Title: Insight from direct in vivo measurements on the force distribution across the human knee in flexion: can it be modified, and can the internal loads be predicted from external measurements?
Abstract: Purpose: An understanding of the medio-lateral distribution of the tibiofemoral (TF) contact forces and the factors influencing this distribution is essential for the targeted development and evaluation of interventions aiming to influence the progression of osteoarthritis. Mechanical models of the human knee have provided a basic understanding of the load distribution at the joint, suggesting a strong correlation between frontal plane alignment and medial loading during walking, when the knee is loaded in a rather extended position. The results obtained here are, however, sensitive to the many assumptions necessary to establish the models, and it remains unclear whether those relationships still hold when the knee is loaded in flexion. Goal of this study was to quantify the medio-lateral force distribution in deep knee flexion using direct measurements of the in vivo knee loading conditions and to examine whether and if so by how much the load distribution can be changed in flexion. In addition, we aimed to assess whether more readily available measures such as the dynamic external knee adduction moment (EAM) or static frontal plane alignment were able to predict the internal loading conditions in deep knee flexion. Methods: Gait analysis was performed on 9 total knee replacement (TKR) patients with a post-op mechanical axis angle (MAA) ranging from 4.5° valgus to 7° varus. Telemetric implants provided access to the in vivo TF contact forces and the ratio of the medial to total contact force (MR), while kinematics and external knee moments were determined using inverse dynamics analyses based on synchronously collected ground reactions forces and skin marker data for variants of squatting. While maintaining the reference position of their feet established for the neutral squat (approximately shoulder-width apart, 9 subjects), additional data from 6 of these subjects was available where they were asked to squeeze their knees together (valgus squat) or push their knees apart (varus squat). To assess whether knee loading can be modified in deep flexion, the medial contact force (Fmed) and the MR were compared between varus and valgus squats, while linear regression analyses assessed the relation between the external adduction moment (EAM) and either Fmed or the MR and whether the MAA explained any variance in the internal forces. All subjects provided written informed consent to participate in the procedures and the study was approved by the local ethics committee. Results: The mean MR for the valgus squats (0.33 ± 0.09) was lower (p<0.05) than for the varus squats (0.47 ± 0.06), but there was no significant difference in Fmed (0.84 ± 0.:31BW vs. 1.08 ± 0.:19BW, p > 0.05). During the neutral squats the patients reached a similar mean peak knee flexion of 95.6°, while the mean MR and Fmed were 0.40 ± 08 and 0.96 ± 0.:32BW, respectively. Linear regression analysis across the 9 subjects revealed a stronger association between EAM and MR (R2 = 0.88) than between EAM and Fmed (R2 = 0.62) (Figure 1). There was no substantial relationship between either MR or Fmed with the EAM (R2 = 0.01 and 0.21 respectively). Conclusions: This study revealed that the MR but not Fmed was substantially modified in deep flexion, a condition under which static frontal plane alignment did not explain any substantial variation in either measure of internal knee loading. Here, the EAM derived from inverse dynamics analysis was a better predictor of the MR than Fmed directly measured in 9 subjects, corroborating the notion that the EAM is a proxy for the medial-to-lateral force distribution rather than for Fmed. Clarification of whether changes in kinematics, muscle activation patterns or their combination can explain the changes in MR between varus and valgus squats could help to better understand key mechanisms that enable modification of knee loading and its control.