Histology of tissues within retrieved human titanium mesh cages

Abstract

Daisuke Togawa, MD, PhD, Thomas W. Bauer, MD, PhD, Cleveland, OH, USA; Gary L. Lowery, MD, PhD, Phoenix, AZ, USAIntroduction: 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, n=4; iliac crest, n=4), autograft (iliac crest) in combination with hydroxyapatite granules (Interpore Cross International, Irvine, CA; n=1) or hydroxyapatite only (n=1). Cage implantation was performed with anterior and posterior segmental instrumentation (plates, pedicle screws and cross-links). Indications for cage retrieval included failed fusion (n=8) or failed fusion with hardware failure (n=2). 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 without 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, r=0.69). 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.