Objective Gout is a chronic inflammatory disease, and cardiovascular disease (CVD) is regarded as one of its complications. The aim of our study was to explore the association between systemic inflammation response index (SIRI) and the risk of CVD in gout. Methods Six cycles of NHANES data were analyzed. Machine learning algorithms were employed to screen covariates, followed by SHAP interpretation to assess variable importance. Participants with gout were stratified by SIRI quartiles, and logistic regression was performed to evaluate CVD risk. RCS were applied to assess nonlinear trends, while discrimination, calibration, and clinical utility were evaluated using ROC, DCA, and calibration curve. Additionally, the Framingham risk score (FRS) model was integrated with SIRI, and model improvement was quantified via net reclassification improvement and integrated discrimination improvement. Results Among 1260 participants with gout, 436 (weighted 28.77%) had CVD comorbidities. A linear positive association was observed between SIRI and CVD risk (P for nonlinear = 0.824), with each 1-unit increase in SIRI corresponding to 29.7% higher CVD risk (OR = 1.297, 95% CI 1.073–1.568). Participants in the highest SIRI quartile Q4 (OR = 2.060, 95% CI 1.141–3.721) exhibited increased CVD risk compared to Q1. The final model demonstrated robust discrimination (AUC = 0.755, 95% CI 0.729–0.783). Incorporating SIRI into the NHANES and clinical datasets improved the discrimination of the FRS model by 5.2% and 1.9%. Conclusion A positive linear association was identified between SIRI and CVD risk in gout patients. The model constructed based on machine learning demonstrated comparable robustness to the FRS model in predicting CVD. These findings provide a theoretical and empirical foundation for early CVD identification, prevention, and management in this population.<Table Float="No" ID="Taba"> <tgroup cols="2"> <colspec align="left" colname="c1" colnum="1" /> <colspec align="left" colname="c2" colnum="2" /> <tbody> <row> <entry align="left" nameend="c2" namest="c1"> Key Points • The positive linear association between the systemic inflammation response index and cardiovascular disease, as well as its subtypes in patients with gout. • SIRI can serve as a valuable complement to the Framingham risk score model. </entry> </row> </tbody> </tgroup> </Table>

Machine learning based on systemic inflammation response index and risk of cardiovascular disease in gout: a retrospective study and clinical validation

期刊论文
出版模式：

开放获取
发表日期：

2025年11月18日

查看全文

Qiang Zhang,
Xuan-hua Yu,
Wei-zhen Zhang,
Xue-bing Lyu,
Hu-han Lin,
Shan-ting Zeng,
Chang-quan Liu,
Hui-juan Huang,
Wei-zhe Deng

摘要

Objective

Gout is a chronic inflammatory disease, and cardiovascular disease (CVD) is regarded as one of its complications. The aim of our study was to explore the association between systemic inflammation response index (SIRI) and the risk of CVD in gout.

Methods

Six cycles of NHANES data were analyzed. Machine learning algorithms were employed to screen covariates, followed by SHAP interpretation to assess variable importance. Participants with gout were stratified by SIRI quartiles, and logistic regression was performed to evaluate CVD risk. RCS were applied to assess nonlinear trends, while discrimination, calibration, and clinical utility were evaluated using ROC, DCA, and calibration curve. Additionally, the Framingham risk score (FRS) model was integrated with SIRI, and model improvement was quantified via net reclassification improvement and integrated discrimination improvement.

Results

Among 1260 participants with gout, 436 (weighted 28.77%) had CVD comorbidities. A linear positive association was observed between SIRI and CVD risk (P for nonlinear = 0.824), with each 1-unit increase in SIRI corresponding to 29.7% higher CVD risk (OR = 1.297, 95% CI 1.073–1.568). Participants in the highest SIRI quartile Q4 (OR = 2.060, 95% CI 1.141–3.721) exhibited increased CVD risk compared to Q1. The final model demonstrated robust discrimination (AUC = 0.755, 95% CI 0.729–0.783). Incorporating SIRI into the NHANES and clinical datasets improved the discrimination of the FRS model by 5.2% and 1.9%.

Conclusion

A positive linear association was identified between SIRI and CVD risk in gout patients. The model constructed based on machine learning demonstrated comparable robustness to the FRS model in predicting CVD. These findings provide a theoretical and empirical foundation for early CVD identification, prevention, and management in this population.

Key Points

• The positive linear association between the systemic inflammation response index and cardiovascular disease, as well as its subtypes in patients with gout.

• SIRI can serve as a valuable complement to the Framingham risk score model.