Cervical Spine Magnetic Resonance Imaging in Alert ...

TRAUMA/ORIGINAL RESEARCH

Cervical Spine Magnetic Resonance Imaging in Alert,
Neurologically Intact Trauma Patients With Persistent Midline

Tenderness and Negative Computed Tomography Results

Helen M. Ackland, MHSc, Peter A. Cameron, MBBS, MD, Dinesh K. Varma, MBBS, MD, Gregory J. Fitt, MBBS,
D. James Cooper, MD, BS, Rory Wolfe, BSc, PhD, Gregory M. Malham, BSc, MB ChB,

Jeffrey V. Rosenfeld, MBBS, MD, Owen D. Williamson, MBBS, GradDipClinEpi, Susan M. Liew, MBBS

From the National Trauma Research Institute (Ackland, Cameron, Rosenfeld), the Emergency and Trauma Centre (Cameron), the Department of
Radiology (Varma), the Intensive Care Department (Cooper), the Department of Neurosurgery (Malham, Rosenfeld), and the Department of
Orthopaedic Surgery (Liew), The Alfred Hospital, Melbourne, Australia; the Department of Epidemiology and Preventive Medicine (Ackland,

Cameron, Cooper, Wolfe, Williamson) and Department of Surgery (Varma, Malham, Rosenfeld), Monash University, Melbourne, Australia; and the
Department of Radiology, Austin Health, Melbourne, Australia, and the Department of Medicine, University of Melbourne, Melbourne,
Australia (Fitt).

Study objective: We aim to determine the prevalence and factors associated with cervical discoligamentous
injuries detected on magnetic resonance imaging (MRI) in acute, alert, neurologically intact trauma patients with
computed tomography (CT) imaging negative for acute injury and persistent midline cervical spine tenderness.
We present the cross-sectional analysis of baseline information collected as a component of a prospective
observational study.

Methods: Alert, neurologically intact trauma patients presenting to a Level I trauma center with CT negative for
acute injury, who underwent MRI for investigation of persistent midline cervical tenderness, were prospectively
recruited. Deidentified images were assessed, and injuries were identified and graded. Outcome measures
included the presence and extent of MRI-detected injury of the cervical ligaments, intervertebral discs, spinal
cord and associated soft tissues.

Results: There were 178 patients recruited during a 2-year period to January 2009. Of these, 78 patients (44%)
had acute cervical injury detected on MRI. There were 48 single-column injuries, 15 two-column injuries, and 5
three-column injuries. Of the remaining 10 patients, 6 had isolated posterior muscle edema, 2 had alar
ligamentous edema, 1 had epidural hematoma, and 1 had atlanto-occipital edema. The injuries to 38 patients
(21%) were managed clinically; 33 patients were treated in cervical collars for 2 to 12 weeks, and 5 patients
(2.8%) underwent operative management, 1 of whom had delayed instability. Ordinal logistic regression revealed
that factors associated with a higher number of spinal columns injured included advanced CT-detected cervical
spondylosis (odds ratio [OR] 11.6; 95% confidence interval [CI] 3.9 to 34.3), minor isolated thoracolumbar
fractures (OR 5.4; 95% CI 1.5 to 19.7), and multidirectional cervical spine forces (OR 2.5; 95% CI 1.2 to 5.2).

Conclusion: In patients with cervical midline tenderness and negative acute CT findings, we found that a subset
of patients had MRI-detected cervical discoligamentous injuries and that advanced cervical spine degeneration
evident on CT, minor thoracolumbar fracture, and multidirectional cervical spine forces were associated with
increased injury extent. However, a larger study is required to validate which variables may reliably predict
clinically important injury in such patients, thereby indicating the need for further radiographic assessment. [Ann
Emerg Med. 2011;58:521-530.]

Please see page 522 for the Editor’s Capsule Summary of this article.

Provide feedback on this article at the journal’s Web site, www.annemergmed.com.
A podcast for this article is available at www.annemergmed.com.

0196-0644/$-see front matter
Copyright © 2011 by the American College of Emergency Physicians.
doi:10.1016/j.annemergmed.2011.06.008

Volume , .  : December  Annals of Emergency Medicine 521

Cervical MRI and Midline Tenderness Ackland et al

Editor’s Capsule Summary cervical CT findings that are negative for acute injury, MRI is
generally indicated in patients with trauma-related neurologic
What is known on this topic symptoms and signs.

Among alert, neurologically intact trauma patients Goals of This Investigation
with midline cervical spine tenderness but negative Our objective was to determine the presence of MRI-detected
computed tomography (CT) results, some have soft
tissue injury on MRI scanning. injury, if any, and the characteristics associated with injury extent in
alert, neurologically intact trauma patients with CT findings
What question this study addressed negative for acute injury. Our scope was not limited to the
detection of instability but focused more broadly on the
How frequent are such injuries, and what clinical characterization of injury that may also result either in delayed
features predict them? neurologic deficit or long-term pain or disability.

