Abstract
Dear Editor,
Based on findings from a previous review of the literature (Dunham et al, 2008), the authors applied a “conservative” assumption of having a 2.5% probability of an “unstable” cervical spine after a negative CT scan of the cervical spine. The authors continue their assumptions by stating that in those patients without an additional MRI scan, the attributable risk of a “tetraplegia” is equal to the probability of having an “unstable spine”, namely 2.5%. This means that 1 out of each 40 surviving comatose, blunt trauma patients with extremity movement and a negative CT scan of the cervical spine would develop a tetraplegia. After reading this, our first question was: “How did the authors come to such an impressively low number needed to harm?” The answer can partly be explained by use of ambiguous definitions and the literature review the authors based their assumptions on (Dunham et al, 2008).
In their review, Dunham and colleagues pooled 2,216 patients from 14 original studies evaluating the number of false-negative cervical spine injuries with non-MRI imaging techniques in the obtunded and comatose blunt trauma population (Dunham et al, 2008). Interestingly, in seven of the included studies, there was no direct comparison between the diagnostic accuracy of the CT and of the MRI. Taking into consideration only those studies with both CT and MRI scans reduces the number of pooled patients to 486. Thirty-four patients of this latter group (7%) were treated with a cervical collar and two (0.4%) underwent surgery. Most of these patients had ligamentous injuries not detected by CT. There was no reported case of tetraplegia occurring after clearance of the cervical spine with CT in any of the studies.
Similar to the exclusion of clinically unimportant fractures in the Canadian C-spine and NEXUS clearance criteria (e.g. transverse process fractures not involving a facet joint, spinous process fractures not involving the lamina) (Hoffman et al., 2000; Stiell et al. 2001), we should also consider the clinical (un)importance of isolated ligamentous injuries as discerned on MRI. Although literature does not provide conclusive data, it is highly unlikely that isolated ligamentous injuries result in serious spinal instability and, as a consequence, neurological sequelae. It may well have been that the indication for cervical collar treatment in the described 34 patients was based on the “defensive medicine” principle (Brown et al., 2010). This would leave only two cases (0.4%) who “required” operation, and more detailed information about these two patients would be of interest (Como, 2010).
In conclusion, given the use of ambiguous definitions of spinal stability and an inaccurate appraisal of previous literature findings, it is likely that the authors overestimated the likelihood of tetraplegia in their probabilistic computer simulation study. Interestingly, a lower likelihood of tetraplegia after clearance of cervical spine with use of CT in obtunded blunt trauma patients would even strengthen the authors' recommendation of not performing an additional MRI for the clearance of the cervical spine in this group of patients.