The aim of this in vitro study was to determine the influence of simulated postoperative neck movements on the stabilizing effect and subsidence of four different anterior cervical interbody fusion devices. Emphasis was placed on the relation between subsidence and spinal stability. The flexibility of 24 human cervical spine specimens was tested before and directly after being stabilized with a WING, BAK/C, AcroMed I/F cage, or with bone cement in standard flexibility tests under 50 N axial preload. Thereafter, 700 pure moment loading cycles (+/- 2 Nm) were applied in randomized directions to simulate physiological neck movements. Additional flexibility tests in combination with measurements of the subsidence depth were conducted after 50, 100, 200, 300, 500, and 700 loading cycles. In all four groups, simulated postoperative neck movements caused an increase of the range of motion (ROM) ranging from 0.4 to 3.1 degrees and of the neutral zone from 0.1 to 4.2 degrees. This increase in flexibility was most distinct in extension followed by flexion, lateral bending, and axial rotation. After cyclic loading, ROM tended to be lower in the group fitted with AcroMed cages (3.3 degrees in right lateral bending, 3.5 degrees in left axial rotation, 7.8 degrees in flexion, 8.3 degrees in extension) and in the group in which bone cement was applied (5.4 degrees, 2.5 degrees, 7.4 degrees, and 8.8 degrees, respectively) than in those fixed with the WING (6.3 degrees, 5.4 degrees, 9.7 degrees, and 6.9 degrees, respectively) and BAK cages (6.2 degrees, 4.5 degrees, 10.2 degrees, and 11.6 degrees, respectively). Simulated repeated neck movements not only caused an increase of the flexibility but also subsidence of the implants into the adjacent vertebrae. The relation between flexibility increase and subsidence seemed to depend on the implant design: subsiding BAK/C cages partially supported stability whereas subsiding WING cages and AcroMed cages did not.