A recent study entitled, “Outcomes at 12 Months After Early Magnetic Resonance Imaging in Acute Trauma Patients With Persistent Midline Cervical Tenderness and Negative Computed Tomography” published in the June 2013 issue of Spine has attempted to shine a small amount of light on a topic that has cast a large shadow of doubt on Emergency Physicians (1). What should be done with the blunt trauma patient with persistent midline cervical tenderness after a negative CT?Those with neurological findings obviously require further investigations (usually MR) and with the obtunded patient we usually default to our institution’s policy, but for the otherwise intact, CT negative patient with persistent midline tenderness the management is unclear.
The authors of the abovementioned study, Ackland et al, attempted to answer this very question. Their initial findings were published in the Annals of Emergency Medicine in December of 2011 (2). At first glance their results appear fairly disheartening. Of the entire cohort of 178 patients with “normal” CTs, 78(44%) had acute cervical injury detected on MRI, 33(18.5%) were treated with cervical collars and 5 patients (2.8%) underwent operative management. Taken at face value, these results imply that CT alone is not acceptable to clear these patients from cervical stabilization. But these findings are not as clear-cut as they initially appear. The real question that this study brings up is how we interpret MR findings in acute cervical trauma. Are these findings truly uncovering occult spinal trauma which, if undetected, would hold true clinical repercussions? Or is this just the inevitable result of applying an overly sensitive test in a low risk population?
As with any diagnostic dilemma in the ED we must ask ourselves what is our end goal? Do we wish to diagnose every clinically irrelevant muscle and ligamentous strain or do we care about injuries with serious clinical consequences? Are we protecting our patients from occult injuries that can lead to functional disability or are we exposing them to the risks of over-testing, over-diagnosis and harmful interventions associated with these type-1-errors?
Contradictory findings – or are they?
A previous retrospective study, “Neurologic Deterioration Secondary to Unrecognized Spinal Instability Following Trauma- A Multicenter Study” from Spine in 2006, estimates the rate of neurologic deterioration to be somewhere around 0.025% (3). The authors were only able to find a total of 24 patients who deteriorated from a missed spinal injury from 8 trauma centers over an 8-year period. Half of these misses seemed to be from inadequate plain films or misreads of a CT rather than a true negative CT.
In the Ackland paper if we were to take the results of the MR at face value, the rate of missed injuries would be 44%. Clearly there are many methodological flaws when retrospectively gathering data, but the one advantage this paper has is the hindsight of clinical outcomes. Whereas in the Ackland paper all we have is an anatomic diagnosis of disease and the assumption of harm. There is little evidence that any of the “injuries” found on MR will cause permanent disability if they are not diagnosed at the initial presentation. In the authors’ follow-up paper (1) they found no difference in functional outcomes at 12 months between the patients with no findings on MR, minor findings on MR, and findings “requiring interventions”.Since this cohort was not randomized to treatment or no treatment it is hard to interpret these findings. We can deduce that the results indicate that MR findings are not an effective tool in determining future prognosis. According to logistic regression, the only factors that correlated with higher disability at 12 months were depression, accident compensation and annual income.Another possible explanation is that the interventions provided for patients with positive MR findings were so effective that they helped bridged the gap between the groups. However, if MR findings truly denoted a clinically relevant disease state then even with the most effective of interventions, the MR positive patients still should have experienced more disability at 12 month follow up when compared to MR negative patients.
To MR or not to MR?
The argument that all neurologically intact patients with a negative CT and persistent midline tenderness require a MR assumes that the MR will accurately identify occult injuries that would benefit from effective treatment. There are no studies directly comparing outcomes of blunt trauma patients with isolated midline tenderness randomized to CT alone or CT followed by MR. This would provide us with valuable answers.
