Interbody cages

Proceedings of the NASS 16th Annual Meeting / The Spine Journal 2 (2002) 3S–44S

31S

pain intensity, concordancy of the pain response and the presence of annular fissures. The data were tabulated and statistically analyzed. Results: Thirty-two patients with a total of 61 IDET-treated disc levels were evaluated. The time interval between pre- and post-IDET secondlook discography was a minimum of 6 months, with a range of 6 to 16 months. Before IDET, 61 disc levels provoked moderately severe to severe concordant pain. After IDET, 29 IDET-treated disc levels (47%) provoked these same pain responses, and 32 of 61 (53%) were painless. (10%) Six of 61 (10%) new (untreated) disc levels were reported as painful. Six of 32 (19%) patients were asymptomatic for 1 year or more before repeat discography. Four of 6 (67%) of these patients reported symptoms at the same level on repeat discography, and 2 of 6 (33%) reported pain at previously untreated levels. Ten of 32 patients (31%) had painless repeat discograms. Eight of 32 (25%) of these patients remained equally symptomatic at all the IDET-treated levels on the repeat discogram. Seven of 61 (11.5%) discs failed to demonstrate annular fissures that were noted on pretreatment discography. These were all discs with contained radial fissures. Discussion: Second-look discography permits a unique evaluation of IDET-treated discs that can not be attained by any other method. In this study, posttreatment repeat discography performed on a group of IDETtreated patients (32) who were unimproved or only partially improved revealed that discogenic pain perception improved after IDET in 32 of 61 treated discs (53%). Ten of 32 patients (31%) had painless repeat discograms. These patients were then treated for their nondiscogenic source of pain and resolved their pain problem. Only 8 of 32 (25%) of these patients remained equally symptomatic at all the IDET-treated levels on the repeat discogram. Six of these patients underwent a second IDET, and two patients underwent interbody fusion. Persisting pain at one disc level in a previously treated multilevel case typically accounted for the partial improvers. Only cases of simple radial fissures failed to be identified on the posttreatment discograms. Conclusion: Second-look discography reveals that even in a group of partially improved patients, IDET-treated discs display a significant reduction in pain perception. Annular morphology was objectively improved in discs with pretreatment radial fissures. The discographic pattern was not changed in cases of more severe disruption, even when the disc was reported as painless after treatment. Further study to evaluate the neurogenic mechanisms responsible for this phenomenon is required. In addition, biomechanical analysis of the annulus to determine if the local stress–strain relationships were impacted by treatment would also be valuable.

out decalcification. Sections were reviewed qualitatively, and the approximate percent of area in the cage occupied by viable bone, necrotic bone (graft), fibrocartilage, fibrous tissue and bone graft substitute were visually estimated. Particles of metal debris were estimated by a semiquantitative scoring system ranging from 0 to 4. Spearman’s rank-order correlation analysis was used to test correlations between duration in situ and the percentage of viable bone, fibrocartilage and the extent of debris. Results: The cage with 2 months duration in vivo showed only fragments of necrotic bone (graft) surrounded by fibrous tissue. All other cages showed evidence of vascular ingrowth and areas of histologically viable bone, almost certainly representing incorporating bone graft. We estimate an average viable bone area of approximately 32% (range, 0% to 80%). The lateral openings of these cages contained approximately 81% fibrous tissue or fibrocartilage (range, 30% to 100%). In spite of the relatively small number of cases, there is a significant correlation between the duration in situ and the proportion of viable bone (p.03, r0.69). At least a few particles of debris were present in 8 of 10 cages, but with the numbers available we did not identify a significant correlation between the debris score and duration in situ (p.36). There was no bone apposition directly to the inner surface of any cage. Instead, a membrane of either fibrous tissue or fibrocartilage separated bone from the inside of each cage. Fibrocartilage in the cages was of two types. Fibrocartilage of probable intervertabral disk origin ranged from 0% to 70% of the available area (average, 22%), and there was no significant correlation between duration and the proportion of fibrocartilage (p.20). In addition, several cages contained small seams of fibrocartilage connecting segments of bone in a pattern that suggests uniaxial motion. There was only focal bone apposition to the hydroxyapatite granules present in one cage. Discussion: In spite of the likelihood of motion in vivo, almost all of these cages showed evidence of at least some bone graft incorporation. The presence of fibrous tissue and necrotic bone (graft), however, suggests that the incorporation was still incomplete. To the best of our knowledge, the seams of fibrocartilage connecting segments of viable bone have not been previously described in this location. A few particles of debris were present in most cages, but there is no histologic evidence of particle-induced bone resorption or inflammation. The study of clinically successful cages retrieved at autopsy would help clarify the safety and efficacy of these intervertebral body fusion cages.