What this study adds to our knowledge MATERIALS AND METHODS
Study Design and Setting
In this prospective study of 178 patients with
midline pain and a CT negative for acute injury, We report here the cross-sectional analysis of baseline
44% had acute injury on MRI, 33 of 78 of whom information collected as part of a prospective cohort study
were treated with cervical collars and 5 of 78 of conducted at a Level I adult trauma center. The institution is a
whom received operative treatment. Advanced referral center for metropolitan and regional major trauma patients
cervical degenerative disease, thoracolumbar and has 15,000 trauma presentations per year, of whom 1,500
fracture, and multidirectional forces predicted patients are categorized as having major trauma. Institutional ethics
greater degree of injury. committee approval was obtained and participating patients’
informed consent was obtained in writing.
How this is relevant to clinical practice
Selection of Participants
These results cannot yet be applied to the general Consecutive, adult, alert trauma patients (Glasgow Coma
population of trauma patients, pending further
study to confirm the applicability of these findings. Scaleϭ15), presenting to the emergency department (ED) during a
Further study may determine which variables are 2-year period to January 2009, were prospectively evaluated for
useful in decisionmaking and which injuries are inclusion in the study. The real-time electronic ED admissions
clinically important. information system was used to alert the study investigators to
potential patients, and the ED medical and nursing staff was
INTRODUCTION involved in the project, assisting by alerting the principal
Background and Importance investigator to potential participants by pager or telephone at any
hour of the day. The daily CT and MRI booking schedules were
In many cases, midline cervical tenderness after trauma is absent also used as a safety net to ensure that all eligible patients had been
or resolves quickly, allowing clearance of the cervical spine in the identified and screened. Those patients with persistent cervical
absence of intoxication, altered conscious state, painful distracting midline tenderness on palpation in the setting of CT findings that
injury, persistent focal neurologic deficit,1-4 and acutely abnormal were negative for acute injury, and who subsequently underwent
cervical computed tomography (CT) findings. However, MRI studies under the institutional protocol, were included (Figure
persistence of midline cervical tenderness on palpation in alert, 1). Patients were initially assessed by a board-certified emergency
neurologically intact patients presents a clinical dilemma because it physician, and referral for spine review occurred when persistence
is unclear which patients require further investigation. The presence of midline cervical tenderness was evident.
of fractures alerts clinicians to the possibility of injury to other
cervical structures, but the absence of acute, positive CT findings is Exclusion criteria comprised:
often erroneously equated with lack of injury. The level of ● fractures of the base of skull or cervical or upper thoracic
efficiency in identifying occult cervical disc or ligamentous injury in
the absence of fracture is uncertain under many cervical spine spine;
clearance protocols. Detection of occult cervical spine injuries and ● painful distracting injury or evidence of intoxication
the determination of their clinical significance are imperative in the
avoidance of missed injuries, the assignment of appropriate according to the criteria of the National Emergency
management, and the mitigation of long-term morbidity. X-Radiography Utilization Study (NEXUS);1-4
● persistent focal neurologic deficit;
In recent years, magnetic resonance imaging (MRI) has emerged ● history of cervical spine surgery or injury; and
as a definitive and highly sensitive tool in the detection of acute ● MRI conducted greater than 96 hours postpresentation.
cervical discoligamentous injury.5 However, in the setting of Patients with minor, single, stable, nondisplaced, low thoracic or
lumbar transverse or spinous process fracture were included, as
these injuries were not considered by the treating physician to be

522 Annals of Emergency Medicine Volume , .  : December 

Ackland et al Cervical MRI and Midline Tenderness

Figure 1. Summary of included and excluded patients.

significantly distracting from the cervical tenderness, and MRI was included early (mild) degeneration and advanced (moderate and
ordered on the basis of the National Emergency X-Radiography severe) degeneration.
Utilization Study criterion of midline tenderness alone. Also
included were patients with transient neurologic deficit, which fully As per the institutional spinal clearance protocol, cervical
resolved shortly after presentation to the ED and before ordering of spine MRI was indicated for persistent midline cervical
the MRI for investigation of midline tenderness. tenderness on palpation in the absence of cervical fracture or
malalignment on CT imaging. MRI scans were performed on a
Methods of Measurement 1.5-T scanner (General Electric Signa Excite; GE Healthcare,
Admission cervical spine evaluation for all NEXUS–positive Milwaukee, WI), using the institutional trauma MRI scanning
protocol: sagittal T1 spin echo sequence: recovery time 440 ms,
patients included 64-slice CT from the skull base to T5 (General echo time 9 ms, spacing 4 mm; short TI inversion recovery
Electric LightSpeed VCT; GE Healthcare, Waukesha, WI). The sequence: recovery time 3,240 ms, echo time 36 ms; T2 fast
admission trauma CT scanning protocol included helical image spin echo fat saturation sequence: recovery time 3,000 ms, echo
acquisition at a 0.625-mm interval, with a pitch of 0.5 mm and a time 99 ms; and axial T2 images: recovery time 500 ms, echo
field of view of 20 cm. The images were transferred to the time 20 ms.
institutional Picture Archiving and Communication System after
2/2-mm axial and sagittal reconstruction on bone algorithm. The initial 100 images were reviewed independently by 2 senior
trauma radiologists from different institutions (D.K.V. and G.J.F.),
For the purposes of the study, the CT images were reviewed by who were not blinded to the purpose of the investigation but were
a senior trauma radiologist for evidence of acute injury and cervical blinded to patient identity, history, mechanism of injury, and
spondylosis. Evaluation of intervertebral disc height, vertebral results of other imaging modalities. The proportions of exact
apophyseal osteophytic morphology, spinal canal patency, vertebral agreement between radiologists for the identification and grading of
body endplate condition, and facet and uncovertebral joint injuries were calculated. Due to the potential for a low rate of
integrity was undertaken in the assessment of the severity of injury and the fact that only exact agreement was measured, the
spondylosis. For study purposes, cervical spondylosis was classified general level of fair agreement according to Altman10 was accepted.
as mild (no disc narrowing, osteophytes Ͻ2 mm and no significant Once these proportions were established as acceptable, the
canal stenosis), moderate (visible disc narrowing, osteophytes Ͼ2 remaining images were reported in the same fashion by a single
mm and visible central or lateral canal stenosis), and severe (complete senior trauma radiologist. With an imaging reporting pro forma
loss of disc height with or without a vacuum phenomenon, specifically designed for the study, the radiology review included
osteophytes Ͼ2 mm and significant canal stenosis).6-9 Further particular attention to the status of the anterior longitudinal
categorization of cervical spondylosis during statistical analysis ligament, intervertebral discs, posterior longitudinal ligament,

