OBJECTIVE: To investigate the cross-sectional association of the obesity with the knee osteoarthritis (OA) in a rural population. METHOD: We studied the cross-sectional asssociation between the obesity and the knee OA by analysis of data (Juam-Study based data) from an epidemiologic survey of a population of 1,100 adults who resided around the Juam lake in Chonnam. The recruited sampling were 475 adults above 40 years old. We diagnosed knee OA by typical clinical features such as bony swelling, crepitus and pain on movement and divided the subjects into non-OA group (143 males:148 females) and OA group (60 males:124 females). There is no difference in the mean ages of two groups. The body mass index (BMI) was calculated from a weight and height (kg/m2) and body fat percent (BFP) was measured by bioelectric impedance fatness analyzer (BIA). We defined obesity as a BMI over 30 kg/m2 in both sex or a BFP over 25% in male and over 30% in female. The BMI and BFP were 23.2 +/- 3.2 kg/m2, 24.0 +/- 7.5% in non-OA group and 23.5 +/- 3.4 kg/m2, 26.0 +/- 7.0% in OA group and were compared in two groups. RESULTS: 1) Prevalence odds ratio (POR) for risk factors of knee OA was high in seventh decade, female, a high cholesterol group, and a high BFP group, and was not high in BMI group. 2) There was no association between BMI and knee OA. 3) There was no association between BFP and knee OA. CONCLUSION: We were not able to prove the cross-sectional association between obesity and knee OA. The results suggested that knee OA was a heterogeneous disorder and more possibly influenced by multiple other factors, such as age, sex, and various metabolic abnormalities than obesity itself.