Study limitations
Some limitations to our longitudinal study need mentioning. First, it is possible that a “floor effect” on activity limitations as measured by the WOMAC-pf is responsible for the nearly significant association with change in muscle strength. Eight persons improved on the WOMAC-pf over 3 years to the minimum score of 0; none deteriorated to the maximum score of 68. After exclusion of these 8 participants in regression analyses, the association with change in muscle strength was significant (B=−8.25, P=.030). Second, our data cover only a 3-year period. The associations we found might be stronger if the intervals between the assessment of muscle strength and activity limitations were longer. Third, the small sample size may limit the ability to estimate associations with sufficient precision in our models. However, the associations of change in muscle strength with change in activity limitations were persistent in all analyses. Fourth, the present study used only a motion detection test as a measurement of proprioception and not a position detection test. Both measurements seem to test different aspects of knee proprioception. 47 and 48 The motion detection test is, however, more reliable than a position detection test. 47 Further studies are needed on the influence of both knee joint motion sense and knee joint position sense on the association between muscle strength and activity limitations. Finally, missing data at t1 might have biased our results. Five participants with total joint replacement dropped out of the study. These 5 participants showed high activity limitations at baseline (data not shown). It is to be expected that these participants would have influenced the results at follow-up by showing stronger associations between change in muscle strength and activity limitations.
The results from the present longitudinal study substantiate that improvement in muscle strength might reduce activity limitations or even prevent the development of activity limitations in the early phases of knee OA. This is in line with the beneficial effects of exercise therapy and physical activity, as observed in clinical trials4 and observational studies22 in more developed knee OA.