Purpose:There are few studies that have explored the relationship between diet and prevalence of osteoarthritis (OA). Although it is usual to study individual nutrients, a dietary pattern approach has the advantage of assessing food combinations, and it avoids the errors associated with nutrient databases. We have shown that dietary patterns were associated with markers of bone health: a healthy diet being related to lower bone turnover; and both a processed food diet, and a diet low in milk and tea consumption, being associated with lower hip bone mineral density (BMD). The aim of this study was to test the relationship between dietary patterns and osteoarthritis.Methods:Data from the 1998-2000 visit of the Aberdeen Prospective Osteoporosis Screening Study (APOSS) were used for this study. Women were aged between 50-62 y. They were asked whether they suffered from OA. This was validated by radiological records in a subset who also stated the site of OA (which included hands, hips, knees, feet, spine, shoulders and elbows).Diet was assessed using food frequency questionnaires (FFQ) (n¼3229) and dietary patterns were generated by principal components analysis using energy-adjusted food intakes. Logistic regression was used to test the relationship between dietary pattern and prevalence of self reported OA. Results:A total of 596 women had self reported OA compared to 2633 who reported no OA. Five dietary patterns (accounting for 26% of the total variance in diet) were tested with OA prevalence, using logistic regression.One dietary pattern, which had negative factor loadings for milk, tea and cereal, and positive loadings for crisps/ nuts, sauce/ jam and confectionery,was associated with lower OA prevalence. This relationship remained significant after adjustment for weight, height, age, national deprivation category, smoking, physical activity and also hip BMD. Further exploration of tea drinking showed that whereas self-reported OA was associated with more tea drinking; coffee drinking appeared to be protective with higher consumption in women who did not report OA. Further, although numbers were small both coffee drinking and being heavier were significantly associated with reduced risk of hand OA (reported hand OA, n 148; did not report hand OA n 96).