Fatigue-induced microdamage in cancellous bone occurs distant from resorption cavities and trabecular surfaces

Abstract

Impaired bone toughness is increasingly recognized as a contributor to fragility fractures. At the tissue level, toughness is related to the ability of bone tissue to resist the development of microscopic cracks or other tissue damage. While most of our understanding of microdamage is derived from studies of cortical bone, the majority of fragility fractures occur in regions of the skeleton dominated by cancellous bone. The development of tissue microdamage in cancellous bone may differ from that in cortical bone due to differences in microstructure and tissue ultrastructure. To gain insight into how microdamage accumulates in cancellous bone we determined the changes in number, size and location of microdamage sites following different amounts of cyclic compressive loading. Human vertebral cancellous bone specimens (n=32, 10 male donors, 6 female donors, age 76 ± 8.8, mean ± SD) were subjected to sub-failure cyclic compressive loading and microdamage was evaluated in three-dimensions. Only a few large microdamage sites (the largest 10%) accounted for 70% of all microdamage caused by cyclic loading. The number of large microdamage sites was a better predictor of reductions in Young's modulus caused by cyclic loading than overall damage volume fraction (DV/BV). The majority of microdamage volume (69.12 ± 7.04%) was located more than 30 μm (the average erosion depth) from trabecular surfaces, suggesting that microdamage occurs primarily within interstitial regions of cancellous bone. Additionally, microdamage was less likely to be near resorption cavities than other bone surfaces (p<0.05), challenging the idea that stress risers caused by resorption cavities influence fatigue failure of cancellous bone. Together, these findings suggest that reductions in apparent level mechanical performance during fatigue loading are the result of only a few large microdamage sites and that microdamage accumulation in fatigue is likely dominated by heterogeneity in tissue material properties rather than stress concentrations caused by micro-scale geometry.

Keywords: Biomechanics; Bone mechanics; Bone quality; Cancellous bone; Microdamage.

Figure 1

Cyclic loading was stopped after different amounts of fatigue loading. Each circle, open and closed, represents one specimen (n = 32, 5 specimen were not loaded). The relationship between microdamage sites and resorption cavities was examined in a subset of specimens loaded to the tertiary phase (open circles only, n=9).

Figure 2

Hypothetical normal distributions of the damage site size are shown. Following additional loading cycles there will be an increase in microdamage, either through an increase in the number of microdamage sites (blue, dashed line) or an increase in the size of microdamage sites (red, dotted line).

Figure 3

(A) Resorption cavities were initially identified in two-dimensional micro-computed tomography images by finding eroded surfaces (arrows) and then (B) traced on three dimensional reconstructions of the micro-computed tomography images (blue).

Figure 4

Although the overall damage volume fraction (DV/BV) was correlated with the reduction in Young’s modulus (A), the number of large microdamage sites showed a stronger correlation to the reduction in Young’s modulus (B). The number of large microdamage sites was also correlated with the overall damage volume fraction (C). The overall number of microdamage sites was not correlated with the reduction in Young’s modulus (D).

Figure 5

Representative histograms of the microdamage site volumes in a specimen loaded to the secondary phase of fatigue life (dashed blue) and that of a specimen loaded to the tertiary phase of fatigue life (solid red) are shown. The grey line shows the cutoff we used for defining large microdamage sites (sites larger than 10⁶ μm³). The specimen loaded to the tertiary phase displayed more microdamage sites classified as large than the specimen loaded to the secondary phase.

Figure 6

The spatial correlations between microdamage sites and resorption cavities using the volume based method are shown. A value of 1.0 indicates no spatial correlation and a value less than 1.0 indicates a negative correlation. (A) An eroded surface was less likely to be near microdamage than regions of bone surface selected at random (at a distance of 8 and 17 μm). (B) No significant spatial correlations were observed when using microdamage as the predictor.

Figure 7

The spatial correlations between microdamage sites and resorption cavities using the object based method are shown. While the majority of resorption cavities were near microdamage sites (A), few microdamage sites were near resorption cavities (B).