OBJECTIVE

While the interpretation of cartilage findings on magnetic resonance imaging (MRI) evolves, plain radiography remains the standard method for assessing progression of knee osteoarthritis (OA). We sought to describe factors that explain variability in published estimates of radiographic progression in knee OA.

DESIGN

We searched PubMed between January 1985 and October 2006 to identify studies that assessed radiographic progression using either joint space narrowing (JSN) or the Kellgren-Lawrence (K-L) scale. We extracted cohort characteristics [age, gender, and body mass index (BMI)] and technical and other study factors (radiographic approach, study design, OA-related cohort composition). We performed meta-regression analyses of the effects of these variables on both JSN and K-L progression.

RESULTS

Of 239 manuscripts identified, 34 met inclusion criteria. The mean estimated annual JSN rate was 0.13 +/- 0.15 mm/year. While we found no significant association between JSN and radiographic approach among observational studies, full extension was associated with greater estimated JSN among randomized control trials (RCTs). Overall, observational studies that used the semi-flexed approach reported greater JSN than RCTs that used the same approach. The overall mean risk of K-L progression by at least one grade was 5.6 +/- 4.9%, with higher risk associated with shorter study duration, OA definition (K-L > or = 2 vs K-L > or = 1) and cohorts composed of subjects with both incident and prevalent OA.

CONCLUSION

While radiographic approach and study design were associated with JSN, OA definition, cohort composition and study duration were associated with risk of K-L progression. These findings may inform the design of disease modifying osteoarthritis drug (DMOAD) trials and assist clinicians in optimal timing of OA treatments.