Comparing NEXUS and Canadian C-Spine decision rules for determining the need for cervical spine radiography

Abstract

In the past 5 years, 2 separate groups of investigators developed clinical decision rules to help the clinician determine which blunt trauma patients require radiographic evaluation of the cervical spine. These decision rules apply to a specific target population: alert, stable patients without neurologic deficit. In this issue of Annals, the Canadian C-Spine Research Group, which developed one of the decision rules, evaluated the “accuracy and reliability” of the National Emergency X-Radiography Utilization Study (NEXUS) decision rule by applying the NEXUS low-risk criteria to a database derived from 10 Canadian emergency departments.1 In essence, the Canadian C-Spine investigators retrospectively applied the NEXUS criteria to the database of the original Canadian C-Spine study. In an accompanying commentary, representatives from the NEXUS group present their perspective on this comparison and discuss the similarities and differences between the 2 decision rules.2 Previous studies by the Canadian C-Spine and NEXUS groups3,4 report that application of either of the decision rules can result in a modest reduction in the number of unnecessary radiographic studies. The most important question raised by the new Canadian C-Spine study, however, is whether these new findings should cause clinicians to question whether the NEXUS criteria can produce results similar to the original NEXUS study in different practice settings. The Canadian CSpine study had 2 conclusions. First, they state that the NEXUS criteria are less sensitive than previously reported. In the Canadian C-Spine study, the NEXUS criteria were reported to have a sensitivity of 92.7% to detect a clinically significant injury as opposed to the 99.6% sensitivity reported in the original NEXUS study. There were 11 patients in their data set who Canadian C-Spine investigators defined as low risk by NEXUS criteria but had clinically important cervical spine fractures. However, there are reasons to question whether all of these patients were accurately classified as low risk. Careful review of the description of these patients suggests that many emergency physicians would classify at least some of these patients as not meeting the low-risk NEXUS criteria. Because the Canadian C-Spine data set had already been collected, the Canadian C-Spine investigators used approximations or surrogate criteria rather than original NEXUS criteria. For example, NEXUS investigators state that any evidence of intoxication is a sufficient reason for exclusion from low-risk classification. They describe a number of specific historical indicators or clinical/laboratory findings that would define a patient as intoxicated: “Patients should be considered intoxicated if they have either of the following: (1) a recent history by the patient or an observer of intoxication or intoxicating ingestion; or (2) evidence of intoxication on physical examination, such as odor of alcohol, slurred speech or ataxia, dysmetria, or other cerebellar findings, or any behavior consistent with intoxication. Patients may also be considered to be intoxicated if tests of bodily secretions are positive for drugs (including but not limited to alcohol) that affect level of alertness.”5 The Canadian study1 is more restrictive; it used “unreliable findings due to drugs or alcohol” as their definition for intoxication. Another example is use of the term “distracting injury.” NEXUS defined distracting injury by giving examples of specific injuries: “No precise definition for distracting injury is possible. This includes any condition thought by the clinician to be producing pain sufficient to distract the patients from a second (neck) injury. Examples may include, but are not limited to, the following: (1) a long bone fracture; (2) a visceral injury requiring surgical consultation; (3) a large lacerT R A U M A / E D I T O R I A L