Association between the Cholesterol, High-Density Lipoprotein, and Glucose Index and All-Cause Mortality in Critically Ill Patients with Atherosclerotic Cardiovascular Disease: A machine learning-based retrospective cohort study from the MIMIC-IV database

Association between the Cholesterol, High-Density Lipoprotein, and Glucose Index and All-Cause Mortality in Critically Ill Patients with Atherosclerotic Cardiovascular Disease: A machine learning-based retrospective cohort study from the MIMIC-IV database | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Association between the Cholesterol, High-Density Lipoprotein, and Glucose Index and All-Cause Mortality in Critically Ill Patients with Atherosclerotic Cardiovascular Disease: A machine learning-based retrospective cohort study from the MIMIC-IV database Jijun Lan, Meihan Liu, Yihua Wang, Jianguo Feng, Maohua Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9035595/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 15 You are reading this latest preprint version Abstract Background: The cholesterol, high-density lipoprotein, and glucose (CHG) index, as a novel comprehensive marker of lipid and glucose metabolism, has yet to be fully elucidated regarding its prognostic value for all-cause mortality in critically ill ASCVD patients. The present study aims to ascertain the correlation between CHG index and mortality in this population and to identify key predictors using machine learning techniques. Methods: Patients diagnosed with ASCVD were enrolled from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Patients were divided into five groups based on CHG index values. The association between the CHG index and mortality was evaluated using Kaplan-Meier curves, Cox proportional hazards models, restricted cubic splines (RCS), and subgroup analyses. Ten machine learning models were applied to predict mortality risk, with the SHapley Additive exPlanations (SHAP) method used to identify key predictors. Results: 1,959 patients were involved (median age 71.99 years; 52.8% male). Following multivariate adjustment, a one-unit increase in the CHG index was found to be significantly associated with an elevated mortality risk (30-day HR: 1.58, 95% CI: 1.22–2.05; 90-day HR: 1.61, 95% CI: 1.28–2.03). In comparison with patients in the first quintile, those in the fifth quintile demonstrated the highest mortality risk (30-day HR: 2.10, 95% CI: 1.34–3.29; 90-day HR: 2.07, 95% CI: 1.39–3.07). RCS analysis demonstrated a linear positive association between CHG index and mortality. Among machine learning models, the Stacking Classifier demonstrated the most optimal predictive performance for 30-day mortality, with an Area Under the Curve (AUC) of 0.882 in the training set and 0.833 in the test set. The SHAP analysis identified the CHG index as a key predictor. Conclusions: The elevated CHG index demonstrated a direct correlation with an augmented mortality risk in critically ill ASCVD patients. The CHG index has the potential to be a valuable predictor for mortality risk assessment in this population. Clinical trial number: not applicable. Atherosclerotic cardiovascular disease All-Cause mortality CHG Machine learning MIMIC-IV Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 1. Introduction Cardiovascular disease (CVD) is a major cause of mortality on a global scale 1 . Atherosclerotic cardiovascular disease (ASCVD), comprising coronary artery disease and ischemic stroke, represents the predominant subtype of CVD and imposes a substantial burden on global healthcare systems 2 . Critically ill patients with ASCVD admitted to the intensive care unit (ICU) constitute a high-risk population with poor prognosis, and metabolic dysregulation frequently complicates their clinical course 3 . Recent studies have highlighted a strong link between metabolic disturbances—particularly in lipid and glucose regulation—and mortality risk among critically ill populations 4 , 5 . Dysfunction of glucose and lipid metabolism has been shown to be associated with disease severity and poor prognosis in ASCVD patients 6 . Insulin resistance (IR) stands at the core