Dynamic computerized tomography (DCT) has been accepted to be the standard diagnostic method of atlanto-axial rotatory subluxation (AARS) although its reliability and reproducibility has not been shown yet. The purpose of this study was to evaluate the intraobserver reproducibility and interobserver reliability of DCT. Standard DCT scans of 18 patients with acute torticollis and 12 normal subjects were examined two times in between a time interval of 1 month by three specialists and a last-year resident to define any existing AARS. The interobserver reliability kappa coefficient was -0.015 (poor) for the first examination and 0.327(fair) for the second one. The intraobserver reproducibility kappa coefficients were 0.135 (slight), -0.204 (poor), 1.00 (almost perfect), and 0.474 (moderate) respectively. It was found that DCT has a poor reliability and reproducibility in diagnosing AARS in patients with acute torticollis. Therefore, its routine use is not cost effective in patients with acute torticollis where the deformity usually resolves by a simple cervical mobilization.