Virtually anywhere in the skeleton, there are three influential variables for bone healing: (a) type of graft, (b) host factors, and (c) stability of the biomechanical environment. Risk factors for development of ventral cervical pseudarthroses can be categorized into these three areas as well.

There has been significant evolution of type of graft material used for ACDF procedures over the decades. Initial descriptions by Robinson and Smith (1) described a tricortical iliac crest type of graft while Cloward (2) utilized a dowel-type corticocancellous graft. Autograft was the gold standard for many years as these operative techniques became common, as harvesting from the patient’s iliac crest provided graft strength as well as the host cell and marrow factors beneficial in generating bone healing.

The inherent morbidity of iliac crest bone graft harvest including pain, fracture, and lateral femoral cutaneous nerve injury led to exploration of the use of allograft to achieve arthrodesis. Early comparative studies suggested allograft union rates were more or less equivalent to autograft in a favorable environment, specifically for one-level ACDF procedures (3).

With the advent of ventral plating in the 1990s, postoperative stability was improved, and reports of high success rates using allograft were published. Samartzis et al. (4), in a retrospective review, found no statistical difference in allograft plus plating versus autograft plus plating for fusion or outcome in multilevel ACDF patients. While there is a lack of prospective studies comparing allograft to autograft with respect to arthrodesis rates and clinical outcomes, generally, the literature has demonstrated comparable success of allograft with the use of modern instrumentation, taking into account other factors influencing fusion to be discussed below.

There are even less rigorous data regarding differing types of structural allo- or autograft such as pure cortical versus corticocancellous versus dense cancellous options. For autograft, options include tricortical iliac crest grafts or solid cortical fibula grafts. Machined allografts present more choices such as cortical rings with a large open center, cortical rings with a minimal open center, iliac crest corticocancellous grafts, as well as composite grafts. Allograft processing techniques such as freeze drying, fresh freezing, and irradiation represent another variable potentially affecting fusion success. There is very little literature regarding the relative success of types of allograft available, although one recent study reported on poor radiographic results and significant resorption in a series of dense cancellous allografts used for ACDF procedures (5).

Synthetic interbody spacers such as titanium or PEEK (polyetheretherketone) cages packed with graft material have also been utilized and are reported in the literature (6,7). Most of these are retrospective case series without comparative groups. Nevertheless, these reports suggest a relatively high rate of arthrodesis with several types of graft filler utilized with the interbody spacer. This probably speaks for the capability of the anterior column in the cervical spine to heal with a variety of operative techniques, as well as the fact that successful clinical outcomes do not always require successful radiographic arthrodesis.

In more recent years, bone morphogenetic protein (BMP) has been utilized in some series for ACDF procedures though this has not been approved at this time by the U.S. Food and Drug Administration (FDA). Perioperative swelling and airway issues have slowed widespread use of BMP in the ventral cervical spine, and there are few data supporting its relative efficacy achieving arthrodesis when compared to autograft or allograft (8,9).

The attributes of the host in promoting successful fusion range from systemic physiology to the local operative environment. Smoking is considered detrimental to bone healing, and this is believed to be true for ventral cervical constructs also. In a retrospective review, Hilibrand et al. (10) noted a higher pseudarthrosis rate in smokers compared to nonsmokers undergoing multilevel ACDF procedures without instrumentation. This study and others to date are retrospective in nature, so the effect of smoking, if any, on subpopulations of fusion patients (e.g., single-level vs. multilevel) awaits further study.

One of the most significant factors affecting pseudarthrosis rate is the number of levels operated on using interbody fusion methods (Table 108.1). Many studies have documented that two-level ACDF procedures have a higher nonunion rate than single-level (11), and three-level procedures are yet even higher than a two-level interbody fusion procedure (12). Using iliac crest autograft without instrumentation, Bohlman et al.’s (13) series documented an 11% pseudarthrosis rate for single-level and 27% for two-level ACDFs. Emery et al. (14), again without instrumentation, described a 45% nonunion rate for three-level stand-alone ACDF procedures. Ventral instrumentation has improved these statistics, but the trend for more levels resulting in higher pseudarthrosis rates continues to hold true as is discussed below.

The propensity for multilevel ACDF to have higher pseudarthrosis rates could be related to the biomechanical environment, but also because of the increased number of bone interfaces, the host must heal as the number of operative levels increases. Hilibrand et al. (15) documented a higher fusion rate in corpectomy and strut graft patients versus multilevel interbody fusion patients. With use of instrumentation, however, Wang et al. (16) found no difference in single-level corpectomy plus strut graft procedures versus two-level discectomy procedures. It is interesting to note that in three-level ACDF procedures (which typically involve C4-C5, C5-C6, and C6-C7), it is usually the C6-C7 that results in pseudarthrosis, perhaps due to the longer lever arm above and the proximity of the stiffer thoracic spine below (Fig. 108.1).