of metabolic dysregulation and is increasingly recognized as a key contributor to poor cardiovascular outcomes 7 . Although the hyperinsulinemic-euglycemic clamp remains the reference standard for assessing insulin sensitivity, its complexity and time demands restrict its use in everyday clinical practice 8 . In response, simpler surrogate indices have emerged—such as the triglyceride-glucose (TyG) index and the metabolic score for insulin resistance (METS-IR)—both of which have shown promise in predicting mortality among critically ill patients 9 . Large epidemiological studies have further linked the TyG index to cardiovascular mortality 10 . That said, these markers rely mainly on triglyceride (TG) and fasting blood glucose (FBG) levels, leaving cholesterol metabolism largely unaccounted for. Of note, low high-density lipoprotein (HDL) cholesterol often accompanies IR, and their combined presence amplifies cardiovascular risk 11 . The cholesterol, high-density lipoprotein, and glucose (CHG) index, which brings together total cholesterol (TC), HDL, and FBG, has been put forward as a more integrated measure of both lipid and glucose status 12 . Whether this index holds prognostic value for critically ill patients with established ASCVD, however, remains an open question. Machine learning approaches, capable of capturing complex, nonlinear relationships among variables, are regarded as valuable tools for forecasting clinical outcomes 13 . Various machine learning models, including random forest (RF), extreme gradient boosting (XGBoost), and gradient boosting (GB), have been developed and validated with high accuracy for clinical prediction tasks 14 . Their application to mortality prediction in ICU settings has grown considerably, and several studies suggest they may outperform established tools like the Acute Physiology and Chronic Health Evaluation (APACHE) and Sequential Organ Failure Assessment (SOFA) scores 15 . Furthermore, the SHapley Additive exPlanations (SHAP) method addresses the "black box" concern often associated with machine learning in clinical settings by quantifying the contribution of each predictor to individual predictions 16 . This interpretable machine learning approach offers a valuable framework for assessing the prognostic relevance of novel metabolic indices in critically ill populations. Using data drawn from the MIMIC-IV database, this study set out to assess whether the CHG index correlates with all-cause mortality in critically ill patients diagnosed with ASCVD. A secondary aim was to build interpretable machine learning models capable of identifying the key factors of prognosis in this vulnerable population. 2. Methods Data source We retrieved data for this retrospective cohort study from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.1). This publicly accessible critical care resource contains de-identified electronic health records from over 360,000 patients and 546,000 hospital admissions at Beth Israel Deaconess Medical Center (Boston, MA, USA) between 2008 and 2022, including 94,458 ICU stays 17 . Patient privacy is safeguarded through the removal of all personal identifiers. The original data collection was approved by the institutional review boards of Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology, with a waiver of informed consent granted due to the anonymized nature of the data. Author Ji-Jun Lan completed the National Institutes of Health (NIH) web-based training course on protecting human research participants (certification number: 69132229) and obtained access to the database. Study population The diagnosis of ASCVD was established through manual review of International Classification of Diseases (ICD), Ninth and Tenth Revisions (ICD-9 and ICD-10) (detailed in Supplementary Material Table S1 ) 18 . Patients were excluded from participation according to the following criteria: (1) age < 18 years; (2) multiple ICU admissions for ASCVD (only the first admission was included); (3) ICU length of stay < 24 hours; (4) missing data on glucose, TC, or HDL; (5) presence of severe liver disease, metastatic solid tumor, or malignant