Volume , .  : December  Annals of Emergency Medicine 523

Cervical MRI and Midline Tenderness Ackland et al

spinal cord, ligamenta flava, and posterior ligamentous complex. associations of such factors with cervical spine injury according to
Findings were categorized and graded according to presence and the number of cervical spine columns injured12,13 were determined
severity of acute injury. Additionally, comments were required to with an ordinal logistic regression (proportional odds) model that
be made about the presence and extent of prevertebral and included the covariates age, sex, mechanism of injury, object
posterior muscle edema; the presence, extent, and maximum impacted, direction of initial spine force, quality of midline
thickness of epidural hematoma; and the presence, extent, and tenderness, transient neurology, additional injuries sustained, and
severity of cervical spondylosis. Associated findings related to the CT-detected cervical spondylosis. The remaining variables were not
integrity of minor ligaments and membranes were reported, as were included in the model because it was considered a priori that these
incidental abnormalities.11 MRI-detected injury was categorized variables were captured in the included covariates. There were no
into spinal columns, adapted according to the thoracolumbar spine missing data on variables in the cross-sectional analysis. The
concept of Denis:12,13 the anterior column is formed by the associations are presented as odds ratios (ORs) with 95%
anterior longitudinal ligament, anterior annulus fibrosus, and the confidence intervals (CIs) (SPSS WIN, version 17.0; SPSS Inc.
anterior component of the vertebral body; the middle column is [now IBM SPSS], Chicago, IL).
formed by the posterior vertebral body, posterior annulus fibrosus,
and the posterior longitudinal ligament; and the posterior column RESULTS
extends from the ligamenta flava to include the posterior bony and Characteristics of Study Subjects
ligamentous complexes.
There were 9,152 trauma patients who underwent admission
Data Collection and Processing cervical spine CT during the 2-year study period to January 2009.
Data collection included the findings of clinical and Of those patients, 741 had CT findings negative for acute injury
and persistent midline cervical tenderness. From this group, 178
radiographic cervical spine assessment, patient medical history, eligible patients were recruited to the study. The
demographic information, accident details, including paramedic inclusion/exclusion information is presented in Figure 1 and
evaluation and management, and patient-reported symptoms patient characteristics are presented in Table 1.
(Table 1). Given the fact that the patients were alert and
recruitment occurred as soon after presentation as practicable, the For the study cohort, cervical spine MRI was conducted at a
primary source of information about mechanism of injury and median of 37.5 hours after presentation to the ED (interquartile
accident details was the patient. Additional information provided range [IQR] 22.5 to 55.1 hours). Of these patients, there were 28
by paramedics and witnesses to the accident was also collected. A (15.7%) who underwent MRI while in the ED. Patients were
study-specific standard data collection tool was used, and a single admitted to the ED short-stay unit for observation of soft tissue
experienced researcher was responsible for data collection. The data injury or ongoing significant pain. The median time to MRI for
collection tool was completed before the MRI, and therefore both this subgroup was 19 hours (IQR 10.0 to 27.1). There were 30
the patients and the investigator were blinded to the MRI patients (16.9%) who were discharged from the ED in cervical
results at this stage. Specific mechanisms of injury were collars to return for outpatient MRI, at a median of 49 hours
examined to ascertain the most dominant direction of force to postpresentation (IQR 31.8 to 62.1 hours), and the remaining 120
the cervical spine during the trauma, and findings were patients (67.4%) were admitted to the hospital at the discretion of
categorized into flexion, extension, axial loading, lateral the trauma unit; the median time to MRI for this group was 40
bending, and axial rotation. Multidirectional biomechanical hours (IQR 24.0 to 59.8 hours).
forces were noted to have occurred when 2 or more such
directions of force were present during the injury mechanism. The proportions of exact agreement between radiologists for
For patients involved in road trauma, the vehicle crash status the identification and grading of injuries to the anterior
was categorized as no impact; collision with nonrigid objects, longitudinal ligament, disc, posterior longitudinal ligament,
including other vehicles, animals, and freeway guardrails; and cord, ligamenta flava, and posterior ligamentous complex for the
collision with rigid objects, including poles and trees. initial 100 patients were found to be 85%, 87%, 96%, 97%,
99%, and 88%, respectively.
Outcome Measures
The primary outcome measures included the presence and Main Results
Seventy-eight patients (44%) had acute cervical spine injuries
extent of MRI-detected injury of the cervical ligaments,
intervertebral discs, spinal cord, and associated soft tissues. detected on MRI. Of these patients, 62% had a single-column
Secondary outcome comprised the association of factors with the injury (nϭ48), 19% had two-column injuries (nϭ15), and 6%
extent of injury according to the number of spinal columns injured, had three-column injuries (nϭ5). Of the remaining 10 patients, 6
as per the three-column theory of Denis.12,13 had isolated posterior muscle edema, 2 had alar ligamentous
edema, 1 had epidural hematoma, and 1 had atlanto-occipital
Primary Data Analysis edema. Of the injured patients, 38 (21% of the total cohort) had
Summary statistics were generated for patient-related, injuries managed clinically, with 33 patients treated in cervical
collars for 2 to 12 weeks and 5 patients undergoing operative
mechanism-related, radiographic, and clinical factors. The management by anterior cervical discectomy and fusion (Figure 1).
The rationale for operative management included cord

524 Annals of Emergency Medicine Volume , .  : December 

Ackland et al Cervical MRI and Midline Tenderness

Table 1. Patient characteristics according to spinal columns injured.12*

Factor No Injury, Minor Injury: No Single-Column Two-Column Three-Column
n‫؍‬100 Columns, n‫؍‬10† Injury, n‫؍‬48 Injury, n‫؍‬15 Injury, n‫؍‬5

Age, y, median (IQR) 31.5 (24.3–44.0) 31.5 (24.0–47.0) 36.5 (27.3–48.0) 48.0 (24.0–54.0) 54.0 (45.5–63.0)
Sex
Male 47 (47) 6 (60) 28 (58.3) 14 (93.3) 5 (100)
Female 53 (53) 4 (40) 20 (41.6) 1 (6.6) 0
Mechanism of injury
MVC 43 (43) 8 (80) 19 (39.6) 4 (26.6) 3 (60)
Sport/recreation 12 (12) 0 6 (12.5) 4 (26.6) 1 (20)
Falls 19 (19) 9 (18.8) 4 (26.6)
MBA 1 (10) 5 (10.4) 0
Pedestrian or cyclist 9 (9) 1 (10) 7 (14.6) 0 0
Assault 9 (9) 2 (4.2) 2 (13.3) 1 (20)
Object struck 8 (8) 0 1 (6.6) 0
None 0 24 (50)
Other vehicle 50 (50) 8 (37.5) 9 (60) 2 (40)
Pole or tree (rigid) 42 (42) 1 (10) 6 (12.5) 4 (26.7) 2 (40)
Direction of initial spine force 7 (70) 2 (13.3) 1 (20)
Single direction 8 (8) 2 (20) 33 (68.8)
Multidirectional 15 (31.3) 11 (73.3) 3 (60)
Vehicle rollover 86 (86) 8 (80) 4 (26.7) 2 (40)
Not in vehicle 14 (14) 2 (20) 28 (58.3)
No 14 (29.2) 10 (66.7) 2 (40)
Yes 52 (52) 2 (20) 3 (20) 3 (60)
Patient-reported neck pain at scene 43 (43) 7 (70) 6 (12.5) 2 (13.3)
No 1 (10) 0
Yes 5 (5) 15 (31.3) 4 (26.7)
Head strike 5 (50) 33 (68.8) 11 (73.3) 2 (40)
No 26 (26) 5 (50) 3 (60)
Yes 74 (74) 8 (16.7) 2 (13.3)
LOC 3 (30) 40 (83.3) 13 (86.7) 1 (20)
No 26 (26) 7 (70) 4 (80)
Yes 74 (74) 22 (45.8) 9 (60)
Midline cervical tenderness 5 (50) 26 (54.2) 6 (40) 2 (40)
Diffuse 62 (62) 5 (50) 3 (60)
Focal 38 (38) 11 (22.9) 2 (13.3)
Transient neurology 7 (70) 37 (77.1) 13 (86.7) 2 (40)
No 34 (34) 3 (30) 3 (60)
Yes 66 (66) 34 (70.8) 11 (73.3)
Additional injuries 6 (60) 14 (29.2) 4 (26.7) 3 (60)
None 77 (77) 4 (40) 2 (40)
Minor limb# 23 (23) 35 (72.9) 13 (86.7)
Minor spinal# 9 (90) 6 (12.5) 1 (6.7) 4 (80)
CT-detected cervical spondylosis 88 (88) 1 (10) 7 (14.6) 1 (6.7) 1 (20)
None 11 (11)
Early 0 28 (58.3) 5 (33.3) 0
Advanced 1 (1) 12 (25) 2 (12.3)
7 (70) 8 (53.3) 0
76 (76) 0 8 (16.7) 0
17 (17) 5 (100)
3 (30)
7 (7)

