Associations of METS-VF and BRI with risk of osteoarthritis in patients with diabetes or prediabetes: evidence from the NHANES 1999–2020

TL;DR

This study finds that METS-VF and BRI are linked to osteoarthritis risk in diabetes or prediabetes patients, but using both together doesn't improve predictions.

Contribution

The study evaluates the predictive value and additive effect of METS-VF and BRI for osteoarthritis in diabetes or prediabetes patients using NHANES data.

Findings

METS-VF and BRI are independently associated with increased osteoarthritis risk in diabetes or prediabetes patients.

Combining METS-VF and BRI does not improve predictive accuracy compared to using either indicator alone.

RCS analysis shows a non-linear relationship between METS-VF and osteoarthritis risk.

Abstract

This study aimed to explore the relationship between METS-VF and BRI with OA risk in DM or Pre-DM patients, evaluate their predictive value, and assess the additive effect of their combined application.

We utilized data from the NHANES 1999-2020, including adult participants diagnosed with diabetes or prediabetes. Multivariable logistic regression was employed to analyze the association between METS-VF and BRI with OA risk, while RCS models were used to explore non-linear relationships. ROC curves were generated to evaluate the predictive value of these indicators for osteoarthritis and determine optimal cut-off values. Based on these cut-offs, participants were divided into four groups to further assess the impact of different combinations on osteoarthritis risk, followed by subgroup analyses. Finally, we analyzed the additive predictive effect of combining METS-VF and BRI.

A total of 2,614 patients with diabetes or prediabetes were included, of whom 450 (17.21%) were diagnosed with osteoarthritis. Multivariable analysis revealed that both METS-VF and BRI were significantly associated with OA risk (METS-VF: OR = 4.02, 95% CI: 2.24-7.19, p<0.0001; BRI: OR = 2.78, 95% CI: 1.60-4.85, p<0.001). RCS analysis demonstrated a non-linear relationship between METS-VF and osteoarthritis risk, while BRI exhibited a linear relationship with OA risk. ROC analysis determined optimal cut-off values of 7.25 for METS-VF and 5.87 for BRI. Compared to the “Low BRI and Low METS-VF” group, the “High BRI and Low METS-VF” group showed an increased risk of OA, although not statistically significant (OR = 1.73, 95% CI: 0.92-3.27, p=0.93). Additive effect analysis found that the combination of BRI and METS-VF (AUC = 0.650) showed no statistical advantage over BRI alone (AUC = 0.632) or METS-VF alone (AUC = 0.650) (DeLong test p>0.05 for both comparisons). Continuous NRI and IDI analyses further confirmed that the combination of BRI and METS-VF did not demonstrate significant additive effects compared to either indicator alone.

METS-VF and BRI are independent predictors of OA risk in patients with diabetes or prediabetes, but their combined application does not enhance predictive value. This suggests that in clinical practice, for OA risk assessment in patients with diabetes or prediabetes, either BRI or METS-VF can be selected individually, without the necessity of evaluating both simultaneously.

Figures