Study Design. In vitro biomechanical testing was performed on 12 cadaveric human lumbar spines. Objective. To determine the initial dislocation resistance, as quantified by the pullout force of three different cage designs. Summary of Background Data. Interbody cage devices frequently are used as stand-alone cages in the surgical treatment of degenerative conditions in the lumbar spine. In contrast to the wide clinical acceptance of interbody fusion cages, there are only a few biomechanical studies of posterior pullout trials. Methods. Cylindrical threaded cages (Ray TFC Surgical Dynamics), bullet-shaped cages (Stryker), and newly designed rectangular titanium cages with an endplate anchorage device (Marquardt) were used for posterior interbody implants. For each device, the pullout test was performed in four specimens on both sides (L3-L4). Results. In the pullout test, the Stryker cages required a median pullout force of 130 N (minimum, 100 N; maximum, 220 N), as compared with the higher pullout force of the Marquardt cages (median, 605 N; minimum, 450 N; maximum, 680 N), and the Ray cages (median, 946 N; minimum, 125 N; maximum, 2230 N). Conclusions. Differences in pullout resistance were noted depending on the cage design, A cage design with threads or a hook device provides superior stability, as compared with ridges. The initial pullout resistance was highest for the Ray cages and lowest for the Stryker cages.