Mechanical characterisation under cycling loading of humerus cortical bone R. Bry a,b,c *, B. Bennani a,b , R. Delille a,b , H. Morvan a,b , A. Hault-Dubrulle a,b and C. Fontaine a,b,c a Laboratory of Automation, Mechanics and Human and Industrial Computing (LAMIH), University of Valenciennes (UVHC), Valenciennes, France; b University of Lille Nord de France, CNRS UMR 8201, Lille, France; c Department of Anatomy, Faculty of Medicine, University of Lille 2, Lille, France Keywords: cortical bone; humerus; behaviour law; experimental characterisation 1. Introduction Mechanical characterisation tests are usually performed on lower limb samples from animals (bovine) or post-mortem human subjects (PMHS) (Abdel-Wahab Adel et al. 2011). Very few studies focus on upper limb bones even though the risks of fracture of these kinds of bones are important (osteoporosis of old women, sportsmen, etc.). Previous results on lower limb cannot be extrapolated to upper limb bones due to the nature of the bones which is different for bones carrying load or not (Sumner and Andriacchi 1996; Jiang et al. 1999). Generally, exper- imental tests are only done under tension or compression, with different sample shapes for both cases and without any study about the influence of the sample localisation. This paper proposes a new experimental framework to test samples in tension and compression with the same sample shape, in longitudinal and transversal directions and with the same initial set-up to avoid intra-variability. The shape of the samples has been specially designed to be closer than a ‘normalised’ one. 2. Methods Twelve samples were harvested from three corpses (subject 1 to 3 – 82, 62 and 45 years old, respectively). The male PMHS are chosen to avoid interference of post- menopause osteoporosis. The embalming technique (methanol, phenol, glycerine and distilled water solution) from the Department of Anatomy, Faculty of Medicine of the University of Lille 2 was used (Unger et al. 2010). The samples were prepared as follows. The upper limbs are dissected. The humeri were first separated from forearm bones and cleaned from their soft tissues. The extremities are then separated from the shaft. Twelve parallelipipedic samples were cut out from three right humeri: 10 in longitudinal direction (< 80 £ 9 £ 2 mm) and two in the transversal direction (< 20 £ 8 £ 3 mm). The two parts [proximal (P) and distal (D)] of the three faces of the humerus diaphysis [anterior – medial (am), anterior – lateral (al) and posterior (post)] were used. The intermediate zone was avoided due to the nutrient artery penetration inside the humerus (Figure 1). The longitudinal samples (HAMD, HALD, HAMP and HPOSTD) were equipped by five strain gauges. Two perpendicular strain gauges are placed on the anterior aspect: one in the longitudinal direction to measure longitudinal elastic modulus E l and the other in the transverse direction for the determination of the Poisson ratio (n lt ¼ 21 t =1 l ). Rosette strain gauges (08/þ 458/–458 from the longitudinal direction) on the posterior aspect to measure the shear modulus G lt . A strain gauge was placed on the anterior aspect of the samples cut in the transversal direction (H1 and H2) to measure the transversal elastic modulus E t . The minor Poisson ratio is deduced from the following expression: n tl ¼ 2ðE t =E l Þn lt : The 12 samples were tested under cycling loading in the elastic domain. Three triangular cycles of ^ 50 mm displacement were applied at 1 mm/min to impose traction and compression behaviour. The length of the sample between the clamping devices was determined to avoid buckling during ISSN 1025-5842 print/ISSN 1476-8259 online q 2012 Taylor & Francis http://dx.doi.org/10.1080/10255842.2012.713616 http://www.tandfonline.com Figure 1. Localisation and cutting of the sample. *Corresponding author. Email: bry.regis@orange.fr Computer Methods in Biomechanics and Biomedical Engineering Vol. 15, No. S1, September 2012, 274–276