Bone mechanical properties are typically evaluated at relatively low strain rates. However, the strain rate related to traumatic failure is likely to be orders of magnitude higher and this higher strain rate is likely to affect the mechanical properties. Previous work reporting on the effect of strain rate on the mechanical properties of bone predominantly used nonhuman bone. In the work reported here, the effect of strain rate on the tensile and compressive properties of human bone was investigated. Human femoral cortical bone was tested longitudinally at strain rates ranging between $0.14–29.1s−1$ in compression and 0.08–17 $s−1$ in tension. Young’s modulus generally increased, across this strain rate range, for both tension and compression. Strength and strain (at maximum load) increased slightly in compression and decreased (for strain rates beyond 1 $s−1$) in tension. Stress and strain at yield decreased (for strain rates beyond 1 $s−1$) for both tension and compression. In general, there seemed to be a relatively simple linear relationship between yield properties and strain rate, but the relationships between postyield properties and strain rate were more complicated and indicated that strain rate has a stronger effect on postyield deformation than on initiation of yielding. The behavior seen in compression is broadly in agreement with past literature, while the behavior observed in tension may be explained by a ductile to brittle transition of bone at moderate to high strain rates.

Glenoid fixation failure has only been grossly characterized. This lack of information hinders attempts to improve fixation because of a lack of methodologies for detecting and monitoring fixation failure. Our goal was twofold: to collect detailed data of glenoid fixation fracture, and to investigate computed tomography (CT)‐scanning as a tool for investigations of fixation failure. Six cadaver scapulas and six bone‐substitute specimens were cyclically loaded and CT‐scanned at clinical settings after 0, 1,000, 5,000, 10,000, 30,000, 50,000 and 70,000 load cycles. The fixation status was evaluated by inspection of the scans. After 70,000 cycles, the specimens were sectioned, and the fixation inspected by microscopy. The results of the microscopy analysis were compared to the CT‐scan analysis. Fracture of the glenoid fixation initiated at the edge of the glenoid rim and propagated towards and around the keel of the implant. The entire process from initiation to complete fracture took place at the polyethylene implant–cement interface, while the cement, the adjacent bone, and the cement–bone interface remained intact. Thus, strengthening the polyethylene–cement interface should improve glenoid fixation. Microscopy results validated the CT methodology, suggesting that the CT technique is reliable. © 2009 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 27:1589–1595, 2009

Total replacement of the glenohumeral joint provides an effective means for treating a variety of pathologies of the shoulder. However, several studies indicate that the procedure has not yet been entirely optimized. Loosening of the glenoid component remains the most likely cause of implant failure, and generally this is believed to stem from either mechanical failure of the fixation in response to high tensile stresses, or through osteolysis of the surrounding bone stock in response to particulate wear debris. Many computational studies have considered the potential for the former, although only few have attempted to tackle the latter. Using finite-element analysis an investigation, taking into account contact pressures as well as glenohumeral kinematics, has thus been conducted, to assess the potential for polyethylene wear within the artificial shoulder. The relationships between three different aspects of glenohumeral design and the potential for wear have been considered, these being conformity, polyethylene thickness, and fixation type. The results of the current study indicate that the use of conforming designs are likely to produce slightly elevated amounts of wear debris particles when compared with less conforming joints, but that the latter would be more likely to cause material failure of the polyethylene. The volume of wear debris predicted was highly influenced by the rate of loading, however qualitatively it was found that wear predictions were not influenced by the use of different polyethylene thicknesses nor fixation type while the depth of wearing was. With the thinnest polyethylene designs $(2mm)$ the maximum depth of the wear scar was seen to be upwards of 20% higher with a metal-backed fixation as opposed to a cemented design. In all-polyethylene designs peak polymethyl methacrylate tensile stresses were seen to reduce with increasing polyethylene thickness. Irrespective of the rate of loading of the shoulder joint, the current study indicates that it is possible to optimize glenoid component design against abrasive wear through the use of high conformity designs, possessing a polyethylene thickness of at least $6mm$.

It is difficult to evaluate glenoid component periprosthetic radiolucencies in total shoulder arthroplasties (TSAs) using plain radiographs. This study was performed to evaluate whether computed tomography (CT) using a specific patient position in the CT scanner provides a better method for assessing radiolucencies in TSA. Following TSA, 11 patients were CT scanned in a lateral decubitus position with maximum forward flexion, which aligns the glenoid orientation with the axis of the CT scanner. Follow-up CT scanning is part of our routine patient care. Glenoid component periprosthetic lucency was assessed according to the Molé score and it was compared to routine plain radiographs by 5 observers. The protocol almost completely eliminated metal artifacts in the CT images and allowed accurate assessment of periprosthetic lucency of the glenoid fixation. Positioning of the patient within the CT scanner as described was possible for all 11 patients. A radiolucent line was identified in 54 of the 55 observed CT scans and osteolysis was identified in 25 observations. The average radiolucent line Molé score was 3.4 (SD 2.7) points with plain radiographs and 9.5 (SD 0.8) points with CT scans (p = 0.001). The mean intra-observer variance was lower in the CT scan group than in the plain radiograph group (p = 0.001). The CT scan protocol we used is of clinical value in routine assessment of glenoid periprosthetic lucency after TSA. The technique improves the ability to detect and monitor radiolucent lines and, therefore, possibly implant loosening also.