According to the National Spinal Cord Injury Statistical Center,<sup>14National Spinal Cord Injury Statistical Center (University of Alabama at Birmingham): nscisc.htm. Spinal cord injury facts and figures, Dec 1998.

Cliquez ici pour aller à la section Références</sup> data from state trauma registries suggest that approximately 10,000 instances of spinal cord injury occur in the United States annually. One of every ten cases of spinal cord injury are fatal, usually attributed to high cervical involvement leading to ventilatory failure or cardiopulmonary arrest. One third to one half of cases of spinal cord injury result in complete paralysis, and one third to one half, in incomplete paralysis. Approximately 200,000 Americans are living with a disability attributable to spinal cord injury. The annual cost of direct medical care for patients with spinal cord injury in the United States is $10 billion.<sup>14National Spinal Cord Injury Statistical Center (University of Alabama at Birmingham): nscisc.htm. Spinal cord injury facts and figures, Dec 1998.

Cliquez ici pour aller à la section Références</sup>

Motor vehicles crashes account for 37% of instances of spinal cord injury, crimes of violence for 26%, falls for 24%, sports mishaps for 7%, and other conditions or illnesses for 6%. <sup>14National Spinal Cord Injury Statistical Center (University of Alabama at Birmingham): nscisc.htm. Spinal cord injury facts and figures, Dec 1998.

Cliquez ici pour aller à la section Références</sup> Because of anatomic proximity, the penultimate cause or mechanism of most spinal cord injuries is spinal fracture or dislocation. One half of spine fractures are cervical, one sixth thoracic, and one third lumbosacral.<sup>14National Spinal Cord Injury Statistical Center (University of Alabama at Birmingham): nscisc.htm. Spinal cord injury facts and figures, Dec 1998.

Cliquez ici pour aller à la section Références</sup> To ameliorate the consequences of spinal cord injury, including permanent neurologic deficit and death, emergency physicians must use an approach with high sensitivity for the detection of spinal fracture. As most spinal fractures are diagnosed definitively on the basis of the results of imaging studies, integral to development of a diagnostic approach for spinal fracture is formulation of a scientifically based decision strategy for the ordering of such imaging studies. The diagnostic approach must have a more sophisticated basis than an a priori decision to sacrifice specificity for sensitivity, which would be harmful to patients (unnecessary discomfort of prolonged immobilization and unnecessary radiation exposure) and economically wasteful.

These authors sought to formulate a decision strategy for the ordering of imaging studies of the spine by reviewing relevant medical literature by means of an English language, human subjects–limited MEDLINE search from 1966 to February 1998. By usual MEDLINE search practice, various combinations of text words (in this case, trauma, spine injury, radiography, and decision analysis) were used to develop search criteria. The 32 articles yielded by the MEDLINE search were reviewed by the authors, as were 5 additional articles identified from the references of the MEDLINE-identified articles. A focus on decision analysis for use of plain radiography, CT, or MRI for detection of acute spine fracture was used as criterion for article inclusion. An article's focus on nonacute fracture or chronic neck or back pain was criterion for exclusion. The 26 articles the data for which is abstracted in detail in Table 1 to Table 6 describe prospective and retrospective cohort and uncontrolled prospective studies. The methodologic quality of the articles abstracted is therefore of level II evidence. These authors found no randomized controlled trials of strategies for diagnosis of acute spinal fracture; none of the articles abstracted yielded level I evidence. Case reports (i.e., articles with level III evidence) were excluded.

Table 1 lists all cervical spine radiographic data from the 13 reviewed articles. Table 2 summarizes data abstracted from the articles regarding the decision to order cervical spine plain radiography. Based on review of the articles abstracted, the authors conclude that the possibility of spinal fracture and therefore spinal cord injury should be taken seriously when victims of motor vehicles crashes, violent crime, falls, and other mechanisms known to cause spinal cord injury present to the emergency department (ED). Detection and treatment of life-threatening injuries would not be delayed by attempts to diagnose spinal fracture definitively, however. The articles abstracted suggest that the following approach to diagnosis of spinal fracture be used after adequate attention to immediate and short-term threat to life (Figure 1).

Table 3 presents data abstracted from the articles regarding the decision to order computerized tomography or magnetic resonance imaging to evaluate patients with possible spinal fracture or spinal cord injury. Based on review of the articles abstracted, the authors conclude that data regarding the decision to utilize CT and MRI in this circumstance is derived almost exclusively from retrospective analyses of uncontrolled actual institutional experiences with CT and MRI. Therefore, rather than make specific recommendations regarding use of CT and MRI for evaluation of spinal fracture or spinal cord injury, the authors suggest that it is most appropriate to limit recommendations to the summary of data in Table 4.