Fatigue crack propagation rates in PMMA bone cement cannot be reduced to a single power law

Cement mantles around metallic implants have pre-existing flaws (shrinkage induced cracks, laminations, and endosteal surface features) and their fatigue failure is related to the fatigue crack propagation (FCP) rate of bone cement. We estimated the relevant in vivo range of cyclic stress intensity factor (DeltaK) around a generic femoral stem (0-1 MPa square root(m)) and determined that previous FCP data did not adequately cover this range of DeltaK. Vacuum-mixed standard bone cement was machined into ASTM E647 standard compact notched tension specimens. These were subject to sinusoidal loading (R = 0.1) at 5 Hz in 37 degrees C DI water, covering a DeltaK range of 0.25-1.5 MPa square root(m) (including a decreasing DeltaK protocol). FCP-rate data is normally reduced to a power-law fit relating crack growth rate (da/dn) to DeltaK. However, a substantial discontinuity was observed in our data at around DeltaK = 1, so two power-law fits were used. Over the physiologically plausible range of DeltaK, cracks grew at a rate of 2.9 E -8 x DeltaK(2.6) m/cycle. Our data indicated that FCP-rates for 0.5 > DeltaK > 0.3 MPa square root(m) are between 10 E -8 and 10 E -8 m/cycle, 1 or 2 orders of magnitude greater than predicted by extrapolating from previous models based on higher DeltaK data.