Background/Purpose: Dynamic joint loads are important in the pathophysiology of knee OA and biomechanical interventions aim to reduce these loads in hopes of improving symptoms and delaying disease progression.  Recent evidence suggests that footwear can influence dynamic knee loading and that flat, flexible shoes may result in lower knee loads compared to more supportive, stiff-soled shoes.   Here, we evaluate the longitudinal effects of flexible footwear (“mobility” shoe) compared to a control shoe on knee loading over 6 months. 
Methods: Subjects with radiographic (KL grades ≥ 2) and symptomatic (at least 30mm pain of 100mm scale while walking) medial compartment knee OA were recruited and randomized to receive a flexible soled shoe (mobility shoe) or identical appearing “control” shoe with stiffer sole.  The stiffness of the soles was evaluated using a biomaterial testing system and was substantially different between the shoes. Investigators and participants were blinded to shoe assignment.  Baseline gait analyses were performed using an optoelectronic camera system and multi-component force plate in subjects’ “own shoes”, study shoes, and barefoot. Subjects were instructed to wear the study shoes at least 6 hours/day for 6 days/week. Gait analysis was repeated at 6, 12 and 24 weeks. The peak knee adduction moment (KAM), a validated marker of medial compartment loading, was evaluated.  An intent-to-treat analysis was performed with imputation of missing data using a hot deck method. Repeated measures analysis of variance compared the two arms and planned contrasts were used to further analyze the data and load reductions at various time points.
Results: 22 participants (13 women, mean age 55±7 years) were assigned to the mobility shoe and 28 (21 women, mean age 55±8 years) to the control shoe.  When evaluating the different walking conditions (study shoe, barefoot, own shoes), there were differences over time in the mobility group vs control.  Compared to their own shoes at baseline, the mobility group experienced a 20% reduction in the KAM by 24 weeks (3.06±1.22 to 2.44±0.72 %BW*ht, p=0.002) while the control group experienced no significant reduction (3.13±0.81 to 3.00±0.66 %BW*ht, p=0.361).  Furthermore, as previously described with use of the mobility shoe, there was a trend for a gait adaptation with reduction in loading during the own shoe condition (7.5%, p=0.154) and barefoot gait (13%, p=0.111) in the mobility shoe group from baseline to 24 weeks (Figure). 
Conclusion: This double-blind randomized controlled trial suggests that use of flat, flexible footwear, in this case the mobility shoe, over 6 months results in significant reductions in medial knee loading compared to a stiff-soled control shoe.  Thus, the use of flexible footwear may be an effective biomechanical intervention for the management of knee OA.