Dietary factors and MRI-Defined synovial inflammation progression in knee osteoarthritis: a longitudinal mixed-effects analysis from the osteoarthritis initiative.

Background: Dietary factors have been implicated in the development and progression of knee osteoarthritis (KOA), yet their role in modulating synovial inflammation remains unclear, particularly when assessed using quantitative MRI. This study aimed to investigate the cross-sectional and longitudinal associations between dietary intake and MRI-defined effusion-synovitis (ES) in knees at risk of KOA. Methods: This was a secondary analysis based on the Pivotal OAI MR Imaging Analyses incidence OA (POMA_INCOA) sub-cohort. Dietary exposures were derived from the Block Brief 2000 food frequency questionnaire. The primary outcomes were baseline ES volume and 12-month change in ES volume (ΔES). Linear mixed-effects models with participant-specific random intercepts were used to account for within-subject correlation due to bilateral knees. Models were adjusted for age, sex, body mass index, baseline ES volume (for longitudinal models), and Kellgren-Lawrence grade (KLG). Sensitivity analyses included additional adjustment for physical activity and diabetes, as well as alternative outcome specifications (winsorization and categorical progression models). Results: A total of 659 knees from 620 participants were analyzed. In cross-sectional analyses, a higher intake of meat, fish, poultry, eggs, and legumes was significantly associated with a larger baseline ES volume (β = 0.452, 95% CI: 0.075 to 0.829, p = 0.019), though this association did not retain statistical significance after stringent FDR correction (FDR-adjusted p > 0.050). In longitudinal analyses, no dietary variables, including percentage of energy from fat (β = -0.005, 95% CI: -0.150 to 0.006, p = 0.383), demonstrated a statistically significant association with 12-month ΔES. This lack of longitudinal association remained consistent across all sensitivity analyses, including models adjusted for physical activity and diabetes, and those using winsorized or binary progression outcomes (all p > 0.05). Conclusion: While specific protein-rich food groups are cross-sectionally linked to baseline synovial inflammation volume, dietary factors are not robustly associated with the short-term progression of effusion-synovitis in knees at risk of KOA. These findings suggest that dietary intake may correlate with the inflammatory "set-point" of the joint but has limited influence on active 12-month inflammatory fluctuations, highlighting the complexity of inflammatory mechanisms in early KOA.