Inflammatory indicators derived from complete blood counts in relation to osteoarthritis prevalence: findings from the NHANES 2007–2020 cross-sectional survey

TL;DR

This study finds that blood count-based inflammation markers are linked to osteoarthritis prevalence, suggesting potential new ways to assess and treat the condition.

Contribution

Identifies novel associations between CBC-derived inflammatory indicators and osteoarthritis prevalence using a large population dataset.

Findings

Five inflammatory indicators (SII, SIRI, MLR, NMLR, NLR) are positively associated with OA prevalence.

Nonlinear associations were found between SII, NMLR, and NLR with OA prevalence.

A prediction model achieved an AUC of 0.735 for OA risk, showing clinical relevance.

Abstract

Although multiple studies have confirmed the importance of chronic low-grade inflammation in the development of osteoarthritis (OA), the association between complete blood count (CBC)-derived inflammatory indicators and osteoarthritis prevalence remains unclear. The present study aims to explore the association between CBC-derived inflammatory indicators and OA prevalence. We used NHANES data from 2007 to 2020 for a cross-sectional analysis. Multivariate logistic regression models were used to evaluate the association between CBC-derived inflammatory indicators and OA prevalence. Restricted cubic spline function (RCS) and threshold analysis were used to assess potential nonlinear associations. In addition, subgroup and sensitivity analyses were performed to assess the stability of the results. Finally, we used LASSO regression to identify the variables most associated with OA outcomes to construct a prediction model, and the model’s validity was verified. Among the 24,112 patients in this study, 3,195 were diagnosed with OA. In the adjusted model, multivariate logistic regression analysis showed that 5 inflammatory indicators (SII, SIRI, MLR, NMLR, NLR) were positively associated with OA prevalence. RCS and threshold analysis showed nonlinear associations between (SII, NMLR, NLR) and OA prevalence. After variable screening, we established an OA risk prediction model with an area under the curve (AUC) of 0.735 (95% CI: 0.726–0.744). Both the decision and calibration curve showed that the model had good clinical significance. The Present study suggests that CBC-derived inflammatory indicators are statistically associated with OA prevalence. Furthermore, MLR and NMLR could be valuable predictors of OA and offer novel perspectives on its assessment and treatment.

Figures