Background Muscle weakness is a common symptom of knee osteoarthritis (OA), related to activity limitations such as standing up, walking and stair climbing. In previous work, relationships between muscle strength and activity limitations were modeled as linear. We hypothesized that in a large group of OA patients, nonlinearities can be observed, as well as clinically relevant thresholds.

Objectives To investigate whether relationships between upper leg muscle strength and activity limitations in patients with knee OA are nonlinear and if so, to determine thresholds for muscle strength at which activity limitations occur.

Methods Baseline data from 573 patients with knee OA from the Amsterdam-Osteoarthritis (AMS-OA) cohort were used. Upper leg muscle strength (Nm/kg) from the quadriceps and hamstrings was measured isokinetically at a velocity of 60°/second. Activity limitations were measured with the Get Up and Go test (GUG) and Stair-Climb (SC) test, subdivided in time to ascend and descend stairs.

First, linear regression analyses were performed to assess the relationships between muscle strength and activity limitations. Subsequently, it was examined whether nonlinear models fit better than linear models. To detect thresholds, linear plus constant models were used. The threshold was defined as the intersection point of both lines. Bootstrapping was performed to estimate the reliability of thresholds.

Results For the relationships between upper leg muscle strength and activity limitations, nonlinear models fitted the data slightly better than linear models. Exponential models improved the fit by 5.9, 8.2 and 5.2 percentage points for respectively GUG time, time to ascend and descend stairs. The linear plus constant models approached the goodness of fit of the exponential models. Thresholds were 0.93 Nm/kg (95% CI=[0.80;1.05], SEM=0.06), 0.89 Nm/kg (95% CI=[0.77;1.00], SEM=0.06) and 0.97 Nm/kg (95% CI=[0.86;1.09], SEM=0.06) for the relationships with respectively GUG time, time to ascend and time to descend stairs.

Conclusions Exponential models describe the relationships between muscle strength and activity limitations slightly better compared to linear models. Thresholds were determined using linear plus constant models which approached the fit of the exponential models. Our results are a step towards better understanding of the relationship between muscle strength and activity limitations. Future research is needed to assess the clinical value of the thresholds determined with respect to the effectiveness of strength training in knee OA patients.

Disclosure of Interest None declared