From theory to tissue: Constitutive modeling and underlying assumptions in cartilage biomechanics.

Abstract

Articular cartilage is a charged, multiphasic tissue whose mechanical response emerges from coupled solid-fluid-ion interactions. Modeling this complexity remains a major challenge in computational biomechanics. This scoping review maps cartilage constitutive models and provides a structured, mechanics-informed appraisal of their physiological representation, constitutive assumptions, and numerical implementation practices. Database searches (1995-2025) identified 84 eligible studies. Models were classified into monophasic, biphasic, triphasic, and other constitutive families. To systematically assess modeling assumptions, a mechanics-oriented appraisal framework structured around five evaluation axes (M1-M5) was applied, addressing constitutive closure, dissipative mechanisms, internal physical admissibility constraints, model-problem coherence, and verification/validation practices. Biphasic models dominate current practice, whereas triphasic formulations better capture osmotic and electrochemical effects. Physiological features were represented unevenly across studies: stress relaxation (86.9%), fluid exudation (69.0%), strain-dependent permeability (48.8%), zonal anisotropy (51.2%), and electrochemical coupling (16.7%). Degeneration mechanisms were incorporated in only 23.8% of studies. Across the corpus, most models demonstrated strong model-problem coherence but frequently lacked explicit admissibility constraints and robust verification and validation practices. Numerical transparency was also limited: although software platforms were often reported, solver configuration, convergence criteria, and computational cost were rarely specified. These findings highlight a persistent gap between constitutive sophistication and empirical validation. Advancing predictive cartilage modeling will require closer integration between constitutive formulation, experimental validation, parameter identifiability, and reproducible numerical implementation.