In both ACDF and ACCF techniques, native structures are removed and reconstructed with grafts. One truism in life is that we can rarely recreate nature’s work. So preserving as much of the human anatomy as possible is naturally advantageous. Also, the more we do in the human body, the more the undesirable collateral effects (such as bleeding, scarring, weakening, alteration of mechanics). If the nerve compression is only at the disk space, then proponents of ACDF state that it is better to just remove the disks and reconstruct the disk spaces. This approach leaves the nonpathologic tissues alone because invariably they cannot be reconstructed the way nature intended them to be, and this potentially may have unintended consequences on adjacent structures. ACDF may be easier than ACCF because extensive bone removal is not necessary, and there is less bleeding (from bony surfaces). The bone work is limited to foraminotomies and preparation of the end plates, so bony bleeding should be less than during a corpectomy. However, this last argument goes both ways since ACDF requires more time spent working in confined and often collapsed disk spaces (Fig. 136.1).

There is also the argument that multilevel ACDF is a more stable construct than ACCF. An ACCF has only the two ends of a long strut graft where loading transpires from the upper vertebra of the cervical spine to lower vertebrae. So the ends of the graft are subject to more destabilizing forces (translational and subsidence) because of the long lever arm of the strut graft or cage and because it spans multiple motion segments. This is as opposed to individual intervertebral grafts, which must resist only single-motion segment loading. Moreover, with anterior cervical plates and screws, each vertebra can be segmentally instrumented to create a much more stable construct (1, 2 and 3). Segmental instrumentation is not possible after ACCF.

Because ACCF consists of a single long graft, there may be a tendency to place it under either too much or too little tension. When the tension during graft placement is too high, subsidence of the graft into the vertebrae can occur. When the tension is too low, dislodgment can occur. In the literature, the prevalence of graft slippage after ACDF surgery is 1% to 2%, while that of a strut graft after ACCF is 6% to 29%, with the inferior end of the graft dislodging most often (4,5). The addition of plate and screw instrumentation reduces this rate, but more so in multilevel ACDF because each segment (vertebrae) can be instrumented (Fig. 136.2) unlike in ACCF where only the end vertebrae can be instrumented (Fig. 136.3).

Sagittal alignment of the cervical spine (lordosis) is important to restore and maintain because it has consequences on the health of the muscles that stabilize the spine and the remaining disks and also prevents any draping of the spinal cord on the dorsal surfaces of the vertebrae in a kyphotic spine (6). The fact that the cervical spine has multiple disks is advantageous because each disk
contributes to the cervical lordosis incrementally, creating a smooth gradual curve. It follows then that if multilevel disk surgery is performed, reconstructing each disk segment separately with grafts of the appropriate height or lordosis will impart a more natural and stable sagittal alignment compared to a long straight strut graft. An analogy may be drawn to segmental spinal instrumentation versus a long Harrington rod instrumentation in the lumbar spine. Though proponents of ACCF will state that cervical lordosis can be reliably restored and maintained, what they cannot promise is that each segment is distracted equally and will thus contribute incrementally to the overall lordosis
as it naturally should. With a single graft spanning multiple disk spaces, the most flexible or unstable segments will contribute most to the lordosis, which may place undue stress and stretch at that segment. Therefore, indirect decompression by distraction may not occur equally in all the segments. This can be better controlled during ACDF by the proper choice of graft size at each segment. This concept is illustrated in Figure 136.4 with the spine represented schematically by a spring.