The main goal of this study was to evaluate the characteristics of bone quality and its microarchitecture in a series of femoral heads that failed at different times for different reasons by adopting an innovative and specific quantitative histomorphometry and μCT methodology. To do this nine failures were considered, which were split into groups depending on the failure time: 3 specimens failed at less than 6 months (Group 1), 3 failed between 6 months and 3 years (Group 2) and 3 failed at more than 3 years (Group 3) after HR surgery. In comparison with other studies, in this one the Groups were divided arbitrarily to highlight the possible presence of a phenomenon that progresses over time.
Histological evaluation showed the presence of focal areas of osteonecrosis with empty lacunae in the Group 1. In Group 2 and Group 3 partial osteonecrosis also was present; nevertheless, newly formed bone was visible on the surface of the necrotic bone trabeculae. These data were in agreement with those of Steffen et al. who showed that the necrotic changes were associated with appositional new bone formation and marrow fibrosis . In fact, proliferating cells spread through the narrow spaces between the dead trabeculae, differentiate into osteoblast, and subsequently form appositional new bone on the surface of dead trabeculae. At the same time, they initiate osteoclastic resorption of necrotic bone. Osteoclastic resorption, modulated by cytokines released from osteoblast, is crucial for the balance of the repair processes. The bone may be markedly weakened if resorption occurs at the interface of the viable and dead bone, or if revascularization and new bone formation in necrotic areas is prevented by the formation of a fibrous scar . Bone atrophy was observed at histological analyses only in Groups 2 and 3 and these results were confirmed by micro-CT (BV/TV, Tb.N, Tb.Sp) thus suggesting a possible role of mechanical factors (stress shielding). Metallosis, with infiltration and accumulation of metallic wear debris, was visible in Group 2 and 3. Therefore, as shown by other authors who studied the failure mechanism of HR prostheses by conventional radiography and qualitative histology, the present histological analyses confirmed that aseptic necrosis and bone rarefaction might play a crucial role in late failures of HR [6, 14, 15, 17, 23–25].
Unlike previous studies, the present one took into consideration 3 groups of patients according to failure times (from 3 weeks to 8 years); quantitative measurements were performed with histomorphometry and μCT and 3 peri-implant bone regions at different distances from the HR dome (within 0.8 cm (top), from 0.8 to 1.6 cm (central) and from 1.6 to 2.4 cm (bottom)) were considered. This was possible through resin embedding of the femoral heads containing the prostheses, cutting along the coronal plane of the macro-sections and subsequent removal by pressure of the prosthesis that permitted the accurate evaluation of bone histology and microarchitecture with both 2D (histomorphometry) and 3D (μCT) techniques. A different bone architecture was highlighted within each group and, in particular, between the Group 1 and Group 3. Both 2D and 3D measurements showed that bone density decreases over time especially in Group 3 if compared with Group 2 and Group 1. 3D data of different ROIs (top, central, bottom) of both lateral and medial compartments showed a significant decrease in bone quality over time in the top ROI near the dome. This was confirmed by the significant differences in BV/TV, Tb.N and Tb.Sp between Group 1 versus Group 2 and Group 3 in the top ROI. This tendency was visible also in the lower ROIs but bone values did not reach statistical significance.
In the present study bone resorption was observed within the resurfaced femoral head and around the proximal part of the stem. Whereas bone remodeling is a feature of normal metabolism in healthy and osteoarthritic bone, the BHR may result in stress shielding with consequent resorption and narrowing of the femoral neck due to altered loading conditions. This stress shielding is probably due to the implant design with long-stems. In fact, Bidyut Pal and coworkers showed that bone resorption was considerably less for short-stem designs; the short-stem design having stem-bone contact not only led to a more physiological stress distribution but also to bone apposition in the superior side of the resurfaced head . Moreover, 2D results showed significant differences also in the percentage of empty lacunae between Group 1 versus Group 2 and Group 3 in the top ROI. The proportion of empty lacunae gradually increased over the time after surgery.
The present results were in agreement with those of Steffen et al. who showed that samples from late fractures had a significantly higher proportion (84%) of empty osteocyte lacunae within the trabecular bone compared with those of samples from fractures occurring within the first month (48%) after HR . Moreover, in the present study the higher mean percentage of empty lacunae in the central and bottom regions of Group 1 was probably due to a vascular injury. This controversial result might be explained by analyzing the surgical technique. During femoral head preparation the top region is always removed, thus eliminating the bone volume more subjected to osteonecrosis. Moreover, the residual blood supplied comes from the lateral femoral circumflex artery and a recent report  shows two more possible sources of blood supply to the femoral head. These two vessels were identified as the anterior nutrient artery of the femoral neck which origins from the lateral femoral circumflex artery and the inferior branch of the deep branch of the superior gluteal artery.
In the present series all the operations were performed through a posterior approach which is known to disrupt the medial circumflex artery; nevertheless, the failure rate due to bone necrosis was low. Similar findings were observed by Mcbride et al. who reported the same implant survival regardless of surgical approach . By using a surgical approach that preserves the blood supply it might be possible to obtain an improved implant survival at longer follow up [14, 19, 29].
The BIC measurement should not be considered as an index of osteointegration because the surgical procedure of HR insertion is not aimed at achieving primary fixation between the bone and the stem as for traditional arthroplasties. However, a progressive decrease of bone in contact around the stem was observed and the difference was significant between Group 1 and Group 3 patients. The decrease in BIC was probably due to the bone rarefaction, which involves the femoral head; it remains to be seen whether it might also be related to a progressive prosthesis loosening over time. In the present study histological and microtomographic analyses suggest that both processes, bone rarefaction and osteonecrosis, start from the bottom of the peri-implant bone and reach the top region adjacent to the HR dome in the Group 2 and 3. Osteonecrosis is expected to start from the top ROI which is far away from blood vessels and probably more influenced by the presence of cement but the findings in the present study showed the contrary. In fact, some sort of stress shielding due to its close relationship with the implant might be the true reason for this particular finding.
The current study has several limitations. First of all the small number of cases prevents any solid conclusions to be drawn about the real failure mechanisms of HR and the progression of femoral head damage. The inter-individual variability between patients and osseous changes should also be taken into account. Nevertheless, it was not the primary objective of this study to define the pathophysiology of HR prosthesis failure. To the present authors’ knowledge, quantitative methodologies for measuring bone quality and its microarchitecture have never been used to study retrieved HR prostheses. In the present study the histomorphometric and microtomographic evaluations allowed bone microarchitecture alterations to be quantified.