Diagnostics: C-Spine Rules


Prior to the creation of clinical decision rules for cervical spine imaging, management was based primarily on physician gestalt and an overall “play it safe” mentality. This essentially amounted to nearly every patient with blunt trauma undergoing some type of cervical spine imaging, especially in the setting of a presumed head injury. In the United States, >13 million patients are evaluated each year for possible cervical spine injuries after trauma. However, only about 0.3% of these patients are found to have significant injury. It’s hard to blame people for being trigger happy on cervical spine imaging though – the tests in isolation are relatively quick and cheap, and the consequence for missing occult injury can be catastrophic. But when all those tests get added together, it’s estimated that over $180 million dollars gets spent each year. [1]


Enter the National Emergency X-Radiography Utilization Study, or NEXUS. A group of physicians at UCLA in 1992 developed a study to test the hypothesis that there are certain low risk criteria that can be applied to blunt trauma patients to determine when cervical spine imaging is not needed. [2] They had the treating clinicians fill out forms consisting of elements of history and physical examination, prehospital treatment, and estimated likelihood of cervical spine injury for all blunt trauma patients for whom the clinician had decided to order cervical spine imaging. Overall, 27 of the 974 patients in the study were found to have a cervical spine fracture. All of the patients with a fracture had at least one of the following:

  • Midline neck tenderness

  • Evidence of intoxication

  • Altered level of alertness

  • Severely painful injury elsewhere

Negative films would have been reduced by 12.5% and no fractures would have been missed if the criteria had been applied to the study population. These results led to the the creation of the NEXUS low risk criteria.

8 years passed until 2000 when the same authors published a study validating the NEXUS low-risk criteria. [3] In the interim, and astutely pointed out by the authors of the Canadian C-Spine Rule, NEXUS decided to add in “no focal neurologic deficit” to their criteria, even though it wasn’t technically a part of the original study and no supporting evidence was offered to justify its inclusion. The validation study included 34,069 patients with blunt trauma who underwent cervical spine imaging. Patients with penetrating trauma or that underwent imaging for reasons unrelated to trauma were excluded. In total, 818 of the 34,069 patients had radiographically documented cervical spine injury.  

The study also distinguished radiographically-documented cervical spine injuries that were not clinically significant. These included: spinous process fractures, simple wedge compression fracture without loss of 25% or more of vertebral body height, isolated avulsion, Type 1 odontoid fracture, end plate fracture, osteophyte fracture, injury to trabecular bone, and transverse process fracture. With these excluded, the sensitivity of the tool increased to 99.6%. Ultimately, 8 patients were found to have cervical spine injury despite a negative result by the NEXUS low risk criteria, and would have been missed by the tool. Only 2 of those patients were deemed to have clinically significant injuries.


In response to some of the perceived limitations of the NEXUS low risk criteria, a group from Canada published and validated their own clinical decision rule in a 2001 paper. [4] These limitations included the low specificity of NEXUS, as well as the subjective nature of some of the criteria (i.e. painful distracting injury and presence of intoxication). The Canadian study enrolled a convenience sample of 8,924 adult patients who presented to the ED with blunt trauma to the head/neck, stable vital signs and a GCS of 15. It is important to recognize the more robust exclusion criteria in this study, as these restrictions weren’t included in NEXUS. Another interesting component of the author’s methods was that not all enrolled patients had cervical spine imaging completed. The imaging was left to the discretion of the treating physician, and if the patient was sent home without imaging, they were placed in a 14-day follow-up protocol. This included a phone call from a registered nurse and the patient was labeled as having no significant cervical spine injury if they met all of the following criteria: neck pain rates as none or mild, restriction of neck movement rates as none or mild, use of cervical collar not require and neck injury that had not prevented return to usual occupational activities.

As in NEXUS, the Canadian study also distinguished between clinically significant injuries and not clinically significant injuries. They defined injuries to not be clinically significant if they did not require stabilized treatment or specialized follow up, and included one of the following four injuries: 1) isolated avulsion fracture of an osteophyte; 2) isolated fracture of transverse process not involving facet joint; 3) isolated fracture of spinous process not involving the lamina; or 4) simple compression fracture with <25% loss of height.

Ultimately, the Canadian study included 151/8,924 patients with a clinically significant cervical spine injury and derived the Canadian C-Spine Rule based on variables that were found to be highly predictive of cervical spine injuries. They then applied their decision rule to the same study population, and found a 100% sensitivity and 42.5% specificity for detecting clinically important cervical spine injuries. Had the decision rule been applied, the overall radiography ordering rate would have been reduced by 15.5%.

Comparison/Further Studies

Since the creation of both the NEXUS and Canadian clinical decision rules, only one study has directly compared the accuracy of the rules against one another. This study was published in 2003 by the same authors of the Canadian C-Spine Rule. [1] They applied the same methods as their validation study in 2001 to a population of 8,283 blunt trauma patients, and ultimately had 169 patients (2%) with clinically important cervical spine injuries.

So in this study, the Canadian C-Spine Rule performed better than NEXUS. One interesting finding, however, was for 845 (10.2%) of the patients, physicians did not evaluate range of motion as required by the CCR algorithm. This may suggest that there is some physician discomfort and decreased compliance when it comes to applying the CCR rule, perhaps due to concern that range of motion testing may exacerbate an injury. An additional 2012 systematic review was published in the Canadian Medical Association Journal to investigate the diagnostic accuracy of the two rules. [5] However, all of the studies included in the review besides the one discussed above were validation studies and did not directly compare the two rules against each other. The review found a combined sensitivity of NEXUS to be 83-100% and Canadian to be 90-100% and the authors concluded that the Canadian C-Spine rule appears to have a better diagnostic accuracy.

Post by Daniel Gawron, MD

Peer Editing by Kevin Randolph, MD and Ryan LaFollette, MD