Computational Sensitivity Analysis of Knee-Joint Tissue Properties on Pulsed Electromagnetic Field Distribution Using COMSOL Multiphysics 6.0

Abstract

Pulsed electromagnetic field (PEMF) therapy is increasingly investigated as a non-invasive adjunct for knee osteoarthritis, yet clinical outcomes remain heterogeneous, partly because the field actually delivered to intra-articular tissues is rarely quantified. This study presents a finite-element sensitivity analysis of how the electromagnetic and geometric properties of knee-joint tissues affect PEMF distribution. A multilayer knee model comprising bone, articular cartilage, muscle, subcutaneous fat, and skin, surrounded by an excitation coil and an air domain, was implemented in the Magnetic Fields interface of COMSOL Multiphysics 6.0 and solved in the time domain. Using a one-at-a-time design, the electrical conductivity (σ), relative permittivity (ε~r~), relative magnetic permeability (μ~r~), and thickness (d) of each tissue were varied around literature-based baseline values, and the response of the volume-averaged magnetic flux density was quantified through percentage change and normalized sensitivity coefficients S~X~. Under baseline conditions, the average flux density ranged from 3.80 mT (skin) to 6.39 mT (cartilage), with peak values between 5.5 and 9.4 mT; the corresponding peak induced electric field and current density were approximately 0.20 V/m and 0.10 A/m^2^. Relative permeability produced the highest normalized sensitivity (|S| = 0.92–0.95), followed by tissue thickness (|S| = 0.11–0.20, inverse relationship) and conductivity (|S| = 0.10–0.19), whereas permittivity had the weakest effect (|S| = 0.06–0.17). Among tissues, muscle and the superficial fat and skin layers were the most sensitive to conductivity and thickness, while cartilage exposure was governed chiefly by geometry. A mesh-convergence study confirmed numerical stability (<0.2% change beyond 1.29 × 10^5^ elements). The results identify the tissue parameters that most strongly control predicted PEMF exposure and provide a quantitative basis for parameter prioritization, uncertainty assessment, and future patient-specific optimization of PEMF devices for knee pain management.