MVC, Motor vehicle crash; MBA, motorbike accident; LOC, loss of consciousness at accident scene; #, fracture.

*Data are presented as No. (%) unless otherwise indicated.
†Minor injury: no columnsϭisolated posterior muscle edema, epidural hematoma, or alar ligamentous edema. Sport/recreationϭfootball, rugby, waterskiing, snow-
boarding. Fallsϭladders, trampolines, from horses, high (Ͼ1 m) and low falls.

compression, combined compression and edema, and delayed (31%) had advanced spondylosis. In the subgroup of patients with
instability (Table 2). There were no cases of neurologic injury to 2 or 3 spinal columns, early and advanced degeneration
deterioration. The CT and MRI images for patient 1, who was was evident on CT images in 2 of 20 (10%) and 13 of 20 (65%)
managed operatively, are presented in Figures 2 and 3. cases, respectively. Highlighting the proportions of patients with
CT-detected spondylosis who were injured, of the patients with
Cervical spondylosis was detected on CT images before MRI in early degeneration, 14 of 31 (45%) had injury detected on MRI,
62 of 178 patients (35%), with early and advanced degeneration in and of the patients with advanced degeneration, 24 of 31 (77%)
equal proportions (Table 1). For patients with MRI-detected had MRI-detected injury. Incidentally, all 5 patients requiring
injury, 14 of 78 patients (18%) had early and 24 of 78 patients

Volume , .  : December  Annals of Emergency Medicine 525

Cervical MRI and Midline Tenderness Ackland et al

Table 2. Operative stabilization cases.

Age/ Mechanism of Degenerative Change Injury Comments Management
Patient Sex Injury Rationale

1 66 M Pedestrian vs Multilevel moderate Acute C5/6 disc extrusion Cleared after CT, ACDF C5/6

vehicle at low spondylosis. Marked with caudal migration of immediate worsening for cord

speed. loss of disc height at disc fragment posterior neck pain and midline compression.

C6/7. to C6 body. Severe tenderness, no abnormal

canal stenosis, partial neurology. MRI

tear ALL, rupture of PLL undertaken.

and edema of

interspinous ligaments.

2 42 M Rear seat Mild to moderate multilevel Acute C5/6 disc edema, During operative procedure, ACDF C5/6

passenger, rear- spondylosis with C4/5 & partial tear ALL, focal PLL found to be torn, for cord

end collision at 5/6 disc space cord edema without with prolapsed disc edema and

60 km/h. narrowing and hemorrhage, material and posterior compression.

osteophytes. interspinous edema. osteophytes causing cord

compression. No

abnormal neurology.

3 50 M Cyclist vs truck. Moderate disc C5/6 disc extrusion with Transient neurology. ACDF C5/6

degeneration most cord flattening and for cord

pronounced at C5/6, edema, disrupted PLL, edema and

where there is a large C3–7 posterior compression.

posterior disc. ligamentous edema.

4 51 M Motor vehicle driver Multilevel moderate disc C4/5: partial tear ALL and Collar 6/52, anterolisthesis ACDF C4/5

vs tree at high degeneration. lig flavum, C3–5 of C4 on C5, with facet for delayed

speed. interspinous edema, joint subluxation instability.

C2–5 prevertebral observed on

edema, C3–6 facet flexion/extension at

capsular strain. 6/52.

5 31 F Thrown from horse. C3–7 moderate to severe Acute C5/6 disc extrusion LOC 3–4 min, transient ACDF C5/6

spondylosis, with cord compression. neurology. for cord

osteophytes. compression.

M, Male; ALL, anterior longitudinal ligament; PLL, posterior longitudinal ligament; ACDF, anterior cervical discectomy and fusion, 6/52ϭ6 weeks; F, female.

Figure 2. A, Sagittal CT image indicating multilevel Figure 3. A, Sagittal short TI inversion recovery sequence
spondylosis with severe loss of intervertebral disc height and image: rupture of PLL suggested at C5/6, and associated
associated osteophytes, particularly at C6/7. B, Sagittal T2 cord edema. B, Axial T2 image at C5/6 level, indicating the
MRI image indicating injury to the C5/6 intervertebral disc, degree of cord compression (patient 1, Table 2).
which is immediately superior to the greatest degree of
spondylosis evident in Figure 2A. C5/6 disc extrusion and associated with a greater extent of injury according to the number
resultant cord compression (patient 1, Table 2). of cervical columns injured. The interpretation therefore is that
patients with advanced spondylosis on CT have an 11.6-fold
operative management of cervical spine injury in our study had increase in the log odds of having a higher number of spinal
advanced cervical spondylosis (Table 2). columns injured, given that all of the other variables in the model

Ordinal logistic regression analysis revealed that advanced CT-
detected cervical spondylosis (OR 11.6; 95% CI 3.9 to 34.3) was

