Background/Purpose: Dietary fiber reduced risks of metabolic diseases in part by reducing systemic inflammation and body weight. These factors are both likely to contribute to causing osteoarthritis (OA) and both are more strongly associated with painful OA than with radiographic disease. In this study, we assessed the relationship between dietary fiber and risk of knee OA.

Methods: We used data from the Osteoarthritis Initiative, a prospective, multicenter cohort of 4,796 U.S. men (41.5%) and women [mean (SD) age: 61.2 (9.2) years and BMI: 28.6 (4.8) kg/m²] with or at risk of knee OA. Dietary fiber was estimated using a validated food frequency questionnaire at baseline and sex-specific quartiles of dietary fiber were created. Total dietary fiber was the sum of fibers from grains, fruits and vegetables, and nuts and legumes. Incident radiographic OA (ROA), symptomatic OA (SxOA), and knee pain worsening were followed annually until 48 months. Incident ROA was defined as a knee newly developing Kellgren and Lawrence grade ≥2. Incident SxOA was defined as a new onset of both ROA and a painful knee on most days in past month. Knee pain was estimated used the WOMAC pain subscale ranging from 0 (no pain) to 20 (most pain) points, and pain worsening was defined if the score difference between the baseline and each annual exam ≥ 14% of the base score according to the published estimates for the minimal clinical important difference in WOMAC. We used Generalized Estimating Equations to account for the correlation between two knees individually and for analysis of pain worsening for each exam. We further assessed to what extent the association between fiber and OA was mediated by BMI using a marginal structural model.

Results: At 48 months, we identified 869 knees with incident SxOA, 152 knees with incident ROA, and 1,964 knees with pain worsening among 5,752 / 3,350 / 7,951 eligible knees, respectively (Table). Dietary total fiber was inversely associated with SxOA and pain worsening (p- trend <0.01); grain fiber was similarly associated with pain worsening (p-trend<0.02). Approximately 34% of the association between total fiber and SxOA and 22% between total fiber and pain worsening was through the mediation by BMI. As a secondary analysis, adjustment for baseline BMI yielded similarly significant results. No associations were found for dietary fiber with ROA or for other fiber with OA phenotypes.

Conclusion: This is the first study demonstrating that dietary fiber is associated with lower risks of symptomatic OA and pain worsening in the knee that may be partially mediated through reduced BMI. The strongest association was found at the highest quartile of fiber intake, which is in line with the recommended daily fiber of 25 grams for Americans.