How tibiofemoral alignment and contact locations affect predictions of medial and lateral tibiofemoral contact forces

Understanding degeneration of biological and prosthetic knee joints requires knowledge of the in-vivo loading environment during activities of daily living. Musculoskeletal models can estimate medial/lateral tibiofemoral compartment contact forces, yet anthropometric differences between individuals make accurate predictions challenging. We developed a full-body OpenSim musculoskeletal model with a knee joint that incorporates subject-specific tibiofemoral alignment (i.e. knee varus-valgus) and geometry (i.e. contact locations). We tested the accuracy of our model and determined the importance of these subjectspecific parameters by comparing estimated to measured medial and lateral contact forces during walking in an individual with an instrumented knee replacement and post-operative genu valgum (6 degrees). The errors in the predictions of the first peak medial and lateral contact force were 12.4% and 11.9%, respectively, for a model with subject-specific tibiofemoral alignment and contact locations determined through radiographic analysis, vs. 63.1% and 42.0%, respectively, for a model with generic parameters. We found that each degree of tibiofemoral alignment deviation altered the first peak medial compartment contact force by 51N (r(2) = 0.99), while each millimeter of medial-lateral translation of the compartment contact point locations altered the first peak medial compartment contact force by 41N (r(2) = 0.99). The model, available at www.simtk.org/homelmed-lat-knee/, enables the specification of subject-specific joint alignment and compartment contact locations to more accurately estimate medial and lateral tibiofemoral contact forces in individuals with non-neutral alignment. (C) 2015 Elsevier Ltd. All rights reserved.