Aims: In this study, the subsidence of different interbody fusion devices was investigated. Hereby, the influence of different designs as well as of the preparation technique was evaluated. Methods: 3 common cervical interbody fusion devices (BAK, Novus and WING) underwent axial compression testing with 4000 cycles in a bovine spine model. The vertebral bodies were prepared in 3 different ways, taking away 0, 1 and 2 mm of the end-plate. So each fusion device was tested in each preparation group in 5 vertebrae. Every 1000 cycles, the subsidance was measured. Results: Taking away 1 and 2 mm of the endplate resulted in a strong increase of the subsidance compared to the situation with intact end-plate. In addition, the design of the interbody device had an influence onto subsidance: In case of intact endplates, the cages with rectangular supporting areas resisted better to axial compression than the cylindrical implant. When the cortical bone of the endplate was taken away, all three implants showed similar subsidance curves. Conclusions: Implants with plane supports seem to provide better stability against subsidance than cylindrical implants. During preparation, the cortical structure of the endplate should be taken care of, especially in the zone, where the implant has its bearing areas