526 Annals of Emergency Medicine Volume , .  : December 

Ackland et al Cervical MRI and Midline Tenderness

are held constant. Similarly, minor thoracolumbar fractures (OR have validated17 and modified18 the three-column theory. The
5.4; 95% CI 1.5 to 19.7) and multidirectional cervical spine forces model remains the most popular and clinically useful and has
(OR 2.5; 95% CI 1.2 to 5.2) were also associated with increased subsequently been widely applied to the cervical spine,19,20
extent of MRI-detected injury in this patient group. allowing the extent of discoligamentous and osseous injury to be
described in terms of the risk to neurologic integrity, with
LIMITATIONS instability defined as injury to 2 or 3 spinal columns.12,13
The main limitation of our study was the potential for Nonfracture-related discoligamentous injury, as found in our study,
although often of less risk to neural structures than injury involving
measurement (observer) bias, given that the study personnel were osseous compromise, can be unstable21-23 and is more difficult to
not blinded to the hypothesis of the study. A single, experienced detect in the acute setting, particularly when normal alignment is
researcher recruited the study participants and was responsible for evident. With a higher number of columns injured representing a
data collection. However, a standard data collection tool was used, greater risk of instability, particularly that related to the disruption
the demographic data were collected before MRI, and the data were of ligamentous integrity, knowledge of the extent of cervical
consistently complete. In addition, the reporting radiologists were column compromise is important in alerting clinicians to the
cognizant that participants had all undergone MRI for injury potential for delayed instability.24
screening as a result of persistent midline tenderness. Despite this,
the radiologists were senior trauma MRI specialists with extensive Although devastating neurologic deficit attributable to acute
experience, were blinded to patient, mechanism, and clinical cervical spine injury is uncommon, the detection of potential
presentation information, and were from different institutions. instability and the prevention of secondary injury remain the
Also, the imaging data collection tool was standardized, and the primary goals of cervical spine clearance protocols. As a result,
proportions of interrater agreement were consistently fair. These comparatively minor symptoms, such as isolated midline cervical
measures assisted in the minimization of observer bias in our study. tenderness, are often underestimated as indicators of injury when
traditional, first-line imaging results are negative. There is no clear
Patient selection also represents a bias, given that the institution consensus on the need for further screening of such patients. The
is a major referral center for trauma from metropolitan and regional cervical spine practice management guidelines of the Eastern
trauma services. Although none of the patients in our study was Association for the Surgery of Trauma25 recommend that MRI be
finally classified as having major trauma, the participants were one of 3 options available to clinicians, along with management
probably representative of patients presenting to metropolitan and with cervical collar and flexion-extension radiography. In contrast,
regional centers. As a result of the reputation of our institution as a the American College of Surgeons advocates that flexion-extension
major trauma center, patients with minor trauma may have elected radiography be performed in such cases,26 whereas the American
to present to the ED for cervical spine assessment, and hence College of Radiology suggests that flexion-extension radiographs are
selection bias may exist. Additionally, only patients who underwent of little value until pain subsides and advocates the use of MRI for
MRI were included in our study, and our institutional practice of possible ligamentous injury without directly recommending MRI
ordering MRI for patients with persistent midline cervical for persistent midline cervical tenderness.27
tenderness and negative CT may not be part of trauma protocols at
other institutions. Much of the evidence regarding MRI-detected cervical spine
injury published to date involves obtunded patients, who are
Further limitations included the single center and the potential presumed to have sustained multiple injuries and lack the ability to
inaccuracy of history reporting from patients, whose quality of participate in clinical assessment.22,28-33 To our knowledge, few
recall about mechanism of injury is often unable to be validated. studies have focused on alert trauma patients with CT findings that
are negative for acute injury, and no studies have specifically
DISCUSSION centered on the relationship between midline cervical tenderness
In 1963, Holdsworth14 defined spinal instability as rupture of and MRI findings in this group. Instead, cervical MRI has generally
been conducted in alert patients as a result of neurologic deficit or
the posterior ligamentous complex, after the development of his patient-reported neck pain.
two-column thoracolumbar theory, in which the anterior column
comprised the anterior longitudinal ligament, intervertebral disc, Two particular variables appear to be of importance in
and posterior longitudinal ligament and the posterior column studies in which MRI findings constitute the primary outcome:
involved all components that were posterior to the posterior the time to MRI from injury and the strength of the magnetic
longitudinal ligament. Subsequent biomechanical studies, however, field, as measured in Tesla (T). In the absence of serial MRI
found that instability in flexion, for example, also required rupture studies, anecdotal evidence suggests that the magnetic properties
of the posterior longitudinal ligament and posterior annulus of edema and hemorrhage reduce over time, such that MRI
fibrosus.15,16 As a result, after a retrospective study of 412 conducted greater than 72 hours after injury is considered less
thoracolumbar injuries, the three-column theory of Denis12,13 was efficient for non-osseous injury detection.22,34-36 Clinical MRI
developed in 1983. This theory introduced a middle column, systems generally tend to have field strengths of 0.2 to 3-T,
including the posterior component of the vertebral body, the whereas research systems of 8-T and above are available.37
posterior longitudinal ligament, and the posterior annulus Generally, the field strength affects the image resolution
fibrosus.12,13 Further cadaver model and biomechanical studies

