Impact of the joint space width measurement method on the design of knee osteoarthritis studies

Recent guidelines recommend measurement of articular loss over several years, determined by conventional X-rays, as the principal outcome measure in clinical trials of potential structure-modifying drugs in osteoarthritis (OA). The aim of this study was to assess the impact of the joint space width measurement method on sample size calculation in knee OA studies.

Standard knee X-rays were taken in 212 patients with knee OA at baseline and after 3 years of follow-up. Mean joint space width (JSW) was measured with an in-house computer-assisted method. Minimum JSW, measured with a graduated magnifying lens, was taken as external standard. After calculation of the intra- and inter-observer reproducibility of the JSW, sensitivity to change was assessed using the standardized response mean (SRM). The number of patients needed to identify a mean significant difference of 0.5 mm in joint space narrowing between the placebo and the treated group, after 3 years of follow-up, was then calculated.

JSW measured with the computer-assisted technique showed better intra- and inter-observer reproducibility than when using the magnifying lens. JSW values measured with our computer-assisted method were significantly correlated with JSW values obtained using the magnifying lens (r=0.87, p<0.001). The SRM were 0.44 and 0.40 for the computer-assisted method and magnifying lens, respectively. The number of patients needed was 131 per group using the computer-assisted method, and 104 using the magnifying lens.

Our method of measurement of JSW may be of potential use in longitudinal studies evaluating the effect of structure-modifying drugs in OA, due to its high level of precision and efficiency. However, although sensitivity to change is markedly better with the digitized method compared with the graduated magnifying lens, we recommend the measurement of mean and minimum JSW in structure-modifying OA trials.