Triangulating evidence for cardiometabolic Index: ROC cutoff, spline nonlinearity, and explainable machine learning for CVD high risk - Summary - MDSpire

Triangulating evidence for cardiometabolic Index: ROC cutoff, spline nonlinearity, and explainable machine learning for CVD high risk

  • By

  • Haoran Wang

  • Zhiwei Huang

  • Jing Bai

  • Haiqin Yuan

  • Qiaotao Xie

  • Bing He

  • Li Guo

  • Li Wu

  • Dongliang Liu

  • Guirang Zhao

  • Jirui Cai

  • Jin Wang

  • May 20, 2026

  • 0 min

Share

Objective:

To evaluate the cardiometabolic index (CMI) as a tool for identifying high cardiovascular disease (CVD) risk specifically in community-dwelling Chinese adults.

Key Findings:
  • 21% of participants were classified as CVD high risk, with higher prevalence in the high-CMI group (≥0.7).
  • CMI showed modest discrimination for CVD high-risk status (AUC: 0.571), but multivariable models improved this (AUC: 0.642).
  • CMI was associated with higher odds of CVD high risk (OR: 1.31 per 1-unit increase).
  • Machine learning analysis indicated CMI's significant role, ranking it third among important features.
Interpretation:

The CMI is a promising composite marker for assessing cardiovascular risk, demonstrating a nonlinear relationship with CVD risk status, which may have significant implications for risk stratification.

Limitations:
  • The study's cross-sectional design limits causal inference.
  • Data collection relied on self-reported measures, which may introduce bias.
  • Findings may not be generalizable to populations outside the study area.
Conclusion:

Higher CMI is associated with increased CVD risk, supporting its potential as an adjunct screening tool in community settings.

Original Source(s)

Related Content