This is one of the first studies to directly compare the efficacy of femoral vs. iliac allograft, as well as autograft vs. allograft bone, in a rat spinal fusion model. Although, no statistically significant differences were observed between grafting groups in terms of initial graft weight, CT fusion score, or fusion mass volume, iliac allografts were found to yield a significantly higher rate of bilateral fusion via manual palpation. The higher palpation fusion rate observed in the iliac allograft group compared to the corresponding iliac autograft group was potentially related to the higher fusion mass volume (23.2 ± 21.6 mm3 vs. 13.3 ± 8.9 mm3) shown in the allograft group, even though this increase was not statistically significant due to high variability. Another probable contributing factor is the slight difference in overall graft compositions, as the syngeneic hip allografts were derived from the whole ilium (including elements of the acetabulum), while the autografts consisted of the iliac crest alone, thus leading to possible differences in cortical-to-trabecular ratios as well as total cellular composition and concentration. Interestingly, the greater palpation scores seen in the allograft hip group compared to the femoral group were reflected in the histological observations of better integration with host bone and increased osteoblasts and osteocytes in hip-derived fusion masses in contrast to those derived from femoral grafts.
The histological differences in the study were echoed in our in vitro analyses. Hip-derived bone marrow cells exhibited a higher CFU-F frequency than femur-derived populations, thus indicating a higher concentration of the osteogenic progenitor cells than can differentiate into osteoblasts and osteocytes. This increase in hip-derived mesenchymal progenitor cell frequency also likely contributed to the improved integration with host bone observed in hip graft-derived fusion masses, as mesenchymal progenitor cells are known to recruit host cell migration via paracrine mechanisms and improve bone repair.[25] The bone-forming capacity of the hip- and femur- derived bone marrow progenitor cells, however, was found to be similar. There were no differences in osteogenic differentiation in vitro between P1 bone marrow cells derived from either bone marrow population, and histology indicated that both hip- and femur- derived graft groups yielded osteoid deposition within fusion masses in vivo.
The comparable CT fusion scores observed in this study between autograft and syngeneic allograft bone, reflect findings in several clinical fusion studies which also used radiographic or CT measures of fusion to compare autograft to frozen allograft [26, 27]. For example, recent meta-analyses by several groups have found comparable lumbar fusion rates between patients treated with allograft and autograft iliac crest bone, with no significant differences in disability or pain scores [28, 29]. As in previous studies, the discrepancy between CT and manual palpation scores in our study further demonstrates that importance of employing multiple modalities to assess fusion, as relying on radiographic diagnosis of pseudarthrosis alone may lead to false-positive or false-negative results.
Although autologous iliac crest bone remains the “gold standard”, femoral- sourced grafts have been commonly used in spinal fusion and bone grafting procedures. In clinical fusion procedures, femoral ring and femoral dowel allografts have been widely employed as osteoconductive structural grafts in interbody procedures [30–34], while femoral head allograft has been successfully used in interbody [35–38], and posterolateral lumbar procedures [15, 39]. With increasing interest in the use of femoral reamer–irrigator–aspirator (RIA) bone autograft, a number of clinical studies have recently compared the efficacy of RIA and iliac crest autograft in nonunion, posttraumatic segmental bone defect, or ankle fusion patients and have observed comparable [40–43], or increased union rates and times to union [44]. Femoral RIA autograft has also been successfully used in clinical interbody and posterolateral fusion procedures [45, 46]. Our study, which is the first to directly compare femur- and hip-derived bone grafts in spinal fusion, suggests that femur-derived grafts at least perform comparably to hip-derived grafts in rats in terms of CT fusion measures.
Since a key element to the clinical success of autograft bone is the presence of osteogenic cells, we compared the colony-forming and osteogenic capacity of bone marrow cells derived from the femur and ilium. Similar to our observations, previous animal studies in dogs and pigs observed higher CFU-F frequency in bone marrow derived from the iliac crest than from the femur [47, 48]. Comparisons of marrow-derived mesenchymal cells from human ilia and femurs, however, have revealed conflicting results, with some yielding similar properties [49–54], while others have observed higher concentrations of osteogenic progenitors in marrow from the iliac crest [55, 56], or from the femur [57, 58]. These seemingly contradictory results may largely be due to differences in study design, particularly: the isolation methods used; the specific regions of the femur and hip from which cells are isolated, e.g. proximal vs. distal femur, iliac crest vs. anterior superior iliac spine; and, the patient populations under investigation.
One of the key limitations of this study was the high variability observed, particularly in the CT fusion scores and fusion volumes. While it is possible that subtle differences in surgical technique may have influenced this variation, the range in age and weight of the animals used as donors and hosts is a more likely contributing factor, as our previous meta-analysis of rat fusion models showed an association between animal age and/or weight and fusion outcomes [2]. Another potential source of variability, especially in the allograft groups, is the possibility that there were slight within group differences in implanted graft composition; for example, one animal may have received more of the epiphysis portion and another more of the diaphysis portion of the femur. Studies in humans, for example, have shown that the proximal femur contains bone marrow with a higher concentration of osteogenic CFU-F than the distal femur [54]. Although the observed rates of fusion in this study were low compared to several previous rat studies, our allograft hip fusion rate was comparable to the only other study to use freshly isolated iliac allograft, which observed manual palpation fusion rates of ~ 40% in Sprague-Dawley rats [59]. Possible reasons for the lower fusion rates observed in this study include: the stringency of our fusion assessment criteria (i.e. only counting solid bilateral fusion) compared to previous studies; the amount of bone graft that we implanted may not have been sufficient to yield higher fusion rates; and rat strain differences in bone healing. In order to optimize the syngeneic allograft rat lumbar fusion model, future studies will control more carefully for these factors, using a tighter age and weight range in donor and host rats, and will examine: increasing volumes of bone graft; site-specific bone graft regions (i.e. epiphysis vs. diaphysis); and, a wider range of donor bone types.