"Dynamic knee valgus" - Are we measuring what we think we're measuring? An evaluation of static and functional knee calibration methods for application in gait and clinical screening tests of the overhead squat and hurdle step

Background: "Dynamic knee valgus" has been identified as a risk factor for significant knee injuries, however, the limits and sources of error associated with existing 3D motion analysis methods have not been well established. Research question: What effect does the use of differing static and functional knee axis orientation methods have on the observed knee angle outputs for the activities of gait, overhead squatting and a hurdle step? Methods: A pre-existing dataset collected from one season (September 2015-May 2016) as part of a prospective observational longitudinal study was used. A secondary analysis of data for 24 male footballers, from a single British University football team, was conducted in order to evaluate the effect of static (conventional gait model) and dynamic (constrained and unconstrained mDynaKAD) methods on knee joint kinematics for flexion-extension and valgus-varus angles. Results: No single calibration method consistently achieved both the highest flexion and lowest valgus angle for all tests. The constrained and unconstrained mDynaKAD methods achieved superior alignment of the knee medio-lateral axis compared to the conventional gait model, when the movement activity served as its own calibration. The largest mean difference between methods for sagittal and coronal plane kinematics was less than 4 degrees and 14 degrees respectively. Cross-talk could not account for all variation within the results, highlighting that soft tissue artefact, associated with larger muscle volumes and movements, can influence kinematics results. Significance: When considering the trade-off between achieving maximum flexion and minimal valgus angle, the results indicate that the mDynaKAD methods performed best when the selected movement activity served as its own calibration method for all activities. Clinical decision making processes obtained through use of these methods should be considered in light of the model errors associated with cross-talk and effect of soft tissue artefact.