Histology of tissues within retrieved human titanium mesh cages Daisuke Togawa, MD, PhD, Thomas W. Bauer, MD, PhD, Cleveland, OH, USA; Gary L. Lowery, MD, PhD, Phoenix, AZ, USA

Introduction: Although most published clinical results of interbody fusion cages are good, areas of controversy remain, including optimum bone graft preparations and the extent of bone graft incorporation in cages of different designs. The purpose of this study is to describe the contents of a large series of retrieved, clinically failed human interbody cages, with special reference to the influence of graft type on the extent of graft incorporation. We also note the presence of hyaline cartilage in some cages packed with local autograft. Methods: As part of a comprehensive device-retrieval program, 57 cages from 34 patients were analyzed. These included five carbon fiber-reinforced polymer I/F Cages (Brantigan; DePuy-AcroMed, Raynham, MA) and 52 threaded metal cages (18 BAK; Spinetech, Minneapolis, MN, 30 Ray-TFC; U.S. Surgical, Norwalk, CT, 2 InterFix; Sofamor Danek, Memphis, TN, and 2 Aesculap; Aesculap, South San Francisco, CA). Preoperative diagnoses of these patients included spondylosis, spondylolisthesis, discogenic disease and/or stenosis. Cage implantation was performed with anterior and/or posterior segmental instrumentation (plate, pedicle screws and cross-links). Information about graft procedure is available for 20 cages. Of these, 14 had been packed with local autograft, 2 received iliac crest autograft and 2 received cancellous allograft. In addition, we recognized demineralized bone matrix material in 2. The cages had been in situ from 4 to 87 months (mean, 23 months). Indications for cage retrieval included failed fusion (18), migration or malposition (13), progressive spondylosis (1) and/or pain (1). Indications for retrieval are unavailable for

Introduction: The use of vertically placed titanium mesh intervertebral body fusion cages is controversial, and there is little histological documentation of the contents of these cages in the human spine. The purpose of this study is to describe the contents of a series of retrieved, clinically failed human titanium mesh cages, with special reference to the viability of bone graft and the presence of debris. Methods: Ten Titanium Surgical Mesh Cages (Harma; DePuy AcroMed, Raynham, MA) retrieved from nine patients were analyzed. Preoperative diagnoses of these nine adult patients included disease conditions in the sagittal plane: spondylosis (eight), discogenic disease (five) and stenosis (two). The spinal levels treated were C4–6 (one patient), C5–6 (three patients), L3–4 (one patient) and L4–5 (four patients). The cages had been packed with autograft (local, n4; iliac crest, n4), autograft (iliac crest) in combination with hydroxyapatite granules (Interpore Cross International, Irvine, CA; n1) or hydroxyapatite only (n1). Cage implantation was performed with anterior and posterior segmental instrumentation (plates, pedicle screws and cross-links). Indications for cage retrieval included failed fusion (n8) or failed fusion with hardware failure (n2). The cages had been in situ from 2 to 47 months (mean, 25 months). Specimens were fixed in 70% ethanol, dehydrated, embedded in plastic and sectioned with-

