Prospective Controlled Study of Rectangular Titanium Cage Fusion Compared With Iliac Crest Autograft Fusion in Anterior Cervical Discectomy

Dimitris Zevgaridis, M.D., Claudius Thomé, M.D., Joachim K. Krauss, M.D., Department of Neurosurgery, University Hospital Mannheim, Mannheim, Germany.

Neurosurg Focus. 2002;12(1) 

In This Article


The anterior approach to the cervical spine has become an accepted route for treatment of a large number of cervical spine disorders including spondylosis, herniated discs, fractures, and neoplastic lesions. Since its introduction by Cloward and by Smith and Robinson, several technical modifications for ACD have been developed.[12,13,33] No consensus regarding the optimum technique, however, has yet been established. Controversial debates in cervical disc surgery are focused on issues such as the use of either autograft or allograft material to maintain disc height, the role of plate-related stabilization, or whether any graft or plate is necessary at all.[3,11,24]

Many surgeons have believed that decompression alone will suffice.[1,6,15,34,37,48] It has been accepted that the surgically treated disc level would undergo settling and some kyphotic angulation and, in the majority of patients, would then fuse. More than 20 years ago, Wilson and Cambell[48] reported good or excellent outcome in 85% of patients after ACD without fusion. Their follow-up period of 6 months, however, was relatively short. Martins[25] reported clinical success in 92% of the patients after ACD with or without fusion at 10-month follow up. In the group undergoing ACD alone, a relevant kyphotic deformity developed in 10% of patients. Savolainen, et al.,[34] recently conducted a randomized study of one-level anterior cervical procedures in which outcomes in patients who were treated with simple ACD were comparable with those who underwent iliac crest graft placement with or without plating.

Other surgeons have argued that microdiscectomy alone is associated with a short-term increase in pain, long-term narrowing of the decompressed foramina, and increased stress on the adjacent disc levels secondary to loss of the normal sagittal alignment of the spine.[37,41] Maintenance of disc height and fusion is desirable, because the loss of disc height and the resultant increased motion is thought to be involved in the pathophysiology of spondylosis. Watters and Levinthal[44] reported that patients who underwent ACD and fusion experienced earlier resolution of their presenting symptoms and a slightly greater chance for long-term success than those who underwent ACD alone. In the fusion group, a greater number of complications (16 compared with four) were demonstrated, which were primarily related, however, to the iliac crest donor site (15 of 16). They concluded that neither procedure was ideal.

Harvesting autogenous bone from the iliac crest can occasionally be associated with marked blood loss. The most important problem, however, is postoperative graft site-related pain. Sawin, et al.,[35] reported a morbidity rate of 25.3% (morbidity included pain, hematoma, fracture, and meralgia paresthetica). The disturbing discomfort in the early postoperative period often resolves with time, yet some authors have shown that residual pain may continue for as long as 24 months after surgery.[4] In the present study, three patients (16.6%) reported ongoing donor site- related pain for more than 6 months.

Avoiding the morbidity associated with the iliac crest has prompted some surgeons to use alternatives such as allografts. The disadvantage of allogenic bone, most commonly iliac crest, is that it may collapse in up to 30% of patients, with a mean height loss of 50%.[7,9,49] When using allografts the surgeon must also take into account the costs of a bone bank and a strict program of quality control. Concerns regarding the risks of transmissible agents such as acquired immune deficiency syndrome virus further restrict the use of allograft bone.

A variety of bone substitutes have also been described. Polymethylmethacrylate is an excellent cement for orthopedic prostheses, but it fails to meet the demands required for an interbody fusion device.[43,45] Although PMMA may be useful in reconstruction after resection of tumors, its use after discectomy is associated with necrosis of adjacent vertebrae.[42] In some studies PMMA fared worse than simple discectomy alone, and it has been suggested that PMMA acts only as a spacer that hinders fusion.[42]

The initial results following the use of HA appeared encouraging. Thalgott, et al.,[40] used treated coral for bone replacement. The combined use of this product and a rigid internal fixation device was strongly recommended. In another study, the authors relied on an alumina core coated in HA, having found HA blocks to be too brittle.[8]

In one frequently quoted report the authors claimed BOP to be a safe, biocompatible, osteoconductive matrix.[22] In subsequent studies, however, authors could not support these claims. The lucency of the BOP graft made radiological assessment difficult. High rates of graft collapse or extrusion were noted.[20] Examination of biopsy specimens of the polymer obtained at reoperation showed a foreign-body, giant cell reaction and no evidence of biodegradation or osseous conduction. It was concluded that BOP acted only as a spacer.[18,19,23]

Left: Lateral radiograph revealing slight iliac crest autograft extrusion (arrow) resulting in pseudarthrosis. Right: Radiograph demonstrating cage subsidence without consequences regarding fusion. Bridging bone can be observed anterior to the cage.

