Increased bone mineral content of bone tissue is a positive outcome of reduced turnover rates and the primary reason osteoporotic individuals are treated with bisphosphonates. However, this positive effect has limitations that may contribute to AFF development in the small cohort of individuals who develop these fractures. In particular, this increase in mineral content is accompanied by an increase in the mean age of the tissue. The absence of remodeling produces not only a greater volume of mineralized tissue but also a reduced volume of newer younger tissue, resulting a more homogeneous tissu e with an increased degree of min eralization.

BP treatment has multiple effects on tissue composition and ultimately results in a more homogenous tissue without the heterogeneity in composition that is normally a hallmark of bone tissue. Changes in bone composition have also been reported both in short-term iliac crest biopsies from alendronate-treated women [28], iliac crest biopsies from individuals with AFF on bisphosphonates [25], and in biopsies obtained adjacent to the fracture site in bisphosphonate-treated women [25, 29]. Compositional variability was reduced in biopsies from individuals with AFFs [25, 29]. When mineral composition was examined as a function of typical or atypical fracture morphology in patients on bisphosphonates, the compositional properties of tissue from patients with AFFs (n = 6) fell within the range of values from patients with typical fractures (n = 14), except the mean cortical degree of mineralization was 8 % greater in AFF tissue (atypical 5.6 ± 0.3 versus typical 5.2 ± 0.5) than in bisphosphonate-treated patients with typical osteoporotic fractures [29]. Biopsies were also included from bisphosphonate-naïve individuals with fragility fractures, none of whom experienced AFFs. Although the mean values of most compositional properties were similar in both fracture groups, the tissue in bisphosphonate-treated patients had a more uniform composition than that of bisphosphonate-naïve patients with typical fractures. A study of iliac crest biopsies from AFF and control patients focused on trabecular tissue and examined similar compositional outcome measures [25]. While AFFs were not only present with bisphosphonate treatment, the biopsies were obtained from four patients on long-term bisphosphonate therapy. Trabecular tissue from the iliac crest of individuals with AFF had increased degree of mineralization, increased collagen maturity, and decreased mineralization heterogeneity. These compositional and morphological features could explain the higher incidence of fracture in these patients. Similar decreases in bone material heterogeneity were reported with treatment by different bisphosphonates including alendronate [28–32], risedronate [33], and zoledronic acid [34]. Oversuppression of remodeling by long-term treatment with bisphosphonates allows the proliferation of microcracks that weaken the bone tissue. Thus, loss of heterogeneity may reflect suppressed bone remodeling and inability to repair microcracks while also resulting in less resistance to crack formation and propagation.

A further material change with bisphosphonate treatment is increased nonenzymatic collagen cross-links [28, 29]. At the tissue level, reductions in post-yield toughness were associated with increased nonenzymatic collagen glycation in cortical tissue of the tibia from dogs treated with high doses of alendronate, but not when clinically equivalent doses were administered [35]. While limited data exist for the composition of bone in individuals with AFF, changes in the collagen maturity, a measure of the ratio of nonreducible to reducible collagen cross-links, were reported in both cortical and cancellous tissue [25, 29] and suggest that they may arise from prolonged bisphosphonate treatment.

Suppression of remodeling by bisphosphonates increases microdamage in cortical bone tissue. Bone microdamage increases due to diminished repair [36–39] and increased crack burden, possibly due to a less heterogeneous tissue, leading to failure at lower energy and in a more “brittle” mode. Reduced post-yield toughness of bone tissue was associated with increased crack lengths and density in dogs treated with high doses of either alendronate or risedronate [40, 41]; however, increased microdamage was not present in animals treated with etidronate [42]. As described above, remodeling suppression reduces or eliminates “normal” bone tissue microstructural heterogeneity that results from having osteons and cement lines of different ages. In healthy tissue, the local differences in material properties produced by microstructural features are essential in dissipating energy and blunting crack propagation. Loss of these natural interfaces and crack-blunting processes can result in brittle-mode-type fractures [36, 37, 42, 43]. In addition, the lack of remodeling limits the repair of this damage.

