METHODS Five different reconstructions of the atlantoaxial complex were biomechanically compared in vitro in a nondestructive test. OBJECTIVE To determine whether non-bone graft-dependent one-point fixation affords stability levels equivalent to three-point reconstructions. BACKGROUND Previous investigations have demonstrated that three-point fixation, using bilateral transarticular screws in combination with posterior wiring, provide the most effective resistance to minimize motion around C1-C2. However, placement of transarticular screws is technically demanding. Posterior wiring techniques affording one-point fixation have failure rates of approximately 15%, with failure considered to be secondary to structural bone graft failures. One-point, non-bone graft-dependent fixations have not been tested. METHODS Eight human cervical specimens, C0-C3 were loaded nondestructively. Unconstrained three-dimensional segmental motion was measured. The reconstructions tested were two one-point fixations, one two-point fixation, and two three-point fixations. RESULTS Under axial rotation two and three-point reconstructions provided better stiffness than the one-point reconstructions (P < 0.05). During flexion-extension, higher stiffness levels were observed in one- and three-point fixations when compared with the intact spine (P < 0.05). In lateral bending no significant differences were observed among the six groups, although the trend was that reconstructions including transarticular screws provided greater stability than one-point fixations. CONCLUSIONS The current findings substantiate the use of three-point fixation as the treatment of choice for C1-C2 instability. [l: atlantoaxial fixation, biomechanics, cervical spine, instability, spinal instrumentation, transarticular screws]