Prognostic Stratification in Myocardial Infarction Using the Modified CONUT Score: A Multidimensional Biomarker from the MIMIC-IV Cohort

Weixu Yu^1,2Yanting Luo²Qingyuan Cai¹Jin Yang²Wanling Deng²Ruimin Dong^2*Xujing Xie^2*

• ¹Department of Cardiology, Zhongshan Hospital, Xiamen University, Xiamen, China
• ²Department of Cardiology, Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, Guangdong Province, China

Background: Malnutrition is increasingly recognized as a modifiable prognostic factor in myocardial infarction (MI), yet traditional nutritional assessment methods often fail to adequately capture lipid-related atherogenic risk. We introduced an innovative modified Controlling Nutritional Status (mCONUT) score, which replaces total cholesterol with non-HDL cholesterol, aiming to improve the stratification of atherosclerotic risk and to assess its prognostic utility in predicting MI outcomes.Methods: In this retrospective cohort study, we analyzed a total of 3,730 patients diagnosed with MI, extracted from the MIMIC-IV database, and stratified them into Normal, Mild, and Worse groups based on the mCONUT score. After performing 1:1:1 propensity score matching, we selected 993 patients for comparative analysis. Multivariable Cox proportional hazards models, adjusted for clinical and demographic confounders, were employed to evaluate all-cause mortality at 180 days and 1 year.Results: Among 993 MI patients, the median age was 75 years (IQR 68-82) with 58.4% males (n=580). Malnutrition severity (mCONUT ≥6) stratified prognostic risk: Worse group demonstrated progressive mortality increases (180-day: 13.9% vs Normal 7.85%, p = 0.037; 1-year: 31.4% vs 21.2%, p = 0.011), alongside lower BMI, extended hospitalization, reduced hypertension, higher CKD incidence, and diminished revascularization (all p < 0.05). Multivariable analyses confirmed graded mortality risk: 180-day (Model 1: adjusted HR 1.58, p = 0.009; Model 2: HR 1.46, p = 0.031) and 1-year (Model 1: HR 1.61, p = 0.002; Model 2: HR 1.52, p = 0.008). Consistency across subgroups was observed, with heightened vulnerability in males, hypertensives (interaction p = 0.004), diabetes and non-white individuals.Conclusions: The mCONUT score has emerged as a robust multidimensional biomarker for predicting MI prognosis, with worse malnutrition (mCONUT ≥ 6) being significantly associated with a 46-61% elevation in mortality risk, demonstrating a clear linear dose-response relationship. Routine screening and tailored nutritional interventions should be prioritized in modern MI management practices.