Long-term effects of bisphosphonate therapy: perforations, microcracks and mechanical properties

Abstract

Osteoporosis is characterised by trabecular bone loss resulting from increased osteoclast activation and unbalanced coupling between resorption and formation, which induces a thinning of trabeculae and trabecular perforations. Bisphosphonates are the frontline therapy for osteoporosis, which act by reducing bone remodelling, and are thought to prevent perforations and maintain microstructure. However, bisphosphonates may oversuppress remodelling resulting in accumulation of microcracks. This paper aims to investigate the effect of bisphosphonate treatment on microstructure and mechanical strength. Assessment of microdamage within the trabecular bone core was performed using synchrotron X-ray micro-CT linked to image analysis software. Bone from bisphosphonate-treated fracture patients exhibited fewer perforations but more numerous and larger microcracks than both fracture and non-fracture controls. Furthermore, bisphosphonate-treated bone demonstrated reduced tensile strength and Young's Modulus. These findings suggest that bisphosphonate therapy is effective at reducing perforations but may also cause microcrack accumulation, leading to a loss of microstructural integrity and consequently, reduced mechanical strength.

Figure 1. Synchrotron X-ray Micro-CT imaging of bone microstructure.

(a) Bone cores were mounted in an X-ray beam line and rotated through 180°. A volume of interest (3.28 × 3.28 × 2.76 mm) was scanned at the centre of the core. (b) 640 X-Ray projections were collected at angular intervals of 0.28° and the projections were used to reconstruct 2000 axial slices with a voxel size of 1.3 μm using filtered back projection. (c) Samples were interrogated from the three planes. (d) Rendered in 3D by masking the perforations and microcracks in three planes using VGStudio MAX.

Figure 2. Classification of microdamage.

2D to 3D rendering of microdamage (orange) within trabecular bone (translucent). (a,b) Perforation (P), (c,d) Microcrack (M). The image analysis software, VGStudio MAX, was used to generate the 3D models using image segmentation.

Figure 3. Microdamage characteristics.

Median density (a,b), volume (c,d) and volume fraction (e,f) of perforations and microcracks in the non-fracture control (n = 5), fracture control (n = 5) and bisphosphonate therapy groups (n = 6). (a) The fracture group had the highest density of perforations at 3.92/mm³ across all groups but this was only significant compared to the non-fracture group (p = 0.005). (b) The bisphosphonate therapy group had a significantly higher microcrack density at 7.38/mm³ compared to the other two groups (fracture p = 0.007, non-fracture p = 0.012). (c) The fracture group had a significantly larger volume of perforations at 24863 μm³ compared to other two groups (bisphosphonate therapy p = 0.011, non-fracture p = 0.012). The bisphosphonate therapy group also showed a significantly higher volume of perforations at 15893 μm³ compared to the non-fracture group at 4220 μm³ (p = 0.019). (d) The bisphosphonate therapy group had the highest microcrack volume at 7173 μm³, which was significant compared to the non-fracture group (p = 0.001). The microcrack volume was also significantly lower in the non-fracture group than the fracture group (p = 0.004). (e) The fracture group had the greatest perforation volume fraction at 15.39 × 10⁻³%, which was significantly greater than both the bisphosphonate-treated group, which was 3.02 × 10⁻³% and non-fracture group, which was 0.50 × 10⁻³% (bisphosphonate therapy p = 0.013, non-fracture p = 0.001). (f) The bisphosphonate therapy group had the highest microcrack volume fraction at 11.03 × 10⁻³% (fracture p = 0.017, non-fracture p = 0.005). Statistically, microdamage data were compared using a Kruskal-Wallis and Mann-Whitney U test. Asterisks denote significant pairwise differences at *p < 0.050, **p < 0.010 and ***p < 0.001.

Figure 4. Tensile stress-strain curves.

(a) An example of a stress-strain curve. The Young’s Modulus was obtained by calculating the gradient of the linear section of the curve. The ultimate tensile strength was obtained from the maximum stress value of the curve. (b) Overall stress-strain curves are shown for the non-fracture control (NFC), fracture control (FC) and bisphosphonate-treated (BP) groups. The shaded area represents the 95% confidence intervals for each group.

Figure 5. Mechanical data characteristics.

Median apparent and normalised ultimate tensile strength and Young’s Modulus in the non-fracture control, fracture control and bisphosphonate therapy groups. (a) The bisphosphonate therapy group had a significantly lower apparent ultimate tensile strength at 0.58 MPa compared to the other two groups (fracture p = 0.021, non-fracture p = 0.002). Non-fracture group also had a significantly higher apparent ultimate tensile strength at 1.62 MPa than fracture group at 0.86 MPa (p = 0.002). (b) The bisphosphonate therapy group showed a significantly lower apparent Young’s Modulus at 0.070 GPa than the fracture group at 0.16 GPa and non-fracture group at 0.20 GPa (fracture p = 0.021, non-fracture p = 0.002). (c) The bisphosphonate therapy group had a significantly lower normalised ultimate tensile strength at 23.78 MPa compared to the other two groups (fracture p = 0.028, non-fracture p = 0.002). Non-fracture group also had a significantly lower ultimate tensile strength at 30.28 MPa than fracture group at 47.86 MPa (p = 0.002). (d) The bisphosphonate therapy group showed a significantly lower normalised Young’s Modulus at 2.88 GPa than the fracture group at 5.33 GPa and non-fracture group at 6.41 GPa (fracture p = 0.028, non-fracture p = 0.002). The mechanical data were compared using a Kruskal-Wallis and Mann-Whitney U test. Asterisks denote significant pairwise difference at *p < 0.050, **p < 0.010, ***p < 0.001.