Volume , .  : December  Annals of Emergency Medicine 527

Cervical MRI and Midline Tenderness Ackland et al

achievable with adequate signal to noise,38,39 and the signal-to- hours of injury, no acute pathology was found.42 The MRI,
noise ratio increases with the field strength on a linear basis.39 however, was conducted on a 0.5-T scanner, which may have
Low-field MRI systems of 0.2 to 0.5-T are most commonly resulted in false-negative findings due to poor image quality.
used for musculoskeletal extremity imaging. Currently, the Similarly, in a larger prospective study, Kongsted et al43 examined
majority of MRI scanners installed in clinical institutions are the MRI findings of 178 patients with acute neck pain, after 0.2-T
high-field systems of 1.5-T,37,39 and there is a trend toward the imaging conducted at a median of 13 days post injury. Traumatic
acquisition of 3-T systems. However, there is considerable disc bulge or protrusion was evident in 0.5% of patients, but again
debate over the optimal field strength, as although the image the magnetic field strength and time to MRI were suboptimal and
resolution provided by the ultra high-field (Ն3 T) technology may have resulted in missed injury.
systems is improved, there is also increased metallic foreign
body artifact, increased chemical shift artifact, and problems in The majority of the remaining evidence is conflicting and based
maintaining homogeneity of the main magnetic field.37,40 on retrospective data, with wide variation in methodologies. In one
Additionally, new symptoms such as headache, vertigo, and such retrospective study, a subset of 20 patients with Glasgow
spine pain have been associated with 3-T MRI scanning.41 As a Coma Scale score of 15, whose only indication for MRI was neck
result, 1.5-T imaging currently appears optimal. pain, was assessed after normal 40-slice CT.44 Three patients
(15%) were found to have 1.5-T MRI-detected cervical spine
Schuster et al33 prospectively examined the findings of a subset injury, 2 of whom were treated in collars for joint capsule strain and
of 93 alert, neurologically intact patients with persistent neck pain interspinous ligamentous edema and 1 of whom was managed
and normal helical CT findings who underwent 1.5-T MRI operatively for a three-column injury. The authors concluded that
scanning. No patients were found to have clinically significant MRI should be included in the evaluation of the cervical spine in
injury detected on MRI, and the authors concluded that normal this patient group.
CT findings are sufficient for cervical spine clearance in this group.
There are 3 main differences in methodology between Schuster’s Our study found a higher proportion of occult
study and ours. First, the purely patient-reported symptom of discoligamentous injury overall, and a higher proportion of
persistent neck pain in the study by Schuster et al33 differs operatively managed injury, than studies in which the rationale for
considerably from the clinician-elicited sign of persistent midline MRI was patient-reported neck pain, in contrast to midline cervical
cervical tenderness in our study. In many cases, patients report tenderness. In addition to the presence of tenderness on palpation,
generalized neck pain lateral to the midline, which may be the prospective nature of our study and the strict selection criteria
attributed to paraspinal muscle edema and is less likely to be an may have contributed to this finding. Additionally, as our patients
indication of disc or ligamentous injury. In addition, minor presented to a Level I trauma center, a high level of expertise in
muscular strain can result in neck pain but may not be sufficiently image interpretation was available, the CT and MRI imaging
significant or extensive to allow detection of edema on MRI. The protocols were consistent, and MRI was conducted in a timely
other main difference between the symptom of neck pain and the manner, thereby minimizing false-negative results. However, in
sign of midline cervical tenderness is that palpation allows for a other centers the need for surgery in some of these patients may be
degree of reproducible elicitation of pain. As a result, patients assessed differently.
reporting neck pain, but without midline tenderness on palpation,
were excluded from our study. Second, the time to MRI, which We found that advanced spondylotic degeneration, the
was not specified by Schuster et al,33 may be an important factor in presence of minor isolated thoracolumbar fracture, and
the lack of injury detected. MRI was conducted at a median time of multidirectional spine forces were associated with a greater
37.5 hours in our study, hence the potential for false-negative extent of injury in terms of the number of spinal columns
results was minimized. Third, Schuster et al33 hypothesized that involved. Intervertebral disc degeneration occurs as the nucleus
significant ligamentous injury without fracture will manifest in pulposus progressively loses hydration over time, resulting in
immediate neurologic deficit. However, following brittleness, loss of disc height, and bulging.45 These changes
discoligamentous injury that initially presents without neurologic occur as a result of reduction in the integrity of the annulus
deficit, particularly with significant flexion/extension mechanisms, fibrosus through fissure development.46 Subsequently, the
delayed instability and resultant secondary neurologic deficit is mechanical characteristics of the cervical vertebral column alter,
possible after subsequent caudal migration of disc matter via a such that the discs and ligaments above and below the
ruptured posterior longitudinal ligament. In our study, MRI- spondylotic area are at risk of increased motion and potential
detected posterior longitudinal ligament rupture constituted the injury during future traumatic mechanisms, as evident in Figure
rationale for operative management in 3 of 5 cases, and in a further 2, and this finding is supported by those of other studies.24,47-50
case, the presence of focal cord edema manifested suspicion of a
lack of posterior longitudinal ligament integrity, which was The association of thoracolumbar injury with cervical spine
confirmed visually during the procedure. fracture has been well documented,51,52 and additional attention to
clinical and radiographic thoracolumbar assessment after fracture
In a further prospective study of 40 alert trauma patients, also detection under meticulous cervical spine protocols has resulted.
with patient-reported neck pain, who underwent MRI within 48 Velocity or impact direction of spine force that is significant
enough to result in thoracolumbar spinous or transverse process
fracture may therefore have also been absorbed to a degree by the

