PMC

Collagen fibril degeneration distributions in (a) baseline (before the degeneration algorithm is applied) and at the end of the simulations with the thresholds of (b) 5 MPa and (c) 7 MPa.

Maximum principal stresses in the subject group 1 (a) and 2 (b) models. Left column presents stress distributions on the tibial cartilage surface during the first and the second peak loading forces (loading response and terminal extension, respectively), whereas right column presents peak maximum principal stresses during the stance phase of gait. Vertical solid lines in the right column indicate the first and the second peak loading forces, while horizontal dashed lines indicate different threshold levels for collagen fibril degeneration.

In our algorithm, cumulative accumulated excessive loading leads to collagen damage and tissue degeneration, leading to reduced stiffness of the collagen fibril network. Specific exclusion criteria for subject groups 1 (BMI  25) and 2 (BMI  30), obtained from the KL0 group in , are presented on top.

The non-linear method (equation ) is indicated with solid lines, whereas the linear method is indicated with dashed lines (equation with x = 4). Non-linear curves represent the collagen fibril degeneration used in the current study. Methods are matched to obtain the same rate of collagen fibril degeneration at 15 MPa stress.

Specific exclusion criteria for 1087 subjects are presented on the right. Subjects with initially zero Kellgren-Lawrence grade at the baseline are highlighted using green and red stripes (these subjects were further divided into normal weight and obese groups, ). Statistically significant differences between adjacent Kellgren-Lawrence grading groups are indicated with a black star (*p < 0.001, one-way ANOVA with Tukey’s Post-hoc test).

Workflow from the subject-specific knee geometry with gait loading (a) to the fibril degeneration algorithm (b) and the results after iterative degenerative simulations (c). (a-left) Knee joint loads obtained from the literature were implemented into the subject-specific whole knee joint models. (a-right) The outcomes (force and rotations) of the whole knee joint model were implemented into the medial compartment model. (b) The fibril degeneration algorithm (left), based on excessive stresses (right), was applied into the medial compartment of the knee joint (most of the experimentally observed degeneration occurred there). The shown stress distributions are from the first peak loading force of the stance phase of gait. (c) Subject-specific fibril degenerations for both subject groups (healthy and obese). Used joint forces and rotations were considered as relative motions of femur with respect to tibia and they were implemented into a reference point located at the middle-central point between the medial and lateral epicondyles of the femur (a) (see details from Methods). In medial compartment models, (b) the location of the reference point was identical with that in the whole knee joint models (a).

(a) The maximum fibril degeneration as a function of time (arbitrary units) with thresholds of 5 MPa (solid line) and 7 MPa (dashed line) in the subject group 1 (normal weight, green) and 2 (obese, red). The curves are obtained at the location where the highest fibril degenerations occurred (see locations from ), when using the threshold of 5 MPa for the collagen fibril degeneration. (b) The fibril degeneration in the subject group 2 with the threshold of 7 MPa along with an experimentally observed increase in Kellgren-Lawrence grade during the follow-up period for 44 knee joints from 35 obese test subjects (mean ± 95% CI, see also ). Since for many patients cartilage degeneration was not initiated during the first year of the follow-up period, the number of patients (N) decreases as a function of time. Nonetheless, the same trend is seen in tissue degeneration both in our novel degeneration algorithm and in experiments. Statistically significant differences in Kellgren-Lawrence grades between adjacent time points are indicated with black stars (*p < 0.001, one-way ANOVA with Tukey’s Post-hoc test). Correlations between experimental follow-up data and simulated cartilage degeneration were calculated with Pearson and Spearman tests using mean experimental values and all observations, respectively.