Both brittleness and loss of heterogeneity allow greater progression of microscopic cracks (Fig. 8.2) that can occur with usual physical activity. Material heterogeneity is a mechanism that normally dissipates crack tip growth energy, thereby limiting crack growth. In a more homogeneous tissue, the energy to grow a crack is reduced and crack progression is less impeded. Targeted repair of cracks by newly activated BMUs appears to be preferentially suppressed by BPs [38]. In classical fracture mechanics, loss of material heterogeneity is associated with increased crack initiation and less resistance to crack propagation, leading to a greater risk of fracture [44, 45]. In cortical bone, transverse cracks are normally deflected longitudinally, limiting the effects of damage when the tissue is loaded. The remarkable straight transverse fracture line seen with AFF is an indicator of the dramatically altered tissue material pro perties and the failure of usual mechanisms to bridge or deflect the crack.

In cortical bone, the functional outcome of these multiple effects of BP treatment can be to alter the mechanical behavior of bulk samples as in the case of AFF, presumably reflecting the combined effects of the increased bone mineral content, reduced heterogeneity, and increased microdamage associated with suppressed bone turnover. In general, bone material strength and stiffness were not altered in cortical bone, but post-yield toughness decreased at high doses [40, 46]. Preclinical studies examining bone properties primarily have been performed in estrogen-replete dog models using supraphysiological BP doses [37]. The majority of these studies have examined alendronate treatment, but these changes are also reported with risedronate and etidronate. The reduced post-yield deformation likely reflects increased damage formation, contributing to the brittle failure evident in AFF. Excessively reduced post-yield toughness produces brittle behavior, which is defined as the abse nce of post-yield deforma tion.

When the mechanical behavior is exam ined in small volumes of cortical bone tissue, the elastic behavior has been reported to be both reduced and unaffected. The reported differences may reflect levels of scale. Reference point indentation (RPI) has previously shown differences in the in vivo bone microindentation properties at the anterior tibial cortex of patients with hip fractures compared to age-matched controls [47]. Using RPI, no differences were present among typical and atypical fracture cases at the mid-tibia for microindentation properties nor were these cases different from patients on long-term bisphosphonate treatment, whose values were intermediate between controls and those who sustained fractures [48]. The similarity of properties among fracture cases suggests that the alterations in tissue-level material properties in individuals with AFF are similar to those of individuals who sustain osteoporotic fractures. Micromechanical properties were reduced in patients using alendronate for 6–10 years, corresponding with decreased mineral crystallinity, elastic modulus, and contact microhardness [49]. Tissue from patients with AFF was not examined. These relatively larger sampled volumes may include damage and other effects, but these effects would also be present with RPI, so in vivo measurement may be a critical difference. Nanomechanical analysis of iliac crest biopsies of individuals with severely suppressed bone turnover and atypical fractures (SSBT) showed no differences in cortical modulus or hardness of cortical tissue from AFF patients relative to age-matched and young female controls and osteoporotic individuals who had experienced vertebral fractures [50]. Plastic deformation resistance was greater in tissue from individuals with SSBT. Nanomechanical differences were present in their cancellous tissue. Tissue-level heterogeneity of the elastic modulus and plastic deformation resistance was reduced in the cortical bone of the bio psies from patients with suppressed bone turnover. While hardness and plastic deformation resistance are inelastic measures, their relationship to the tissue-level toughness has not been established.

Finally, if AFFs are due to impaired remodeling and associated mechanisms, as described, a similar fracture pattern might be expected in individuals with pycnodysostosis, a rare disorder with mutations in cathepsin K, the enzyme that digests the organic matrix of bone during the remodeling process [51]. In fact, AFFs have been reported in some patients with pycnodysostosis [52]. However, AFFs have not been reported in other cases of pycnodysostosis or in patie nts with other defects in remodeling, such as osteopetrosis.