528 Annals of Emergency Medicine Volume , .  : December 

Ackland et al Cervical MRI and Midline Tenderness

cervical spine, resulting in occult disc or ligamentous injury relationships in any way related to the subject of this article
presenting as midline cervical tenderness, which may account for as per ICMJE conflict of interest guidelines (see
our finding. www.icmje.org). The authors have stated that no such
relationships exist. Funded by the Transport Accident
The association of occult cervical spine injury with Commission, Victoria, Australia (grant No. NV16).
multidirectional spine forces found in our study is supported in
human cadaver model studies, in which force and velocity have Earn CME Credit: Continuing Medical Education is available for
been correlated with cervical spine structure responses during this article at www.ACEP-EMedHome.com.
various biomechanical mechanisms.24,47,53 The combining of
one force vector with another has the potential to magnify the Publication dates: Received for publication December 8,
energy acting on the cervical structures, and the result is the 2010. Revisions received February 14, 2011, and April 15,
induction of additional changes in the components of the spine. 2011. Accepted for publication June 14, 2011. Available
When the tolerance of cervical structural components is online August 5, 2011.
exceeded by the strength of the force, injury occurs.
Additionally, certain structures offer greater protection against Presented as preliminary findings at the Asian Oceanic
force and vector direction than others. For example, in Congress of Radiology, March 2010, Taipei, Taiwan; the Spine
comparison with the supraspinous ligament, a component of Society of Australia Annual Scientific Meeting, April 2010,
the posterior ligamentous complex that strongly attaches to the Christchurch, New Zealand; the 13th International Conference
spinous processes, the intervertebral discs and other cervical on Emergency Medicine, June 2010, Singapore; and the
ligaments offer relatively weak protection against injury.24 Annual Scientific Meeting of the Spine Society of Europe
(Eurospine 2010), September 2010, Vienna, Austria.
Despite the absence of cervical spine injury on CT images,
persistent midline cervical tenderness can indicate the presence of Address for correspondence: Helen M. Ackland, MHSc, E-mail
occult discoligamentous injury, detectable on MRI. In our study, [email protected].
advanced cervical spondylosis, minor isolated thoracolumbar
fracture, and multidirectional spine forces were found to be REFERENCES
associated with greater degrees of cervical spine trauma, as defined 1. Hendey GW, Wolfson AB, Mower WR, et al. Spinal cord injury
by the number of spinal columns injured. However, further study is
required to determine which variables reliably predict injury in without radiographic abnormality: results of the National
patients with persistent midline cervical tenderness and which of Emergency X-Radiography Utilization Study in blunt cervical
these injuries are clinically important and require identification at trauma. J Trauma. 2002;53:1-4.
the time of the ED visit. 2. Hoffman JR, Mower WR, Wolfson AB, et al. Validity of a set of
clinical criteria to rule out injury to the cervical spine in patients
The authors gratefully acknowledge the assistance of Rosie Bushnell, with blunt trauma. National Emergency X-Radiography Utilization
RN, and the ED medical and nursing staff for support with patient Study Group. N Engl J Med. 2000;343:94-99.
recruitment, Thomas Kossmann, MD, for assistance in project protocol 3. Hoffman JR, Wolfson AB, Todd K, et al. Selective cervical spine
development and patient review, Natalie Carr for support with MRI radiography in blunt trauma: methodology of the National
booking, and Robert Wills, DipAppSci(MedRad), DipEd(MedSci), for Emergency X-Radiography Utilization Study (NEXUS). Ann Emerg
advice regarding radiology project administration. Med. 1998;32:461-469.
4. Panacek EA, Mower WR, Holmes JF, et al. Test performance of
Supervising editor: Allan B. Wolfson, MD the individual NEXUS low-risk clinical screening criteria for cervical
spine injury. Ann Emerg Med. 2001;38:22-25.
Author contributions: HMA and PAC conceived the study and 5. Ackland HM, Cooper DJ, Malham GM, et al. Magnetic resonance
were responsible for recruitment and managed the data. HMA, imaging for clearing the cervical spine in unconscious intensive
PAC, DKV, GJF, DJC, RW, GMM, JVR, and ODW provided input care trauma patients. J Trauma. 2006;60:668-673.
into the study design. HMA secured research funding. HMA, 6. Peterson C, Bolton J, Wood AR, et al. A cross-sectional study
PAC, DKV, GJF, DJC, RW, GMM, and JVR supervised the correlating degeneration of the cervical spine with disability and
conduct of the trial. HMA was responsible for primary data pain in United Kingdom patients. Spine. 2003;28:129-133.
collection. DKV and GJF were responsible for review of 7. Peterson CK, Bolton JE, Wood AR. A cross-sectional study
images. GMM, SML, and JVR were responsible for patient correlating lumbar spine degeneration with disability and pain.
review and management decisions. RW provided advice on Spine. 2000;25:218-223.
statistical design. HMA and RW analyzed the data. PAC 8. Rosen RS, Fayad L, Wahl RL. Increased 18F-FDG uptake in
chaired the steering committee. HMA drafted the article, and degenerative disease of the spine: characterization with 18F-FDG
all authors contributed substantially to editing and revision. PET/CT. J Nucl Med. 2006;47:1274-1280.
HMA and PAC take responsibility for the paper as a whole. 9. Yu SW, Haughton VM, Ho PS, et al. Progressive and regressive
changes in the nucleus pulposus. Part II. The adult. Radiology.
Funding and support: By Annals policy, all authors are required 1988;169:93-97.
to disclose any and all commercial, financial, and other 10. Altman D. Practical Statistics for Medical Research. London,
England: Chapman & Hall; 1991.
11. Ackland HM, Marshall NL, Hall NDP, et al. Unilateral internal
jugular venous thrombosis in a young trauma patient. Injury Extra.
2009;40:246-248.
12. Denis F. Spinal instability as defined by the three-column spine
concept in acute spinal trauma. Clin Orthop Relat Res. 1984;
189:65-76.