Interbody cages: be careful what you pack Daisuke Togawa, MD, PhD, Thomas W. Bauer, MD, PhD, Cleveland, OH, USA

32S

Proceedings of the NASS 16th Annual Meeting / The Spine Journal 2 (2002) 3S–44S

7 cages. Specimens were fixed in 70% ethanol, dehydrated, embedded in plastic and sectioned, resulting in at least one section through the center of each cage. Sections were hand sanded to a final thickness of approximately 50 mm and stained with Giemsa. Slides were reviewed qualitatively and the approximate areas occupied by viable bone, necrotic bone (graft), fibrocartilage, hyaline cartilage, fibrous tissue and bone graft substitutes were visually estimated and expressed as percent of available area. Comparisons of these variables were performed using analysis of variance with Fisher’s PLSD test. Spearman’s rank-order correlation analysis was used to test correlations between duration in situ and the percentage of viable bone, fibrocartilage and the extent of debris. P values less than .05 were considered to be statistically significant. Results: All cages showed evidence of vascular ingrowth and areas of histologically viable bone, almost certainly representing incorporating bone graft. Average viable bone area was approximately 46% (range, 0% to 80%). We were unable to identify a correlation between duration in situ and the extent of viable bone in this group of failed cases (p.79), and there were no obvious differences based on cage design (p.99). Fibrocartilage in the cages ranged from 0 to 50% of the available area (average, 16%), and there was no significant difference in the extent of fibrocartilage based on cage design (p.08). Thirty-one percent of the cages that had been packed with local autograft showed more than 5% of the available area occupied by hyaline cartilage. This cartilage was not associated with bone formation. No hyaline cartilage was identified in either of the two cages that had been packed with iliac crest autograft or in the 26 cages for which information about graft origin is unavailable. In some cages, fragments of cortical bone graft were associated with only minimal new bone formation, whereas smaller pieces of cancellous bone showed greater evidence of incorporation. In several cases, demineralized bone matrix was present and not associated with new bone formation. One cage contained abundant fibrillar foreign material. Discussion: The extent of bone formation in interbody fusion cages will reflect, in part, the quality of the bone graft materials at the time of cage insertion. Previous basic science studies have demonstrated greater bone formation associated with cancellous compared with cortical graft, reflecting primarily differences in surface area. Our qualitative observations from this series of failed cages support that concept and suggest that segments of cortical bone may be less effective graft materials than cancellous bone from the iliac crest. In addition, to our knowledge, this is the first study to recognize the presence of hyaline cartilage in cages that received local autograft. The cartilage is presumably derived from the vertebral end plates or facets and emphasizes the importance of careful graft and implant site preparation. Finally, these cages contain a relatively high proportion of fibrocartilage. It is unclear whether this fibrocartilage made its way into the cages after they had been inserted or was inadvertently packed into the cages along with local autograft. In either case, the fibrocartilage is not associated with bone formation and is undesirable. The histologic findings in failed cages are likely to be different from those in clinically successful cages, but our results still emphasize potential differences in quality of graft preparations from different sites and illustrate the importance of implant site preparation to maximize bone graft incorporation.

The effect of pulsed electromagnetic fields on the quality of cageinstrumented lumbar vertebral fusion in the ovine model Alex Ghanayem, MD, Maywood, IL, USA; Mike Larson, DVM, Logan, UT, USA Introduction: Threaded interbody cages have come into widespread use. Fusion rates are high but not perfect. Problems noted on postoperative radiographs after cage insertion include radiolucency around the cage, indicating a fibrous tissue at the bone–cage interface. Other problems include normal-appearing postoperative radiographs but failure of bony consolidation. Intrinsic and extrinsic factors have been modified in search of increasing fusion rates in cage constructs, including the use of bone graft substitutes and implantable direct current stimulation in a sheep model. External pulsed electromagnetic fields (PEMF) is another way to provide electrical