cancer; (6) outliers of CHG index, defined as values below Q1 − 3IQR or above Q3 + 3IQR, where Q1 and Q3 respectively represent the first and third quartiles, and IQR stands for the interquartile range. Figure 1 illustrates the patient selection process. Data acquisition The data extraction process was executed utilizing the structured query language (SQL) via Navicat Premium (version 17), with a particular emphasis placed on six distinct categories: (1) Demographics: age, sex, race; (2) Vital signs: heart rate, weight, temperature, respiratory rate (RR), systolic blood pressure (SBP), diastolic blood pressure (DBP), peripheral oxygen saturation (SpO2); (3) Clinical severity scores: Glasgow Coma Score (GCS), Acute Physiology Score III (APSIII); (4) Laboratory parameters: blood urea nitrogen (BUN), creatinine, potassium, sodium, anion gap, bicarbonate, white blood cell (WBC), red blood cell (RBC), platelets, hemoglobin, red blood cell distribution width (RDW), hematocrit (HCT), glucose, TG, TC, HDL, low-density lipoprotein cholesterol (LDL); (5) Comorbidities: sepsis, congestive heart failure, dementia, chronic pulmonary disease, rheumatic disease, diabetes, renal disease, mild liver disease, paraplegia, hypertension; (6) Medications: vasopressin, aspirin, clopidogrel, warfarin, and statins. We computed the CHG index according to the formula: CHG index = Ln [TC (mg/dL) × FBG (mg/dL) / (2 × HDL (mg/dL))] 19 . Vital signs and laboratory parameters were based on the first measurement within 24 hours of ICU admission to capture baseline physiological status. Clinical severity scores were calculated based on the mean values of measurements taken on the first day of ICU admission. The utilization of pharmaceuticals was defined as the record of whether the patient received the specified medications during their ICU stay. The main endpoint was designated as 30-day all-cause mortality, whereas the secondary endpoint was designated as 90-day all-cause mortality. Statistical analysis The CHG index was then grouped into quintiles: Q1 (3.44–4.98), Q2 (4.98–5.25), Q3 (5.25–5.50), Q4 (5.50–5.84), and Q5 (5.84–7.53). The continuous variables were reported as the mean ± standard deviation (SD) or the median (interquartile range (IQR)), and the group comparisons were performed using either the t-test or the Mann-Whitney U test. Categorical variables were expressed as frequencies (percentages) and compared using chi-square or Fisher's exact test. Multicollinearity was detected using the Variance Inflation Factor (VIF). Covariates with a VIF greater than 5 were omitted from multivariate analyses. Kaplan-Meier curves were employed to visualize survival differences, and Cox proportional hazards regression was applied with three adjustment models: The first model was a crude model; the second model was adjusted for age, sex and race; the third model was further adjusted for comorbidities, vital signs, medications, laboratory parameters and severity scores. The selection of covariates was based on clinical relevance and univariate significance, with multicollinearity controlled (VIF < 5). Restricted cubic splines (RCS) based on Cox proportional hazards model (Model 3) were used, along with histograms, to evaluate the dose-response relationship between CHG index and all-cause mortality at different prognostic time points. Subgroup and sensitivity analyses were conducted to evaluate result robustness. The missing data were then imputed using multiple imputation by chained equations (MICE), which generated five imputed datasets with pooled estimates calculated via Rubin's rules 20 , 21 . Machine learning model development and performance assessment In the context of machine learning-based prediction, the dataset was initially partitioned into a training and testing set at an 8:2 ratio, with the sets being randomly selected. The Boruta algorithm was subsequently employed on the training set to rank the importance of predictive features for 30-day mortality 22 . A total of nine base models were developed, including logistic regression (LR), RF, XGBoost, light gradient boosting machine (LightGBM), linear discriminant