The ideal cervical fusion substitute should result in fusion in all patients and offer maximum comfort for the patient. It should avoid painful autograft sites and associated soft-tissue morbidity, obviate the need for cervical orthosis, and not impede subsequent radiological investigations. It should provide immediate stability in compression and resist axial displacement, minimize neck pain, and maintain spinal alignment and foraminal height. In addition to a high fusion rate, clinical success depends on the maintenance of interspace height and sagittal balance. The latter is accomplished by avoiding collapse of the graft during its incorporation.

All these factors led many authors in recent years to study the results associated with fusion cages. Most of the studies, however, have not been controlled, are retrospective, or lack adequate follow-up periods.[2,5,10,21,26,31] There are only few prospective reports with the appropriate number of patients and longer follow-up periods in which cages are compared with controls. Hacker, et al.,[17] conducted a prospective, controlled, randomized, multicenter clinical study of a threaded titanium cage, the BAK/C cage. They reported successful fusion for one-level procedures at 12 months in 97.9% of patients in the BAK/C group and in 89.7% of those in a bone graft group. The complication rate in the bone graft group was 20.4% compared with an overall complication rate of 11.8% in those treated with BAK/C. The authors concluded that clinical outcome after threaded cage-assisted cervical fusion cage is similar to that achieved by bone autografting after ACD. Cage-assisted fusion, however, has a lower risk of complications.[17] Bartels, et al.,[5] recently examined the use of rectangular carbon fiber cages in 13 consecutive patients with symptomatic cervical disc herniation and reported good results with regard to restoration of the foraminal height and of lordosis, as well as a low incidence of subsidence at 1 year postoperatively.

Pilot studies with the RABEA cage reported by al-Hami[2] and Lange, et al.,[21] yielded promising results. In neither study was a control group included, and there were no available data about lateral flexion-extension radiographs. Cage technology obviates the need for harvesting iliac crest or using allografts. Rectangular cages with spiky indentations are autostabilizing and further immobilization by a plate is not necessary. An interface with the esophagus, which is common to all plate devices and may provoke dysphagia, is not present with cervical cages. Because the cages are sunk just below the vertebral surface, they constitute essentially a "no profile" device for stabilization. In situ stabilization of the cages by axial compression is further enhanced by the spikes firmly secured into the adjacent endplates. This fact has now led us to change our policy regarding postoperative orthosis. All patients in whom cages are implanted wear only a soft collar for 6 weeks.

One drawback of titanium cages is the potential for artifacts in the vicinity of the cage observed on magnetic resonance imaging studies. This could be relevant in the case of a neurological complication. Nevertheless, lesser field strength and the use of fast spin-echo techniques can reduce titanium artifacts.[16,29,39] A second source of criticism could be the fact that solid fusion cannot be easily and definitely determined from simple radiographic analysis alone. Because the device is not radiolucent, it is difficult to determine on radiographs whether solid osseous fusion has occurred (osseous trabeculation, evidence of bone formation in and around the device). Nevertheless, histological examinations of the tissue obtained in the hollow spaces from removed cages after second-look surgery in the studies of Lange, et al.,[21] and Carvi, et al.,[10] confirmed sufficient bone growth in these areas. Newly formed bone, as a result of a differentiation process in different stages, could be demonstrated. Although solid incorporation of the cage into adjacent bone is desired for stability, it may complicate revision surgery. Partial corpectomy would then be necessary.

A final point of criticism could be the relatively high incidence of cage subsidence (in the present study eight [35%] of 23 levels). This seems to be a common complication with the use of any metallic device.[47] In our study, cage subsidence occurred without causing any clinical significance. It was more frequently observed early in the series and is attributed to overaggressive removal of the endplates to "overpromote" fusion, because no bone material was inserted in the cage. It has been observed rarely since we began applying a less aggressive technique, without any consequences concerning fusion. All cases of subsidence occurred within 3 months of surgery and showed no progression over time, probably as a result of fusion. Thus far, no long-term data on subsidence of cages in the cervical spine are available.

In the present study, similar fusion rates and clinical outcomes were present in both groups of patients. The obvious advantage of the cage is the immediate relative stability (in compression) resisting axial displacement, the ability to act as an incompressible spacer maintaining spinal alignment and foraminal height, and most importantly the absence of harvest site morbidity. The benefit related to the use of a cage compared with traditional material may be more a matter of avoiding the indirect morbidity rather than obtaining superior results.