stimulation to improve the fusion/bone graft incorporation process and may enhance the cage fusion rate. This study examines the efficacy of using an externally applied PEMF in promoting bone growth through and around an anterior lumbar cage construct in the ovine model. Methods: Twelve sheep underwent an anterior cage fusion procedure at the L4–5 level using a left retroperitoneal surgical approach. The interspace was distracted and a single working tube placed on the left lateral surface of the disc space. The interspace was then reamed and tapped. A threaded titanium fusion cage (Interfix; Medtronic-Sofamor Danek, Memphis, TN) was then placed transversely across the disc space. Cages were packed with morsilized autogenous ICBG. Six animals were randomly selected and fitted with a customized external PEMF device (Spinal Stim; Orthofix) that provided 4 hours of stimulation every 24-hour cycle. The other six animals did not receive stimulation and served as a control group. At 16 weeks after surgery, animals were euthanized. Specimens were removed en bloc from T12 to S1 and radiographed in the arteroposterior and lateral projections. Lateral flexion and extension radiographs were also obtained. The radiographs were blindly evaluated for the following characteristics: obvious lucency at the cage–vertebral end plate junctions, sclerosis or reactive bone at the cage–vertebral end plate junctions, no lucency or reactive bone at the cage–vertebral end plate junctions, definite bridging bone across the interspace and the absence or presence or motion, or the appearance of lucencies at the cage–vertebral end plate junctions on flexion-extension radiographs. Specimens were prepared for histologic evaluation and surface-stained with hemotoxylin and eoson. Histologic analysis was completed blindly using the following criteria: the presence or absence of a tight interface between the cage and the surrounding bone without fibrous tissue interposition, the presence or absence of bridging bone from one vertebral body, through the cage, to the other body and the presence or absence of bridging bone from one vertebral body, through the cage or around a portion of the cage, that extended to the other body. Criteria for a successful fusion included no lucency or motion on plain radiographs, the presence of tight cage–bone interfaces without fibrous tissue interposition and bridging bone through the cage on histologic evaluation. The lack of lucencies and motion on plain radiographs would not qualify as a successful fusion unless there were confirmation of fusion on the histologic analysis. Lucencies around the cages, motion on flexion-extension radiographs, fibrous tissue interposition at the cage–vertebral body interface and lack of bridging bone on histology would be classified as not fused. Results: All 12 sheep completed the protocol. One animal experienced transient weakness in the hind limb opposite to the side of cage insertion, which resolved spontaneously over the course of 3 days. Radiographs with lucencies always correlated with fibrous tissue at the cage–vertebral body interface on histologic evaluation. The absence of lucencies, however, did not always predict histologic fusion. Four of the six animals in the PEMF group were judged to have a successful fusion. In the control group, one animal was judged to be successfully fused and one was thought to have a partial fusion. If this control animal is considered a success, then the fusion rates was two of six in the control group. Discussion: Previous studies in sheep have shown recombinant human bone morphogenetic protein or internal direct current stimulation of an anterior cage construct can increase the fusion rate. It should be noted that radiographs without evidence of lucencies or motion did not always predict a successful fusion histologically, thus reaffirming a shortcoming of plain radiographs in assessing the success of a cage fusion. Radiographic bridging bone, however, was predictive of a successful fusion. This study has shown that the use of a nonimplantable PEMF device, with 4 hours of stimulation per day, also yields a significant increase in the fusion rate of an anterior cage–autogenous graft fusion construct.

Revision lumbar arthrodesis for the treatment of cylindrical cage pseudarthrosis: perioperative complications and pitfalls James D. Schwender, MD, Francis Denis, MD, Timothy A. Garvey, MD, John E. Lohnstein, MD, James W. Ogilvie, MD, Joseph H. Perra, MD, Manuel R. Pinto, MD, Ensor E. Transfeldt, MD, Robert B. Wood, MD, Minneapolis, MN, USA