Introduction:
Osteoarthritis (OA) is a common chronic joint disease that severely affects patients’ quality of life and causes a significant socioeconomic burden. The aim of this study was to investigate the association between relative fat mass (RFM) and OA and to assess the diagnostic efficacy of RFM in predicting OA risk.

Material and methods:
We conducted a cross-sectional analysis using data from the National Health and Nutrition Examination Survey (NHANES) from 2001 to 2020. Thirty-nine thousand six hundred and fifty-eight study participants were included in the study, which used multifactorial logistic regression analyses, stratified analyses, restricted cubic spline curves (RCS), and ROC curves to explore the association between RFM and OA.

Results:
RFM was significantly and positively associated with OA, which remained statistically significant after correction for confounders (OR = 1.062, 95% CI: 1.056–1.069, p < 0.0001). Restricted cubic spline (RCS) analysis showed a J-shaped relationship between RFM and OA (p = 0.024 for non-linear test). Stratified analyses further confirmed that the association between RFM and OA was positive in all subgroups, and the strength of this association varied by age and ethnicity (p < 0.05 for interaction). ROC curve analyses showed that RFM was significantly more diagnostic of OA than body weight, waist circumference (WC) and body mass index (BMI), with areas under the curve (AUC) of 0.646, 0.550, 0.621 and 0.550, respectively.

Conclusions:
RFM may be an important risk factor for OA and has a diagnostic efficacy superior to traditional anthropometric indices in predicting OA risk. Future studies should further explore the mechanisms linking RFM and OA and validate its clinical applicability, with a view to providing new insights and methods for the prevention and diagnosis of OA.