EFFECTS OF COLLAGEN FIBER ANISOTROPY ON THE HYDRAULIC PERMEABILITY OF THE MENISCUS

INTRODUCTION: Fluid pressurization of the meniscus is an important mechanism for load support and distribution in the knee joint (1). The hydraulic permeability is an important parameter governing fluid pressurization and flow-dependent mechanical effects in the meniscus. Collagen fiber orientation contributes to a fiber-induced anisotropy for the meniscus, which gives rise to transversely isotropic material symmetry (2). Fiber-induced anisotropy has been shown to affect the material properties of the extracellular matrix, such as the Young's modulus and Poisson's ratio (3,4,5). While the hydraulic permeability has been experimentally determined for the meniscus in uniaxial compression tests (6,7), it has not measured in a configuration for which there is a mechanical contribution from the collagen fibers (i.e., tension). Thus, the effect of fiber orientation on the hydraulic permeability is not known. For a transversely isotropic model, two permeability coefficients are required, k1 and k2, which describe fluid flow parallel and transverse to the collagen fibers, respectively. In this study, tensile testing was conducted to study the transversely isotropic material properties of the meniscus. A 3-D linear biphasic and anisotropic FEM was developed to model the meniscal samples in simple tension, and optimization was used to determine k1 and k2. Values for the hydraulic permeability were obtained for varying strains to detect potential nonlinear effects. The results for permeability coefficients and matrix properties (i.e., Young's moduli, Poisson's ratios) were evaluated for significant effects of collagen fiber anisotropy and strain level on the material properties of the meniscus. EXPERIMENTAL METHODS: Tensile testing was conducted on meniscal samples from skeletally mature dogs (n=10) in circumferential and radial directions FEM OPTIMIZATION: The tensile experiments were modeled with a 3-D custom-written FE code based on biphasic mixture theory (2). A u-p formulation of the governing equations was implemented. The aspect ratio of the samples permitted the assumption that fluid flow was dominant in the thickness direction (width:thickness of 3.6:1, length:thickness of 15:1). Thus, individual permeability coefficients could be determined from each of circumferential and radial test configurations. The meniscus was modeled as a linear elastic material with transverse isotropy based on the principal collagen fiber orientations. The subject-specific 3-D geometry and experimentally determined transversely isotropic properties were incorporated into a FE model of each sample. The Young's moduli E1 and E2 from mechanical testing were used to define the solid matrix in the model, and isotropic behavior was assumed in shear. To obtain a thermodynamicallypermissible dataset, ν23 was set to 0.5, the limit for an isotropic material. For circumferential samples, an iterative least-squares optimization routine was performed on axial stress during stress-relaxation to determine k2 and ν12. For radial samples (r-z orientation), the optimization procedure was repeated to determine k1 using ν12 from the corresponding circumferential sample. RESULTS: Anisotropy did not have a significant effect on the determined permeability coefficients (p>0.05, ANOVA