# Prediction of cardiovascular mortality based on iron metabolism in patients with cardiovascular–kidney–metabolic syndrome: development and validation of clinical predictive model

## TL;DR

This study shows that measuring iron and ferritin levels improves predicting heart disease deaths in patients with a syndrome involving heart, kidney, and metabolism issues.

## Contribution

The novel contribution is a validated clinical model that integrates iron metabolism indicators to better predict cardiovascular mortality in CKM patients.

## Key findings

- Serum iron and ferritin levels were significantly associated with cardiovascular mortality in CKM patients.
- Adding iron metabolism indicators improved model performance metrics like Harrell’s C and AIC in both training and validation sets.
- The model demonstrated improved discrimination and reclassification for predicting CVD mortality compared to traditional risk factors.

## Abstract

Iron metabolism is associated with cardiovascular disease (CVD) mortality in patients with cardiovascular–kidney–metabolic (CKM) syndrome. However, its prognostic significance remains unclear. This study was aimed at evaluating incorporation of iron metabolism indicators as a predicting tool for CVD mortality in CKM.

This study was conducted on data of CKM patients collected from the NHANES between 1999 and 2018. The outcome was CVD mortality. Predictors included iron metabolism indicators and traditional risk factors. Cox proportional hazard models were performed to construct the model. The performance of model was assessed through Harrell’s C, NIR, IDI, AIC, BIC and decision curve analysis curves. The nomogram was depicted to predict CVD mortality risk.

Total of 26,609 patients were included, among which 12,808 (48.1%) were male, and the mean age was 51 years. Over a median follow-up of 9.7 years, there were 911 (3.4%) CVD deaths. Among the iron metabolism indicators, iron and ferritin were associated with CVD mortality, and the performance of model improved after adding to the basic model (Harrell’s C, 0.8353 vs 0.8880; NRI, < 0.001; IDI, < 0.001; AIC, 1667.7860 vs 324.8295; BIC, 1740.5520 vs 390.4293) in the training set. These were confirmed in the validation set. The model also demonstrated improved reclassification and discrimination (Harrell’s C, 0.8332 vs 0.8946; NRI, < 0.001; IDI, 1.333; AIC, 769.3590 vs 195.0582; BIC, 834.3515 vs 250.9120).

The incorporation of serum iron and ferritin levels enhanced the predictive value for CVD mortality in CKM patients without pre-existing CVD.