Clinical fracture is often attributed to a single traumatic event, and the magnitude of trauma required to produce such a fracture decreases substantially with age (Hedlund & Lindgren, 1987) . In addition to age-related bone mass reductions, the lower magnitude of fracture-inducing loads may be accounted for by bone quality factors such as tissue-level material (Boyce & Bloebaum, 1993) and architectural changes (Heaney, 1993). Matrix damage effects from cyclic loading (Schaffler et al., 1994) , with their attendant reduction in cortical stiffness and strength (Carter & Hayes, 1977) suggest that bone fatigue may be a pre-disposing factor in fractures (Biewener 1993). The purpose of this study was to evaluate how matrix damage caused by cyclic loading influences the residual stiffness, strength and toughness of compact bone in a subsequent single-cycle failure.

PROCEDURES

Specimens were taken from paired femoral diaphyseal regions from one individual (49 y.o. male) to minimize inter-specimen variability. Diaphyseal pieces were lathe-turned to produce specimens with 3mm dia. gage region, and were moistened during all steps of fabrication. Half of the specimens, randomly chosen, were removed for use in a complementary microdamage histological study. The remaining specimens were randomly assigned to three groups: no fatigue (n=7), fatigue to 20% strain increase (n=6), fatigue to 40% strain increase (n=6), corresponding to 0%, 18.6(0.6% and 30.9(0.9% modulus loss, respectively. Two Hz cyclic loading, at a tensile load corresponding to 5000 ((, was performed on a servohydraulic testing machine. Specimens were tested at 37(C under constant wetting with calcium supplemented saline solution (Gustafson et al., 1995). Tests were halted when the pre-determined strain increase was reached. Following cyclic loading, without disturbing the set-up, specimens were loaded at 20000 ((/s in tension to failure. Initial and final elastic moduli were determined from the slope of linear trend lines through the first and last cycles of stress-strain data, except for the 0% group, where the initial elastic modulus was taken from the monotonic stress-strain curve. Residual properties (ultimate strength, ultimate strain, and energy absorption to failure) were determined. Data were analyzed by ANOVA with (=0.05 with post-hoc testing by Scheffe's Test.

RESULTS

Initial moduli of all three specimen groups were alike (Table 1, p=0.768). Stress-strain curves (fig. 1) from specimens loaded to 20% strain increase were qualitatively similar to non-fatigued specimens from comparable regions, albeit with somewhat reduced material properties. By contrast, 40% strain increase specimens had a radically different stress-strain relationship, characterized by an elastic region that deviates early from linearity, poorly-defined yield and a truncated post-yield region. Between the no fatigue and 20% strain fatigue groups, residual properties were moderately reduced: ultimate strain (fig. 2) decreased by 13%, strength (fig. 3) fell by 9%, and failure energy (fig. 4) reduced by 16%. However, in the 40% deformation group, reductions were dramatically higher: 64%, 27% and 74% respectively. For all three material parameters, the 20% group was not significantly different from the control (0%) group (0.125 < p <0.215). The 40% group differed very significantly from both the control group (p=0.0001 for each material parameter) and the 20% group (0.0001 < p < 0.0037).

Figure 1 - Single-cycle stress-strain relationships for specimens fatigued to 0%, 20% and 40% displacement increase. Specimens taken from the anterior femoral diaphysis.

Figure 2 - Residual ultimate strain. Bars indicate SD.

Figure 3 - Residual tensile strength.

Figure 4 - Residual energy absorption to fracture.

Fatigue Group %Displacement Increase	0% (n= 7)	20% (n= 7)	40% (n= 6)
Initial Modulus GPa	16.6 ± 1.0	16.3 ± 1.0	16.2 ± 1.3
% Modulus Degradation	---	18.6 ± 0.6	30.9 ± 0.9
Fatigue Cycles	---	177 ± 147	1279 ± 1810
Residual Strain	3.34 ± 0.35%	2.92 ± 0.31%	1.20 ± 0.40%
Residual Tensile Stress, MPa	126.8 ± 10.3	115.9 ± 10.9	92.2 ± 11.7
Energy Absorption MJ/m3	3.25 ± 0.56	2.73 ± 0.36	0.83 ± 0.36

Table 1: Fatigue and monotonic overload properties. (Mean ?SD)

DISCUSSION

The present study shows that damage from low levels of fatigue, to about 20% modulus reduction (~20% strain increase), are well-tolerated and do not significantly compromise the bone tissue properties. When fatigue damage reaches about 30% modulus reduction (( 40% strain increase) however, the material response changes dramatically. The elastic region of the stress-strain curve becomes nonlinear, and the plastic region is all but eliminated (fig. 1); these effects together account for reduced energy absorption to about one-quarter that of non-fatigued bone. With higher fatigue levels, stored energy is released in small amounts as local matrix damage (Carter & Hayes, 1977) and little additional energy may be absorbed during the subsequent single-cycle overload; by contrast, non-fatigued bone can store a large amount of energy during monotonic loading, which is released at once in catastrophic failure. The current data indicate a nonlinear dependence of residual properties on matrix damage in bone, severely compromising its ability to withstand overload following higher levels of damage. These data also suggest that a buildup of unrepaired microdamage in bone, as in aging (Schaffler et al., 1994), will severely alter bone's residual properties and will contribute substantially to increased fragility in the aging skeleton.

REFERENCES

Hedlund, R., Lindgren U. J Orthop Res 5:242, 1987

Boyce, T.M., Bloebaum, R.D. Bone 14:769, 1993

Heaney, R.Calcif Tissue Int 53:S3, 1993

Schaffler, M.B. et al., Trans ORS 19:190, 1994

Carter, D.R., Hayes W.C. J Biomech 10:325, 1977

Biewener, A.A., Calcif. Tissue Int 53:S68, 1993

Gustafson, M.B., et al., Trans ORS 20:297, 1995

ACKNOWLEDGMENT

Supported by NIH grants AR41210 (MBS) & NRSA AR0832 (TMB). Materials provided by the Musculoskeletal Transplant Foundation.