In lieu of a study comparing outcomes of CT and MR to CT alone, let us examine the test characteristics of MR in acute spinal trauma. In an article by Rhin et al (5) the authors examined the MR findings on 89 prospective blunt trauma patients who presented for surgery to stabilize spinal fractures identified on CT. Inherently this is a much higher acuity population, as they are already undergoing surgical management of a spinal injury. This should help the physicians reading the MR, as the known fractures would likely facilitate their search for ligamentous injuries. The MR findings were read by two radiologists and two spinal surgeons, and were then compared to the gold standard of findings reported by the surgeon during surgery. Overall there was a large variation in the findings on MR compared to what was discovered during surgery. The sensitivities for identifying a ligamentous injury on MR, ranged from 78.4-100% and the specificities from 59.0-80.5%. In a second paper, the same authors (6) examined the accuracy of MR in a cohort of 47 blunt trauma patients who were scheduled to undergo spinal surgery. On this occasion they found the sensitivity ranged from 80-100% and the specificity ranged from 56-67%.
Neither the sensitivity nor specificity are acceptable for use in the clinical arena in this context, but the relatively low specificity is most problematic. Since neurological intact trauma patients with negative CTs are very low risk, missing a nonexistent injury should not be our concern. More troubling is the over diagnosis and unnecessary treatment that will result from a specificity as low as 56%.
The signal vs. the noise
In a study by Anderson et al (7) the authors examined 100 blunt trauma patients with persistent neck pain and negative CT imaging. They compared the findings on MR to 100 asymptomatic healthy volunteers. It is important to note that case-control study is not the ideal methodology to base our clinical decisions upon. However, this study allows us to examine the baseline noise in the general population and how MR may lead us astray.
Overall there were a greater number of injuries noted on the case patients but there were a large number of abnormal findings in the control patients (507 vs. 237 total findings). These included 24 occult vertebral body fractures in the case group compared to 1 in the control group, 54 occult facet joint fractures compared to 26, and 10 ligamentum nuchae tears compared to 4. Cleary there is a great deal of baseline noise that will influence decision making.
The actual utility of MR in blunt cervical trauma is still unknown. It is certainly not the panacea it is currently perceived to be and, if used in the wrong context, may lead our clinical decision making far off course.
While the MR itself has very little risk for the individual patient, the overtreatment that will result from the false positives is concerning. Consider the main intervention performed on the patients with unstable injury found on MR in the Acker’s studies (1,2), anterior cervical decompression and fixation (ACDF). A large review of 1015 stable patients undergoing first-time ACDF for cervical radiculopathy or myelopathy counted death (0.1%) secondary to an esophageal rupture, postoperative dysphagia (9.5%), postoperative hematoma (5.6% – half required surgical interventions), recurrent laryngeal nerve palsy (3.1%), esophageal perforation (0.3%), and worsening of myelopathy secondary to surgery (0.2%) as complications. When done on a population that has no neurologic findings and probably requires no intervention this procedure will surely lead to harm.
Is early intervention a benefit?
Finally, what is the benefit in intervening early in a true occult spinal injury?There have been many retrospective studies that show a benefit to early surgical intervention the most notable of which is the STASCIS trial (9). All of these studies fail to account for the fact that patients who had a delay prior to surgical intervention were sick patients who were not stable enough to go to surgery. It is obvious that this group would fare worse than their healthier counterparts.The only1 RCT comparing early versus delayed surgical intervention in patients with known acute spinal injuries with neurological findings failed to show any benefit of early treatment (10).
If there is no benefit to the more severe symptomatic population it is hard to imagine that early treatment is important in a population with cord compression so subtle it has yet to manifest neurological symptoms. The more important question is whether a patient with an occult unstable cervical injury is cleared from cervical stabilization, will go on develop permanent neurological disability.This is a hard question to answer. In an article by Sanchez et al published in The Journal of Trauma in 2005 (11) the authors examine a protocol in which patients received CT first if they could not be cleared clinically (NEXoid like criteria). If the CT was negative and the patient had neurological findings an MR was ordered. For the patients with no neurological findings and a negative CT, spinal precautions were removed. Among the 2,500 patients who were cleared clinically or following a negative CT none were found to have neurological deterioration at two week follow-up.
