Purpose (the aim of the study): Suboptimal articular loading contributes to hip osteoarthritis (OA) onset and progression. Both insufficient and excessive articular loading can drive osteoarthritic processes and symptom worsening. Determining whether hip loads, and the muscle forces underpinning these loads, vary from at-risk to established OA could identify potentially modifiable mechanisms of disease early in the OA process. Clearly defined treatment targets that optimise articular loading could improve efficacy of conservative management for cohorts at-risk for (i.e., femoroacetabular impingement syndrome (FAIS) [1]), and with established [2], hip OA. This cross-sectional study aimed to compare hip contact forces and hip muscle forces during walking between individuals with mild-to-moderate hip OA, FAIS, and healthy controls.

Methods: Nineteen participants with mild-to-moderate hip OA (age=61.3±6.4 yrs, body mass index (BMI)=29.8±4.1 kg.m-2, 26% male) walked on an instrumented split-belt treadmill, 24 participants with FAIS (age=27.3±6.0 yrs, BMI=24.5±2.5 kg.m-2, 79% male), and 39 healthy, pain-free controls (age=29.0±5.2 yrs, BMI=23.1±2.6 kg.m-2, 62% male) walked overground at a self-selected speed while three-dimensional whole-body motion, ground reaction forces, and electromyography (EMG) from 4 hip-spanning muscles (gluteus maximus, gluteus medius, medial hamstring, tensor fascia latae) were synchronously recorded. Hip contact forces and muscle forces were calculated for 3 gait cycles per participant using an EMG-assisted neuromusculoskeletal model [3], following a synergy-based calibration [4]. Hip muscles were grouped into hip flexors, extensors, abductors, or adductors based on their primary function. Hip contact forces (bodyweights=BW) and hip muscle forces (BW) were compared between groups across the stance phase using an analysis of variance via statistical parametric mapping. A secondary analysis of peak values was performed using a general linear model with sex and walking speed as covariates (P<.05).

Results: Participants with hip OA walked with lower magnitude hip contact forces than the other groups during loading response (0-25% stance, mean difference: -0.75 BW 95%CI (-0.34, -1.16) (FAIS), -0.61 BW 95%CI (-0.22, -1.00) (controls), P<.01) and terminal stance (81-100% stance, mean difference: -1.06 BW 95%CI (-0.62, -1.50) (FAIS), -0.97 BW 95%CI (-0.57, -1.38) (controls), P<.01) (Figure 1). Compared to the other groups, participants with hip OA generated less peak hip extensor (-22.6% FAIS, -22.8% controls) and peak hip adductor (-26.4% FAIS, -27.1% controls) muscle group forces during loading response, and less peak hip flexor (-33.1% FAIS, -27.9% controls) muscle group force during terminal stance. Hip contact forces and hip muscle forces were not significantly different between FAIS and control groups. After accounting for sex and walking speed (hip OA=0.9±0.2 m/s, FAIS=1.4±0.2 m/s, control=1.4±0.2 m/s), the hip OA group remained different to the other groups for peak hip flexor muscle force (P=.03).

Conclusions: Hip articular loading did not get progressively lower from at-risk to established hip OA. Only those with established hip OA had lower loading than healthy controls, driven primarily by lower extensor, flexor, and adductor muscle forces, and largely explained by slower walking speed. These hypothesis-generating findings suggest the biomechanical mechanisms of OA onset and progression could differ. Future investigation of regional articular loading and tissue-level cartilage biomechanics (e.g., stresses and strains) may provide further insights into biomechanical markers of disease onset and severity. Longitudinal studies are needed to unravel the complex interaction between hip loading, symptoms, and cartilage health, and to confirm relevant biomechanical targets for early treatment to prevent/slow disease progression.