Runners' lower limb kinematics and training characteristics interactions associated with the occurrence of patellofemoral pain.

Patellofemoral pain (PFP) may result from different factors. Thus, we investigated the interactions among kinematic variables, training characteristics and demographics with PFP occurrence in runners. In this secondary analysis from cross-sectional data, we included 83 runners, 42 with PFP and 41 without PFP. 3-dimensional kinematics (peak hip adduction and internal rotation, peak knee abduction and flexion, peak ankle dorsiflexion and peak foot eversion), cadence, running speed, running experience, weekly running distance, age and sex were analysed using the Classification and Regression Tree (CART). The CART model identified 3 subgroups of runners with PFP: 1) weekly running distance ≤29 km, peak knee abduction ≤8.09° and peak ankle dorsiflexion ≤23.14°; 2) weekly running distance ≤29 km, peak knee abduction ≤8.09° and peak ankle dorsiflexion >23.14° and weekly running distance ≤18.5 km/week; 3) weekly running distance >29 km/week, self-selected running speed <8.25 km/h. Three subgroups without PFP were also identified: 1) weekly running distance >29 km and self-selected running speed >8.25 km/h; 2) weekly running distance ≤29 km and peak knee abduction >8.09°; 3) weekly running distance ≤29 km, peak knee abduction <8.09°, peak ankle dorsiflexion >23.14° and weekly running distance >18.5 km. In conclusion, running training characteristics (weekly running distance and self-select running speed) interacted with lower limb kinematics variables (peak knee abduction and peak ankle dorsiflexion) in distinct ways to characterize groups with and without PFP. The identified interactions could be used in clinical decision-making aiming to personalize training and treatment recommendations when dealing with runners with PFP.