There is no consensus over the optimal criterion to define the fatigue life of bone cement in vitro. Fatigue testing samples have been made into various shapes using different surface preparation techniques with little attention being paid to the importance of these variations on the fatigue results. The present study focuses on the effect of test sample shape and surface production method on the fatigue results. The samples were manufactured with two cross sectional shapes: rectangular according to ISO 527 and circular according to ASTM F2118. Each shape was produced using two methods: direct moulding of the cement dough and machining from oversized rods. Testing was performed using two different bone cements: SmartSet GHV and DePuy CMW1. At least 10 samples of each category were tested, under fully reversed tension-compression fatigue stress at ±20MPa, to allow for Weibull analysis to compare results. The growth of fatigue cracks was observed by means of the changes in the absorbed energy and apparent modulus. It was found that fatigue crack growth can be altered by the sample shape and production method; however it is also dependent on the chemical composition of the cement. The results revealed that moulded samples, particularly those based on the ASTM F2118 standard, can lead to up to 5.5 times greater fatigue lives compared to the machined samples of the same cement. It is thus essential, when comparing the fatigue results of bone cement, to consider the effect of production method along with the shape of the test sample.

No abstract

Poly(methyl methacrylate) (PMMA) bone cement is used to anchor the majority of total joint replacements (TJRs). Many brands of cement are used, both with and without the addition of antibiotics to reduce the risk of infection. The present study involved determination of various parameters in tensile fatigue loading: 1) energy absorbed (U) vs number of loading cycles (N) and creep strain (e) vs N, during fatigue tests on specimens of an antibiotic-containing cement (SmartSet GHV) and a plain cement (CMW1) and 2) crack length (a) vs fatigue loading cycles (N) and crack growth rate (da/dN) vs Mode I stress intensity factor range (DKI), during Fatigue Crack Propagation (FCP) tests. In the fatigue tests, four different sample types (round, machined; round, directly moulded; rectangular, machined, and rectangular, directly moulded) and tensiontension loading were used. In the FCP tests, compact tension specimens under tension-tension loading were used. It was found that there were limited effects of sample type, except at the highest stress levels, but that these two cements had different rates of crack propagation. These differences were reflected in the fracture surfaces with SmartSet GHV showing accumulation of opacifier around the particles and crack progression around the intial beads, while for CMW1 the opacifier was evenly distributed and the cracks went through the initial beads.

The fatigue performance of bone cement is influenced by the testing parameters. In previous in vitro fatigue studies, different testing conditions have been used leading to inconsistencies in the findings between the studies, and consequent uncertainties about the effects of testing specimen specifications and stress parameters. This study evaluates the role of specimen variables (namely; specimen cross-section shape, surface production method and cement composition) in a range of in vitro stress amplitudes (±12.5, ±15, ±20, ±30 MPa), using S-N (Wöhler) analysis. The two main findings are: while specimen cross-section configuration and fabrication method (specimen type) played a key role in controlling the fatigue longevity of the same cement, the stress amplitude was seen as the dominant controlling variable to affect the fatigue behaviour of different cements when using the same specimen type. Thus, considering the effect of specimen type, testing at high stress amplitudes should be treated with caution, particularly in tension-compression loading, to ensure fatigue failure occurs due to mechanical rather than thermal effects and thus models the in vivo behaviour.