analysis (LDA), GB, adaptive boosting (AdaBoost), naive bayes (NB), and quadratic discriminant analysis (QDA). A stacking ensemble was also built, combining logistic regression (LR), linear discriminant analysis (LDA), XGBoost, RF, and adaptive boosting (AdaBoost) as base learners, with LR again serving as the meta-learner. Each model underwent hyperparameter tuning via Bayesian optimization—100 iterations per model—with the area under the receiver operating characteristic curve (AUC) set as the objective function to maximize 23 . Performance was evaluated using five repetitions of 10-fold stratified cross-validation. Several metrics were used to assess model performance, including AUC, accuracy, specificity, sensitivity, and F1 score. The calibration curves were utilized to assess the concordance between the observed and predicted outcomes. Decision curve analysis (DCA) was employed to evaluate the clinical net benefit. The best-performing model was further interpreted using SHAP for determining crucial predictive factors. All data analyses were performed using Python 3.12.4 and R 4.5.0 software. A two-tailed P value < 0.05 was considered statistically significant. 3. Results Baseline characteristics The study population comprised 1,959 critically ill patients with ASCVD were included, with a median age of 71.99 years (IQR: 61.33-82.63) and 52.8% (n=1,034) being male. The statistics of the missing ratio before missing value interpolation are shown in Supplementary Material Table S2. The median CHG index was 5.38 (IQR: 5.05-5.73). The mortality rates within 30 and 90 days of ICU admission were 16.7% and 21.0%, respectively. Analysis revealed that non-survivors were older [80.74 (69.25-87.89) vs. 70.29 (59.83-81.17) years], had higher CHG index levels [5.44 (5.15-5.88) vs. 5.37 (5.04-5.71)], higher APSIII scores [48 (34-61) vs. 33 (25-43)], lower GCS scores [14.67 (12.46-15.00) vs. 14.84 (13.99-15.00)], higher heart rate, RR, WBC count, glucose, BUN, and RDW, lower RBC count, hemoglobin, TC, and LDL, higher prevalence of comorbidities including sepsis, congestive heart failure, and renal disease, higher vasopressin use, and lower use of aspirin, warfarin, and statins (all P<0.05). The baseline characteristics of the two groups are exhibited in Table 1 and Supplementary Material Table S3. Table 1 Baseline characteristics according to 30-day mortality Variable Overall(n=1959) Survivors(n=1632) Non-survivors(n=327) P value CHG index 5.38 (5.05-5.73) 5.37 (5.04-5.71) 5.44 (5.15-5.88) <0.001 Demographic variables Age (years) 71.99 (61.33-82.63) 70.29 (59.83-81.17) 80.74 (69.25-87.89) <0.001 Male (n, %) 1034 (52.8) 882 (54.0) 152 (46.5) 0.015 Race (White) 1031 (52.6) 881 (54.0) 150 (45.9) 0.009 Vital signs Heart rate (b/min) 80.00 (69.00-93.00) 79.00 (69.00-92.00) 85.00 (71.00-97.50) <0.001 RR (b/min) 18.00 (15.00-22.00) 18.00 (15.00-22.00) 19.00 (16.00-23.00) 0.003 SBP (mmHg) 140.00 (122.00-156.00) 140.00 (123.00-156.00) 137.00 (117.00-155.25) 0.05 DBP (mmHg) 77.00 (66.00-90.00) 78.00 (67.00-91.00) 74.00 (61.75-86.00) <0.001 Spo2 (b/min) 98.00 (96.00-99.00) 97.00 (96.00-99.00) 98.00 (96.00-100.00) 0.009 Weight (kg) 78.00 (65.50-92.70) 78.70 (66.68-93.85) 73.30 (59.95-86.75) <0.001 Temperature (℃) 36.78 (36.50-37.00) 36.78 (36.50-37.00) 36.72 (36.44-37.06) 0.129 Clinical severity scores GCS 14.83 (13.79-15.00) 14.84 (13.99-15.00) 14.67 (12.46-15.00) 0.003 APSIII 35.00 (26.00-46.00) 33.00 (25.00-43.00) 48.00 (34.00-61.00) <0.001 Laboratory parameters BUN (mg/dL) 17.00 (13.00-25.00) 16.00 (12.00-23.00) 22.00 (16.00-34.00) <0.001 Creatinine (mg/dL) 0.90 (0.80-1.20) 0.90 (0.70-1.20) 1.10 (0.80-1.60) <0.001 Potassium (mEq/L) 4.10 (3.70-4.40) 4.00 (3.70-4.40) 4.20 (3.80-4.70) <0.001 Sodium (mEq/L) 139.00 (137.00-141.00) 139.00 (137.00-141.00) 139.00 (137.00-142.00) 0.282 Anion gap (mmol/L) 14.00 (12.00-16.00) 14.00 (12.00-16.00) 15.00 (13.00-18.00) <0.001 Bicarbonate (mmol/L) 23.00 (21.00-25.00) 23.00 (21.00-25.00) 22.00 (19.00-24.00) <0.001 WBC (×10 3 /µL) 10.00 (7.70-13.00) 9.60 (7.50-12.50) 11.90 (9.20-15.60) <0.001 RBC (×10 6 /µL) 4.17 (3.70-4.61) 4.21 (3.75-4.63) 3.99 (3.49-4.46) <0.001 Platelets (×10 3 /µL) 215.00 (171.00-266.00) 216.00 (173.00-265.00) 213.00 (161.00-270.50) 0.29 Hemoglobin (g/dL) 12.50 (11.00-13.80) 12.60 (11.10-13.90) 11.80 (10.50-13.30) <0.001 RDW (%) 13.70 (13.00-14.60) 13.60 (12.90-14.50) 14.20 (13.40-15.50) <0.001 HCT (%) 38.10 (33.80-41.50) 38.40 (34.20-41.60) 36.10 (32.40-40.55) <0.001 Glucose (mg/dL) 125.00 (103.00-162.00) 122.00 (100.75-154.00) 145.00 (119.00-196.00) <0.001 TG (mg/dL) 102.00 (74.00-143.00) 103.00 (75.75-145.00) 94.00 (68.50-135.50) 0.014 TC (mg/dL) 156.00 (127.00-189.00) 159.50 (129.00-191.00) 140.00 (112.00-172.00) <0.001 HDL (mg/dL) 45.00 (36.00-57.00) 46.00 (37.00-57.00) 45.00 (35.00-54.00) 0.022 LDL (mg/dL) 83.00 (60.00-113.00) 86.00 (62.00-114.00) 69.00 (51.00-101.00) <0.001 Comorbidities Sepsis (%) 607 (31.0) 421 (25.8) 186 (56.9) <0.001 Congestive heart failure (%) 625 (31.9) 483 (29.6) 142 (43.4) <0.001 Dementia (%) 110 (5.6) 74 (4.5) 36 (11.0) <0.001 Chronic pulmonary disease (%) 327 (16.7) 264 (16.2) 63 (19.3) 0.198 Rheumatic disease (%) 48 (2.5) 39 (2.4) 9 (2.8) 0.848 Diabetes (%) 659 (33.6) 533 (32.7) 126 (38.5) 0.047 Renal disease (%) 380 (19.4) 289 (17.7) 91 (27.8) <0.001 Mild liver disease (%) 48 (2.5) 40 (2.5) 8 (2.4) 1 Paraplegia (%) 961 (49.1) 770 (47.2) 191 (58.4) <0.001 Hypertension (%) 1525 (77.8) 1254 (76.8) 271 (82.9) 0.02 Medications Vasopressin (%) 55 (2.8) 16 (1.0) 39 (11.9) <0.001 Aspirin (%) 1314 (67.1) 1119 (68.6) 195 (59.6) 0.002 Clopidogrel (%) 357 (18.2) 311 (19.1) 46 (14.1) 0.04 Warfarin (%) 190 (9.7) 177 (10.8) 13 (4.0) <0.001 Statins (%) 1339 (68.4) 1164 (71.3) 175 (53.5) <0.001 Data are expressed as median (IQR), or n (%). CHG index: Cholesterol, high-density lipoprotein, and glucose index; RR: respiratory rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; Spo2: peripheral oxygen saturation; GCS: Glasgow Coma Score; APSIII: Acute Physiology Score III; BUN: blood urea nitrogen; WBC: white blood cell; RBC: red blood cell; RDW: red blood cell distribution width; HCT: hematocrit; TG: triglyceride; TC: total cholesterol; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol Association between CHG index and all-cause mortality Table 2 The association between the CHG index and all-cause mortality Variables Model 1 Model 2 Model 3 HR (95%CI) P HR (95%CI) P HR (95%CI) P 30-day mortality CHG (per 1 unit) 1.62 (1.33-1.97) <0.001 2.24 (1.82-2.76) <0.001 1.58 (1.22-2.05) <0.001 Quintile groups Group 1 1.00 (Reference) 1.00 (Reference) 1.00 (Reference) Group 2 1.55 (1.06-2.26) 0.023 1.61 (1.10-2.35) 0.014 1.75 (1.19-2.59) 0.005 Group 3 1.37 (0.93-2.02) 0.114 1.54 (1.04-2.27) 0.03 1.54 (1.03-2.30) 0.037 Group 4 1.52 (1.04-2.22) 0.032 1.88 (1.28-2.76) 0.001 1.76 (1.16-2.67) 0.008 Group 5 2.13 (1.49-3.05) <0.001 3.12 (2.16-4.51) <0.001 2.10 (1.34-3.29) 0.001 P for trend 1.62 (1.27-2.07) <0.001 2.22 (1.73-2.87) <0.001 1.60 (1.16-2.20) 0.004 90-day mortality CHG (per 1 unit) 1.53 (1.28-1.83) <0.001 2.13 (1.77-2.58) <0.001 1.61 (1.28-2.03) <0.001 Quintile groups Group 1 1.00 (Reference) 1.00 (Reference) 1.00 (Reference) Group 2 1.49 (1.07-2.08) 0.019 1.56 (1.12-2.18) 0.009 1.70 (1.21-2.40) 0.002 Group 3 1.38 (0.98-1.94) 0.061 1.57 (1.12-2.20) 0.01 1.61 (1.13-2.29) 0.008 Group 4 1.48 (1.06-2.07) 0.021 1.86 (1.33-2.61) <0.001 1.78 (1.23-2.58) 0.002 Group 5 1.96 (1.42-2.70) <0.001 2.91 (2.10-4.04) <0.001 2.07 (1.39-3.07) <0.001 P for trend 1.53 (1.24-1.91) <0.001 2.12 (1.69-2.66) <0.001 1.60 (1.20-2.12) 0.001 HR: Hazard Ratio, CI: Confidence Interval. Model 1: unadjusted; Model 2: adjusted for age, gender, race; Model 3: adjusted for Model 2 plus Sepsis, Congestive heart failure, dementia, diabetes, renal disease, paraplegia, hypertension, heart rate, DBP, RR, weight, BUN, creatinine, potassium, anion gap, bicarbonate, WBC, RBC, RDW, LDL, vasopressin, aspirin, clopidogrel, warfarin, statins, GCS, APSIII. Kaplan-Meier survival analysis showed significant differences in survival rates among CHG index quintiles, with patients in the highest CHG index groups demonstrating higher 30-day and 90-day all-cause mortality than those in the lowest groups (log-rank test, P = 0.00079 and P = 0.0012, respectively); the detailed