False-negative plain cervical spine x-rays in blunt trauma.

Although considered very accurate, false-negative plain cervical radiographs of blunt trauma patients will occur with potentially devastating complications. We sought to define the population of patients who fall into this category and the overall accuracy of adequate three-view cervical spine radiography in the blunt trauma population. A retrospective search was carried out of blunt trauma patients entered into our trauma registry. All patients with the ICD-9 codes indicating cervical spine injury with a negative three-view cervical spine radiograph reading had their charts and radiographs reviewed. Institutional statistics for blunt cervical trauma evaluation and injury were obtained from the trauma registry. Fifty-eight of 936 blunt trauma patients (6.2%) were diagnosed with cervical spine injury over the 9-month study period. Of 649 patients with adequate three-view plain radiographs, three patients were identified with negative plain radiographs and significant cervical spine injury, a false-negative rate of 0.5 per cent. Sensitivity was 90.3 per cent, specificity was 96.3 per cent, positive predictive value was 54.9 per cent, and negative predictive value was 99.5 per cent. Three-view plain radiograph series of the cervical spine remains a highly sensitive and specific test for cervical spine injury following blunt trauma. However, the fact that we identified three patients with significant fractures after negative plain radiographs suggests that serious consideration of computed tomography must be applied in treating symptomatic, high-risk blunt trauma patients when plain radiographs do not reveal an injury.