In the Ackland study only one patient had what was considered “delayed instability”(2). In the Levi cohort they found only two patients who had serious neurological deterioration and death due to cervical spine clearance and mobilization. However, these occurred after negative 3 view cervical plain film series. These were not cases of CT with recons missing an unstable fracture. Even had they been, it would not justify exposing all of our low risk CT negative patients to the over-diagnosis, over-treatment and complications that are bound to occur with a mandatory MR protocol.
The authors of the Acker study come to the same conclusion. All 5 of the “unstable” cervical injuries requiring surgery had subtle positive findings of CT (2) – they all had severe spondylosis. Because of these findings the authors changed the protocol used to manage blunt trauma patients in their institution so that patients with high clinical suspicion and severe spondylosis on CT will have an MR. Otherwise, neurologically intact patients with a negative cervical CT and persistent midline tenderness have their collar removed without further testing.
Where this leaves us and how we manage the cervical trauma patient is still not entirely clear. What is clear is that using MR to evaluate all patients with persistent midline tenderness after a negative CT will lead to more harm than good. Instead of continuing the quixotic quest for perfection it is important to define a realistic standard of practice and understand that continuing to propagate the “no miss” mentality will further strain an already overburdened system and hurt more patients than it helps.
- Outcomes at 12 Months After Early Magnetic Resonance Imaging in Acute Trauma Patients With Persistent Midline Cervical Tenderness and Negative Computed Tomography.SPINE Volume 38, Number 13, pp 1068–1081, 2013.
- Ackland HM,Cameron PA,Varma DK, et al.Cervical spine magnetic resonance imaging in alert, neurologically intact trauma patients with persistent midline tenderness and negative computed tomography results.Ann Emerg Med 2011 ; 58 : 521 – 30
- Levi et al. Neurologic Deterioration Secondary to Unrecognized Spinal Instability Following Trauma- A Multicenter Study. SPINE Volume 31, Number 4, pp 451-458, 2006
- Stephen G. Pauker, M.D., and Jerome P. Kassirer, M.D. The Threshold Approach to Clinical Decision Making N Engl J Med 1980; 302:1109-1117May 15, 1980
- Rhin, J et al. Using Magnetic Resonance Imaging to Accurately Assess Injury to the Posterior Ligamentous Complex of the Spine: A Prospective Comparison of the Surgeon and Radiologist. J. Neurosurgery Spine 12;391-396
- Rhin, JA et al. Assessment of the Posterior Ligamentous Complex Following Acute Cervical Trauma. J Bone Joint Surg Am. 2010 Mar;92(3):583-9
- Anderson, S et al Are there cervical spine findings at MR imaging that are specific to acute symptomatic whiplash injury? A prospective controlled study with four experienced blinded readers. Radiology. 2012 Feb;262(2):567-75. doi: 10.1148/radiol.11102115. Epub 2011 Dec 20.
- Fountas, Kostas N. MD, PhD*; Kapsalaki, Eftychia Z. MD*; Nikolakakos, Leonidas G. MD; Smisson, Hugh F. MD, FACS; Johnston, Kim W. MD, FACS; Grigorian, Arthur A. MD, PhD; Lee, Gregory P. PhD; Robinson, Joe S. Jr MD, FACS Anterior Cervical Discectomy and Fusion Associated Complications Spine:1 October 2007 – Volume 32 – Issue 21 – pp 2310-2317
- Fehlings MG, Vaccaro A, Wilson J, et al. Early versus delayed decompression for traumatic cervical spinal cord injury: results of the Surgical Timing in Acute Spinal Cord Injury Study (STASCIS). PLoS ONE 2012;7:e32037.
- A. R. Vaccaro, R. J. Daugherty, T. P. Sheehan et al., “Neurologic outcome of early versus late surgery for cervical spinal cord injury,” Spine, vol. 22, no. 22, pp. 2609–2613, 1997.
- Sanchez, Barry, et al. “Cervical spine clearance in blunt trauma: evaluation of a computed tomography-based protocol.” The Journal of Trauma and Acute Care Surgery 59.1 (2005): 179-183.
Thanks for all your help!
- Peer reviewed by Brent Thoma
- Special thanks to Rebecca Talmud