Allografts promote skeletal regeneration of periprosthetic femoral bone loss

Abstract

Background: Periprosthetic bone loss is a common clinical problem in hip arthroplasty that must be addressed during revision surgery to achieve adequate implant stability. Although bone allografts represent the clinical standard among substitute materials used, evidence of their regenerative potential at the microstructural, cellular, and compositional level is lacking.

Methods: A multiscale imaging approach comprising contact radiography, undecalcified histology, scanning electron microscopy, and nanoindentation was employed on human femoral explants obtained postmortem many years after allograft use during revision surgery.

Results: The degree of skeletal regeneration through allograft incorporation between host bone and allograft bone was highly dependent on the defect depth (R² = 0.94, p < 0.001), while no association between the allograft time in situ and incorporation (R² = 0.06, p = 0.61) was apparent. The host bone-allograft interface showed a high overlap of 4.0 ± 2.9 mm and was characterized by active bone remodelling, as indicated by osteoid accumulation, high abundance of bone cells and vasculature. While bone cement generally limited the incorporation process, the osteocytic canalicular system of the host bone reached the allograft interface to guide bone remodelling.

Conclusion: This is the first multiscale, histomorphometry-based evaluation of bone allografts used in revision hip arthroplasty for femoral bone loss in humans, demonstrating that they adequately facilitate skeletal regeneration through osteoconduction and subsequent remodelling.

The translational potential of this article: This study identified the mechanisms and determinants of femoral defect regeneration through allografts on the basis of a unique sample collection. While our results support their favourable clinical outcomes, the scientific basis for incomplete incorporation is also demonstrated.

Keywords: Allograft; Bone regeneration; Bone transplantation; Osseointegration; Osteocyte; Revision arthroplasty.

Graphical abstract

Fig. 1

Specimen acquisition and processing of femoral sections containing allograft bone. (A) Contact radiographs of the proximal femur after revision arthroplasty with impacted allograft bone, and hydroxyapatite (HA) and glass ionomer cement (GI) in two individual cases were analysed. The red colouring indicates the region of the prepared cross-section for further analysis. (B) Contact radiographs of the determined cross sections containing implants and periprosthetic bone to identify and localize allografts and additional bone substitute material. (C) Incorporation was analysed on undecalcified histological sections of the whole cross-section stained with toluidine blue. Blue lines reflect the host bone border, red lines reflect the allograft bone border. Both lines imply the overlap (i.e., ingrowth). The green double arrow exemplifies the defect depth (i.e. the length between the cement interface and allograft bone border).

Fig. 2

Interface analysis of allograft and host bone indicates an association of defect depth and incorporation. (A) At the interface of allograft (AB) and host bone (HB) an overlap comprising the fusion of viable and non-viable bone was detected. (B) Allograft bone presented with cell-free osteocyte lacunae and bone cement or fibrotic tissue surrounding bone fragments. The overlap region was characterized by a fusion of non-viable AB and viable host HB. Host bone presented with viable osteocytes and osteoblast as well as osteoclasts at the bone surface. (C) The mean overlap of HB and AB in the proximal femur (Fe) was higher than previously reported overlap in the acetabular (Ace) component. (D–F) Analysis of the associations of patient age, allograft time in situ, and defect area with overlap showed no significant results. (G) Representative illustrations of the overlap of HB and AB in cases with a high and low defect depth (green arrow). While a large overlap was found at high defect depth (left) with hydroxyapatite (HA), a low defect depth (right) was associated with a small overlap and fibrosis development. FT: fibrous tissue, BC: bone cement. (H) Strong significant association between the defect depth and overlap. (I) The defect depth was not associated with fibrosis thickness. (J) Left: Signs of demineralizing effects in the presence of bone cement, but also osteoconductive potential of allograft bone and additional hydroxyapatite or solely with allograft bone (right) in the presence of bone cement.

