The cross-sectional and longitudinal associations of dietary patterns with knee symptoms and MRI detected structure in patients with knee osteoarthritis

Abstract

To examine the cross-sectional and longitudinal associations of dietary patterns with knee symptoms and structures in patients with knee osteoarthritis (OA). Participants with symptomatic knee OA were recruited from a randomised, placebo-controlled trial conducted in Tasmania (N=259) and Victoria (N=133). Diet was assessed by the Anti-Cancer Council of Victoria food frequency questionnaire. Factor analysis was used to identify dietary patterns. Knee symptoms were assessed using Western Ontario and McMaster Universities Arthritis Index (WOMAC) and structures using MRI. Multivariable linear regressions were used to examine associations. Three dietary patterns ("high-fat", "healthy" and "mixed") were identified in whole sample. Participants with higher "healthy pattern" score had lower total WOMAC, pain, function and stiffness scores at baseline but the associations were not significant over 24 months. Three ("western", "vegetable and meat" and "mediterranean") and two ("processed" and "vegetable") patterns were identified in Tasmania and Victoria, respectively. Cross-sectionally, only "mediterranean pattern" and "vegetable pattern" scores were significantly and negatively associated with total WOMAC or function scores. Longitudinally, participants with higher "western pattern" had worsening function (β: 0.35, 95%CI: 0.03, 0.67) and total WOMAC (β: 0.40, 95%CI: 0.07, 0.72) scores; furthermore, "vegetable pattern" was associated with decreased WOMAC stiffness score (β:-0.47, 95%CI:-0.93,-0.02). In contrast, dietary patterns were largely not associated with structural changes. Some healthy dietary patterns were associated with reduced joint symptoms but dietary patterns were not associated with joint structure in this sample with knee OA. Further studies are required to confirm these findings.