry WS, Jones PW. High dose methylprednisolonein the management of acute spinal cord injury: a systematic reviewfrom a clinical perspective. Spinal Cord. 2000;38:273–86.

84. Guidelines of the American Association of Neurologic Surgeonsand the Congress of Neurologic Surgeons. Pharmacologicaltherapy after cervical spinal cord injury. Neurosurgery. 2002;50(suppl):63–72.

85. Pointillart V, Petitjean ME, Wiart L, et al. Pharmacological therapyof spinal cord injury during the acute phase. Spinal Cord. 2000;38:71–6.

86. Geisler FH, Dorsey FC, Coleman WP. Recovery of motor functionafter spinal cord injury: a randomized, placebo-controlled trial withGM-1 ganglioside. N Engl J Med. 1991;324:1829–38.

87. Geisler FH, Coleman WP, Grieco G, et al. The Sygen Study Group:the GM1 ganglioside multi-center acute spinal cord injury study.Spine. 2001;26(suppl):87–98.

88. Papia G, McLellan BA, El-Helou P, et al. Infection in hospitalizedtrauma patients: incidence, risk factors, and complications. JTrauma. 1998;47:923.

89. Nash MS. Immune dysfunction and illness susceptibility after spi-nal cord injury: an overview of probable causes, likely conse-quences, and potential treatments. J Spinal Cord Med. 2000;23:109–10.

90. Cook D, Guyatt G, Marshall J, et al. A comparison of sucralfateand ranitidine for the prevention of upper gastrointestinal bleedingin patients requiring mechanical ventilation. Canadian Critical CareTrials Group. N Engl J Med. 1998;338:791–7.

91. Elliott TR, Frank RG. Depression following spinal cord injury.Arch Phys Med Rehabil. 1996;77:816–23.

92. Radnitz CL, Tirch D. Substance misuse in individuals with spinalcord injury. It J Addict. 1995;30:1117–40.

93. Krause JS. Spinal cord injury and its rehabilitation. Curr OpinNeurol Neurosurg. 1992;5:669–72.

94. Hambly PR, Martin B. Anaesthesia for chronic spinal cord lesions.Anaesthesia. 1998;53:273–89.

95. Schonwald G, Fish KJ, Perkash I. Cardiovascular complicationsduring anesthesia in chronic spinal cord injured patients. Anesthe-siology. 1981;55:550–8.

96. Hodgson PS, Neal JM, Pollock JE, et al. The neurotoxicity of drugsgiven intrathecally (spinal). Anesth Analg. 1999;88:797–809.

97. MacEwan GD, Bunnell WP, Krishnaswami S. Acute neurologiccomplications in the treatment of scoliosis: a report of the ScoliosisResearch Society. J Bone Joint Surg Am. 1975;57:404–8.

98. Nuwer MR, Dawson EG, Carlson LG, et al. Somatosensory evokedpotential spinal cord monitoring reduces neurologic deficits afterscoliosis surgery: results of a large multicenter survey. Electroen-cephalogr Clin Neurophysiol. 1995;96:6–11.

99. Bhardwaj A, Long DM, Ducker TB, et al. Neurologic deficits aftercervical laminectomy in the prone position. J Neurosurg Anesthe-siol. 2001;13:314–9.

100. Kumar A. Bhattacharya, Makhija N. Evoked potential monitoringin anaesthesia and analgesia. Anaesthesia. 2000;55:225–41.

101. de Haan P, Kalkman CJ. Spinal cord monitoring: somatosensoryand motor evoked potentials. Anesthesiol Clin North Am. 2001;19:923–45.

102. Noonan KJ, Walker T, Feinberg JR, et al. Factors related to false-versus true-positive neuromonitoring changes in adolescent idio-pathic scoliosis surgery. Spine. 2002;27:825–30.

103. Mackenzie CF, Shin B, Krishnaprasad D, et al. Assessment ofcardiac and respiratory function during surgery on patients withacute quadriplegia. J Neurosurg. 1985;62:843–9.

104. Karlsson AK. Autonomic dysreflexia. Spinal Cord. 1999;37:383–91.

105. Teasell RW, Arnold JM, Krassioukov A, et al. Cardiovascular con-sequences of loss of supraspinal control of the sympathetic nervoussystem after spinal cord injury. Arch Phys Med Rehabil. 2000;81:506–16.

106. Rampil IJ, Mason P, Singh H. Anesthetic potency (MAC) is inde-

Stevens et al228


pendent of forebrain structures in the rat. Anesthesiology. 1993;78:707.

107. Rampil IJ. Anesthetic potency is not altered after hypothermic spi-nal cord transection in rats. Anesthesiology. 1994;80:606–11.

108. Hoffman WE, Edelman G, Kochs E, et al. Cerebral autoregulationin awake versus isoflurane-anesthetized rats. Anesth Analg. 1991;73:753.

109. Hitchon P, Lobosky J, Yamada T, et al. The response of spinal cordblood flow to high dose barbiturates. Spine. 1982;7:41–5.

110. Stone WA, Beach TP, Hamelberg W. Succinylcholine: danger inthe spinal-cord-injured patient. Anesthesiology. 1970;32:168–9.

111. Schwab ME. Repairing the injured spinal cord. Science. 2002;295:1029–31.

112. Namiki J, Kojima A, Tator CH. Effect of brain-derived neuro-trophic factor, nerve growth factor, and neurotrophin-3 on func-tional recovery and regeneration after spinal cord injury in adultrats. J Neurotrauma. 2000;17:1219–31.

113. Merkler D, Metz GA, Raineteau O, et al. Locomotor recovery inspinal cord-injured rats treated with an antibody neutralizing the

myelin-associated neurite growth inhibitor Nogo-A. J Neurosci.2001;21:3665–73.

114. Bradbury EJ, Moon LD, Popat RJ, et al. Chondroitinase ABCpromotes functional recovery after spinal cord injury. Nature.2002;416:636–40.

115. Ramón-Cuet A, Cordero MI, Santos-Benito FF, et al. Functionalrecovery of paraplegic rats and motor axon regeneration in theirspinal cords by olfactory ensheathing glia. Neuron 2000;25:425–35.

116. McDonald JW, Liu XZ, Qu Y, et al. Transplanted embryonic stemcells survive, differentiate and promote recovery in injured rat spi-nal cord. Nat Med. 1999;5:1410–2.

117. Bracken MB, Collins WF, Freeman DF, et al. Efficacy of methyl-prednisolone in acute spinal cord injury. JAMA. 1984;251:45–52.

118. Bracken MB, Shepard MJ, Hellenbrand KG, et al. Methylprednis-olone and neurological function 1 year after spinal cord injury:results of the National Acute Spinal Cord Injury Study. J Neuro-surg. 1985;63:704–13.



critical care and perioperative management in …...review article critical care and perioperative...

Documents