Fig. 3

The host bone-allograft bone overlap is characterized by fusion and high remodelling. (A) Von Kossa- and toluidine blue stained sections were analysed indicating osteoid seams (pink) and bone cells, i.e., multi-nucleated bone resorbing osteoclasts (red arrow) and bone forming osteoblasts, at the host bone (HB) – allograft (AB) overlap (Ov). (B) Hydroxyapatite (HA, red asterisks) was nearly fully incorporated by newly formed host bone, while osteoid seams (red arrows) were frequently apparent near host bone. (C) Glass ionomer cement (GI, red asterisks) presented with thick seams of unmineralized bone matrix (pink on the left, light blue on the right) as well as fibrosis near GI. (D) Overall, the bone volume fraction (BV/TV) was increased in the overlap region. (E) This was partly recapitulated by an elevated trabecular thickness (Tb.Th) compared to allograft bone. (F) Allograft bone, host bone, and overlap showed a similar trabecular number (Tb.N). (G) The amount of unmineralized bone matrix (osteoid) represented by the fraction of osteoid volume per bone volume (OV/BV) was elevated in the overlap region compared to allograft bone. Of note, AB with additional GI cement presented with an OV/BV over 40 %. (H) At the HB-AB overlap, a higher proportion of bone surface covered with osteoblasts (Ob.S/BS) was observed in comparison to AB and to HB. (I) Similarly, the overlap region presented a higher proportion of bone surface covered with osteoclasts (Oc.S/BS) compared to AB. (J) The number of vascular channels per marrow area (N.VC/Ma.Ar) was highest in the overlap region, specifically compared to allograft bone. However, even in the allograft bone region few blood vessels were apparent. (K) Similarly, the overlap region presented a higher proportion of vascular channel area per marrow area (VC.Ar/Ma.Ar) compared to HB and AB.

Fig. 4

Histomorphometric analysis of the overallskeletalstatus in trabecular bone of the iliac crest. (A) Representative images of iliac crest sections stained with von Kossa/van Giesson indicating an age-related normal to low bone mass. (B) The bone volume fraction (BV/TV) and osteoid to bone volume fraction (OV/BV) were in the range of an age- and sex-matched control cohort [18]. (C) No significant association between iliac crest BV/TV and host bone-allograft bone overlap was apparent.

Fig. 5

Allograft bone presents with a higher degree of bone mineral content. (A) Quantitative backscattered electron imaging (qBEI) was performed to determine the local bone mineral density distribution (BMDD), which is indicated by grey values representing the calcium content. In color-coded images of the overlap (Ov), fusion of host bone (HB) and highly mineralized allograft bone (AB) is apparent. (B) Hydroxyapatite (HA), which is characterized by a high calcium content, is completely incorporated and forms a phase contrast to host bone. (C) The contact zone of GI and bone was apparent with near absence of calcium in bone and a changed composition of GI compared to its central region. (D) By quantification of the calcium content, no change in the mean calcium content (CaMean) between the three region was detectable. (E) The peak calcium content (CaPeak) in AB was increased compared to HB and the overlap region. (F) Similarly, the mineralization heterogeneity (CaWidth) was elevated in AB. (G,H) Nanoindentation of the mineralized bone matrix in all regions was performed to analyze mechanical characteristics. HB, AB, and the overlap region did not present with differences in Hardness or Modulus.

Fig. 6

The osteocyte lacuno-canalicular network (LCN) is limited by the allograft-host bone border. (A) QBEI measurements were further used to identify osteocyte lacunae and determine the extent of the LCN. The number of osteocyte lacunae per bone area (N.Ot.Lc/B.Ar) as well as the mean osteocyte lacunar area (Ot.Lc.Ar) was similar in host bone (HB), allograft bone (AB) and overlap (Ov). (B) Additionally, acid etching and subsequent imaging with a scanning electron microscope was facilitated to depict the osteocyte lacunae and branching canaliculi. The interfaces of HB, AB, and bone marrow (BM) were clearly distinguishable. (C) Osteocyte lacunae were only found in HB, while AB was completely absent of lacunae. At the interface canaliculi attached to non-viable AB, but did not penetrate. (D) In the overlap region osteocyte lacunae presented with a higher number of canaliculi (N.Ca/Ot.Lc) compared to both other regions. (E,F) In samples with additional hydroxyapatite (HA) and glass ionomer (GI), osteocytes also arranged parallelly to its interface while canaliculi did also not penetrate the bone substitute materials.