Statistical distribution of the fatigue strength of porous bone cement

Abstract This paper reports on the damaging effects of different percentage porosities on the fatigue life of acrylic bone cement as used in the fixation of orthopaedic implants. Both hand-mixed (HM) and vacuum-mixed (VM) specimens containing different levels of porosity were fatigue tested to failure. A negative correlation between porosity level and fatigue life was demonstrated for both techniques. Considerable scatter was present in the data. Using the pore size distributions for HM and VM cement virtual HM and VM specimens were created containing various levels of porosity. Incorporating the effect of pore size and pore clustering quantified previously using the theory of critical distances a fatigue life prediction could be obtained for the virtual specimens. The virtual data agreed strongly with the experimental findings, predicting the correlation and more significantly the scatter in the experimental results. Using the virtual porosity failure model, it was demonstrated that given a constant porosity the fatigue life can vary by over an order of magnitude in both HM and VM cement. This suggests that not only porosity level but pore size distribution is extremely important in controlling the fatigue life of bone cement. It was verified that pore clustering and pore size are the major contributors to failure in HM and VM cement respectively. Furthermore, given the beneficial effects of porosity it has been proposed that an even distribution of small pores would provide an optimal bone cement mantle. Using the virtual model, it was determined that neither technique was capable of achieving such a distribution indicating a need for a new more reliable technique. The TCD based virtual porosity failure model should prove to be a powerful tool in the design of such a technique.