Volume , .  : December  Annals of Emergency Medicine 529

Cervical MRI and Midline Tenderness Ackland et al

13. Denis F. The three column spine and its significance in the patients with normal computed tomographic results and no motor
classification of acute thoracolumbar spinal injuries. Spine. 1983; deficits. Arch Surg. 2005;140:762-766.
8:817-831. 34. Albrecht RM, Malik S, Kingsley DD, et al. Severity of cervical
spine ligamentous injury correlates with mechanism of injury, not
14. Holdsworth FW. Fractures, dislocations and fracture-dislocations with severity of blunt head trauma. Am Surg. 2003;69:261-265.
of the spine. J Bone Joint Surg Br. 1963;45B:6-20. 35. Muchow RD, Resnick DK, Abdel MP, et al. Magnetic resonance
imaging (MRI) in the clearance of the cervical spine in blunt
15. Panjabi MM, Hausfeld JN, White AA 3rd. A biomechanical study of trauma: a meta-analysis. J Trauma. 2008;64:179-189.
the ligamentous stability of the thoracic spine in man. Acta 36. Spitzer WO, Skovron ML, Salmi LR, et al. Scientific monograph of
Orthop Scand. 1981;52:315-326. the Quebec Task Force on Whiplash-Associated Disorders:
redefining “whiplash” and its management. Spine. 1995;20:1S-
16. Panjabi MM, White AA 3rd. Basic biomechanics of the spine. 73S.
Neurosurgery. 1980;7:76-93. 37. Wilkinson ID, Paley MNJ. Magnetic resonance imaging: basic
principles. In: Adam A, Dixon AK, Grainger RG, et al, eds.
17. Panjabi MM, Oxland TR, Kifune M, et al. Validity of the three- Grainger and Allison’s Diagnostic Radiology: A Textbook of
column theory of thoracolumbar fractures. Spine. 1995;20:1122- Medical Imaging. Vol 1. 5th ed. Philadelphia, PA: Elsevier Limited;
1127. 2008.87-107.
38. Semple JC. Magnetic resonance imaging. In: Easton S, ed. An
18. Louis R. Spinal stability as defined by the three-column spine Introduction to Radiography. Philadelphia, PA: Elsevier Limited; 2009:
concept. Anat Clin. 1985;7:33-42. 237-254.
39. McRobbie DW, Moore EA, Graves MJ, et al. MRI: From Picture to
19. Bozkus H, Ames CP, Chamberlain RH, et al. Biomechanical Proton. New York, NY: Cambridge University Press; 2007.
analysis of rigid stabilization techniques for three-column injury in 40. Schmidt GP, Wintersperger B, Graser A, et al. High-resolution
the lower cervical spine. Spine. 2005;30:915-922. whole-body magnetic resonance imaging applications at 1.5 and
3 Tesla: a comparative study. Invest Radiol. 2007;42:449-459.
20. Cybulski GR, Douglas RA, Meyer PR Jr, et al. Complications in 41. Weintraub MI, Khoury A, Cole SP. Biologic effects of 3 tesla (T)
three-column cervical spine injuries requiring anterior-posterior MR imaging comparing traditional 1.5 T and 0.6 T in 1023
stabilization. Spine. 1992;17:253-256. consecutive outpatients. J Neuroimaging. 2007;17:241-245.
42. Borchgrevink G, Smevik O, Haave I, et al. MRI of cerebrum and
21. Benzel EC, Hart BL, Ball PA, et al. Magnetic resonance imaging cervical columna within two days after whiplash neck sprain
for the evaluation of patients with occult cervical spine injury. injury. Injury. 1997;28:331-335.
J Neurosurg. 1996;85:824-829. 43. Kongsted A, Sorensen JS, Andersen H, et al. Are early MRI
findings correlated with long-lasting symptoms following whiplash
22. D’Alise MD, Benzel EC, Hart BL. Magnetic resonance imaging injury? a prospective trial with 1-year follow-up. Eur Spine J.
evaluation of the cervical spine in the comatose or obtunded 2008;17:996-1005.
trauma patient. J Neurosurg. 1999;91:54-59. 44. Menaker J, Stein DM, Philp AS, et al. 40-Slice multidetector CT:
is MRI still necessary for cervical spine clearance after blunt
23. Dusseldorp JR, Ow-Yang M, Mobbs RJ. Unrecognized ligamentous trauma? Am Surg. 2010;76:157-163.
instability due to high-energy, low-velocity mechanism of injury. 45. Castillo M, Smith JK, Mukherji SK. The spine. In: Lee JKT, Sagel
J Clin Neurosci. 2010;17:139-141. SS, Stanley RJ, et al, eds. Computed Body Tomography With MRI
Correlation. 3rd ed. Philadelphia, PA: Lippincott-Raven; 1998:
24. Cusick JF, Yoganandan N. Biomechanics of the cervical spine 4: 1449-1491.
major injuries. Clin Biomech. 2002;17:1-20. 46. Czervionke LF, Haughton VM. Degenerative disease of the spine.
In: Atlas SW, ed. Magnetic Resonance Imaging of the Brain and
25. Como JJ, Diaz JJ, Dunham CM, et al. Practice management Spine.Vol 2. 3rd ed. Philadelphia, PA: Lippincott Williams &
guidelines for identification of cervical spine injuries following Wilkins; 2002:1633-1713.
trauma: update from the Eastern Association for the Surgery of 47. Cusick JF, Yoganandan N, Pintar F, et al. Cervical spine injuries
Trauma Practice Management Guidelines Committee. J Trauma. from high-velocity forces: a pathoanatomic and radiologic study.
2009;67:651-659. J Spinal Disord. 1996;9:1-7.
48. Davis SJ, Teresi LM, Bradley WG Jr, et al. Cervical spine
26. American College of Surgeons Committee on Trauma. Spine and hyperextension injuries: MR findings. Radiology. 1991;180:245-251.
spinal cord trauma. In: Advanced Trauma Life Support (ATLS) 49. Yoganandan N, Kumaresan S, Pintar FA. Biomechanics of the
Student Course Manual. 8th ed. American College of Surgeons. cervical spine part 2. Cervical spine soft tissue responses and
2008:157–173. biomechanical modeling. Clin Biomech (Bristol, Avon). 2001;16:1-27.
50. Kumaresan S, Yoganandan N, Pintar FA, et al. Contribution of
27. American College of Radiology. ACR appropriateness criteria on disc degeneration to osteophyte formation in the cervical spine: a
suspected spine trauma. Available at: http://www.acr.org. biomechanical investigation. J Orthop Res. 2001;19:977-984.
Accessed July 7, 2010. 51. Hsu JM, Joseph T, Ellis AM. Thoracolumbar fracture in blunt
trauma patients: guidelines for diagnosis and imaging. Injury.
28. Schoenwaelder M, Maclaurin W, Varma D. Assessing potential 2003;34:426-433.
spinal injury in the intubated multitrauma patient: does MRI add 52. Winslow JE 3rd, Hensberry R, Bozeman WP, et al. Risk of
value? Emerg Radiol. 2009;16:129-132. thoracolumbar fractures doubled in victims of motor vehicle
collisions with cervical spine fractures. J Trauma. 2006;61:686-687.
29. Tomycz ND, Chew BG, Chang YF, et al. MRI is unnecessary to 53. Huelke DF, Nusholtz GS. Cervical spine biomechanics: a review of
clear the cervical spine in obtunded/comatose trauma patients: the literature. J Orthop Res. 1986;4:232-245.
the four-year experience of a Level I trauma center. J Trauma.
2008;64:1258-1263.

30. Hogan GJ, Mirvis SE, Shanmuganathan K, et al. Exclusion of
unstable cervical spine injury in obtunded patients with blunt
trauma: is MR imaging needed when multi-detector row CT
findings are normal? Radiology. 2005;237:106-113.

31. Como JJ, Thompson MA, Anderson JS, et al. Is magnetic
resonance imaging essential in clearing the cervical spine in
obtunded patients with blunt trauma? J Trauma. 2007;63:544-
549.

32. Stassen NA, Williams VA, Gestring ML, et al. Magnetic resonance
imaging in combination with helical computed tomography
provides a safe and efficient method of cervical spine clearance
in the obtunded trauma patient. J Trauma. 2006;60:171-177.

33. Schuster R, Waxman K, Sanchez B, et al. Magnetic resonance
imaging is not needed to clear cervical spines in blunt trauma

530 Annals of Emergency Medicine Volume , .  : December 