Background: Prompt identification of cervical spine injuries has been a critical issue in trauma management. In 1998, the authors developed a new protocol to evaluate cervical spines in blunt trauma. This protocol relies on clinical clearance for appropriate patients and helical computed tomography instead of plain radiographs for patients who cannot be clinically cleared. The authors then prospectively collected data on all cervical spine evaluations to assess the sensitivity and specificity of their approach.
Methods: Any patient without clinical evidence of neurologic injury, alcohol or drug intoxication, or distracting injury underwent cervical spine evaluation by clinical examination. Patients who did not meet these criteria underwent helical computed tomographic scanning of the entire cervical spine. For patients who had neurologic deficits, a magnetic resonance image was obtained. If the patient was not evaluable secondary to coma, the computed tomographic scan was without abnormality, and the patient was moving all four extremities at arrival in the emergency department, the cervical spine was cleared, and spinal precautions were removed. Data were collected for all patients admitted to Santa Barbara Cottage Hospital trauma service between 1999 and 2002. The authors selected for analysis patients with blunt trauma and further identified those with closed head injuries (Glasgow Coma Scale score < 15 and loss of consciousness). In addition, all blunt cervical spine injuries were reviewed.
Results: During the period of study, 2,854 trauma patients were admitted, of whom 2,603 (91%) had blunt trauma. Of these, 1,462 (56%) had closed head injuries. One hundred patients (7% of patients admitted for blunt trauma) had cervical spine or spinal cord injuries, of which 99 were identified by the authors' protocol. Only one injury was not appreciated in a patient with syringomyelia. Fifteen percent of patients with spinal cord injury had no radiographic abnormality; all of these patients presented with neurologic deficits. The sensitivity for detecting cervical spine injury was thus 99%, and the specificity was 100%. The risk of missing a cervical spine injury in these blunt trauma patients was 0.04%. The authors missed no spine injuries in patients with head injuries.
Conclusion: The use of the authors' protocol resulted in excellent sensitivity and specificity in detecting cervical spine injuries. In addition, it allowed early removal of spinal precautions.