OBJECTIVE To determine whether flexion-extension cervical spine radiography (FECSR) is abnormal in children who have sustained blunt cervical spine injury (CSI) when standard cervical spine radiography (SCSR) demonstrates no acute abnormalities. METHODS This was a blinded radiographic review of 129 patients < or = 16 years of age evaluated at an academic pediatric trauma center during July 1990-March 1996. All patients had SCSR (anteroposterior/lateral views) and FECSR performed for a trauma-related event within seven days of injury. RESULTS Of 46 patients without acute abnormalities on SCSR, one patient (with final clinical diagnosis of "no CSI") had acute abnormalities on FECSR (95% CI = 0.06% to 11.5%). Of 50 patients with isolated loss of lordosis on SCSR, no patient had acute abnormalities on FECSR (95% CI = 0% to 5.8%). The FECSR review revealed no acute abnormalities in 75 of 83 patients (90.4%) with suspicious findings for CSI viewed on SCSR (95% CI = 81.9% to 95.7%). Complications during FECSR were noted in one patient with transient paresthesias (0.8%) (95% CI = 0.02% to 4.2%). CONCLUSIONS In children who underwent acute radiographic evaluation of blunt cervical spine trauma, FECSR was unlikely to be abnormal when no acute abnormality or isolated loss of lordosis was evident on SCSR. In a subset of patients with suspicious findings for occult CSI on SCSR, FECSR was useful in ruling out ligamentous instability in the acute, posttrauma setting.