Retrieved human allografts - A clinicopathological study

Background: We studied seventy-three massive preserved human allografts, retrieved from two to 156 months after implantation, to provide insight into the mechanisms of their repair. Methods: The specimens were studied with radiographic and histological techniques that permitted time-related quantitative analysis of the reparative mechanisms of union, cortical repair, soft-tissue attachment, fracture, and characteristics of the allograft-cement interface and the articular cartilage. Results: Union at cortical-cortical junctions occurred slowly (approximately twelve months) by host-derived external callus that bridged the junction and filled the gap between abutting cortices. The bone in the gap did not undergo stress-oriented remodeling even after many years, and, when the union was intentionally disrupted, failure occurred at the cement line that marked the allograft-host junction. Repair of the necrotic graft matrix was both external and internal. External repair consisted of the apposition of a thin seam of host bone on the outer surface of the graft, coating about 40% of the surface art one year and 80% at two years. Internal repair was confined to the ends and the periphery of the cortices and penetrated so slowly that only 15% to 20% of the graft was repaired by five years, after which deeper repair seldom occurred. Graft fractures in specimens retrieved soon after fracture showed only necrotic bone adjacent to the fracture site, whereas those retrieved after fracture-healing showed a marked increase in internal repair of the bone about the fracture site. When bone cement had been used to fix a prosthesis, there was no evidence of bone resorption or loosening of the device. The osteoarticular specimens showed no survival of chondrocytes in the articular cartilage. However, the architecture of the acellular cartilage was well preserved after two to three years and occasionally after as many as five years. Late degenerative changes in the articular cartilage coincided with subchondral revascularization and fragmentation, and the articulating surfaces became covered by a pannus of fibrovascular reparative tissue. Degenerative changes in articular cartilage occurred earlier and were more advanced in specimens retrieved from patients with an unstable joint than in those retrieved from patients with a stable joint. Conclusions: Repair of massive human allografts is an indolent process that follows a fairly predictable course during the first few years and is influenced by other biological activities, such as fracture repair, supplementary autografting, and tumor recurrence. Clinical Relevance: These observations provide a clear, detailed picture of the extent, timing, and deficiencies in the incorporation and repair of large human allografts preserved by conventional banking techniques. As such, they provide a basis for comparative studies of the efficacy of the recently developed osteoinductive substances currently under investigation.