Background Knee osteoarthritis (KOA) related to metabolic stress and its risk factors, so called metabolic KOA, is a well-known subtype. Recent epidemiologic studies showed that poor dietary carbohydrate quality, such as a high proportion of carbohydrate energy intake as sugar, is associated with a higher risk for metabolic complications and cardiovascular mortality.

Objectives The purpose of this study was to analyze dietary nutritional factors associated with KOA, focusing on the dietary glycemic index (DGI), a higher score of which means poor dietary carbohydrate quality.

Methods This was a cross-sectional study including 9,203 participants aged ≥50 years from The Fifth Korean National Health and Nutrition Examination Survey, who were selected using two-step stratified clustered equal-probability systematic sampling. Radiographic KOA (RKOA) was defined as a Kellgren-Lawrence grade of ≥2. Dietary information was obtained using the 24-hour dietary recall method. The quality of dietary carbohydrate was calculated using glycemic index (GI) based on glucose (GI of glucose being 100). The amount of carbohydrate (g) from each food item participants recalled having had in the previous 24 hours was multiplied by the GI of each item, and the products were summed. The summation was divided by total carbohydrate intake (g) in the 24 hours, and the resulting average GI, the DGI was used as the measure of dietary carbohydrate quality. Sensitivity analysis was performed in a sub-population not undergoing diet modification for weight loss, and excluding participants having extremely high or low dietary energy intake (beyond 2 standard deviation of the distribution). All analyses were stratified by sex and weighted for the sampling design.

Results The prevalence of each knee-state group was as follows: normal (NO) 72.0% (95% CI 70.2%–73.8%), radiographic KOA (RKOA) 16.8% (15.3%–18.3%), chronic knee pain without RKOA (CP) 6.8% (5.9%–7.8%) and symptomatic RKOA (SRKOA) 4.4% (3.8%–5.2%) in men, and NO 47.3% (45.4%–49.2%), RKOA 23.9% (22.4%–25.4%), CP 9.6% (8.7%–10.6%) and SRKOA 19.2% (17.8%–20.6%) in women. In women, DGI was significantly higher in CP (61.2 [SE 0.4]) and SRKOA group (63.1 [0.3]) than in NO group (59.3 [0.2]), and the estimates for age-adjusted mean difference were 1.3 (SE 0.4, P-value 0.007) for CP group and 1.2 (0.3, 0.001) for SRKOA group. In analysis adjusted for age, obesity and the day of week of the dietary recall, the association with SRKOA across quintiles of DGI was Q1 1.00 (reference), Q2 OR 1.20 (95% CI 0.81–1.76), Q3 1.54 (1.09–2.18), Q4 1.63 (1.14–2.32) and Q5 1.73 (1.22–2.44; P value 0.025). CP did not show a linear relationship across the quintiles of DGI. RKOA was not significantly associated with DGI. The association was consistent in the sub-population of no dietary modification and no extreme energy intake. In men, there was no significant association between DGI and any knee state.

Conclusions There was a significant association between poor dietary carbohydrate quality and symptomatic KOA in women. This study provides the foundation for clinical research that addresses the causality of the relationship and the effect of dietary modification on KOA and its co-morbidities.

Disclosure of Interest None declared