findings are illustrated in Figure 2. The results of multivariate Cox proportional hazards regression analysis are shown in Table 2, and the selection criteria for covariates in Model 3 are provided in Supplementary Materials Table S4-6. In the fully adjusted model (Model 3), a 1-unit increase in the CHG index showed a 58% increased risk of 30-day all-cause mortality (HR=1.58, 95% CI: 1.22-2.05) and a 61% higher risk of 90-day all-cause mortality (HR=1.61, 95% CI: 1.28-2.03) (both P<0.001). For 30-day mortality, utilizing Q1 group as the reference, the hazard ratios (HR) and 95% confidence intervals (CI) for Q5 group in Model 1, Model 2, and Model 3 were 2.13 (1.49-3.05), 3.12 (2.16-4.51), and 2.10 (1.34-3.29), respectively. For 90-day mortality, the HR and 95% CI for Q5 group in Model 1, Model 2, and Model 3 were 1.96 (1.42-2.70), 2.91 (2.10-4.04), and 2.07 (1.39-3.07), respectively. The trend test indicated a dose-response relationship between CHG index and all-cause mortality risk (P = 0.004 and P = 0.001, respectively). RCS regression analysis confirmed a linear dose-response relationship between CHG index and all-cause mortality risk (Figure 3). In the fully adjusted model (Model 3), the CHG index exhibited a linear correlation with both 30-day mortality (P-nonlinear=0.250) and 90-day mortality (P-nonlinear=0.076). Subgroup and sensitivity analyses The results of the subgroup analysis are shown in Figure 4. Subgroup analysis revealed a significant interaction between the CHG index and 30-day mortality risk depending on clopidogrel use (P for interaction = 0.049) and statin use (P for interaction = 0.008). Among patients not receiving clopidogrel, a higher CHG index was associated with increased mortality risk (HR = 1.38, 95% CI: 1.09–1.74); this association was not observed in those on clopidogrel (HR = 0.79, 95% CI: 0.40–1.53). Similarly, the association was stronger in patients not using statins (HR=1.69, 95% CI: 1.19-2.39) compared with those using statins (HR=1.14, 95% CI: 0.84-1.56). For 90-day mortality, significant interactions were found across age (P for interaction=0.033), aspirin use (P for interaction=0.030), and statin use (P for interaction=0.013) subgroups. The CHG index demonstrated a significant association with mortality risk in patients aged ≥65 years (HR=1.35, 95% CI: 1.08-1.70), but not in those aged <65 years (HR=1.47, 95% CI: 0.95-2.27). We conducted multiple sensitivity analyses to evaluate the stability of our results. Firstly, following the elimination of missing values, the outcomes of the Cox proportional hazards regression analysis were found to be in alignment with the primary analysis (Supplementary Material Table S7). Secondly, utilizing the technique of logistic regression, the correlation between the CHG index and all-cause mortality persisted in alignment with the primary findings (Supplementary Material Table S8). Model performance evaluation Feature importance ranking was performed on the training set using the Boruta algorithm (Figure 5). We identified the top 16 important features, including APSIII, vasopressin use, sepsis, GCS score, WBC count, age, glucose, BUN, weight, temperature, CHG index, bicarbonate, SpO2, HCT, TC, and statin use, and subsequently constructed 10 machine learning models for predicting 30-day mortality. The performance metrics of each model on the training and test sets are presented in Supplementary Material Table S9, and the optimal hyperparameter settings for each model are shown in Supplementary Material Table S10. On the training set, RF achieved the highest AUC (0.912, 95% CI: 0.898-0.926), followed by GB (AUC=0.907) and Stacking Classifier (AUC=0.882). However, on the test set, Stacking Classifier demonstrated the most stable performance with an AUC of 0.833 (95% CI: 0.796-0.870), which was comparable to its training set result (AUC difference=0.049), indicating good generalization ability (Figure 6A-B). In contrast, RF and GB showed decreased AUC values of 0.821 and 0.795 on the test set (AUC differences of 0.091 and 0.112, respectively), suggesting some degree of overfitting. Considering multiple performance metrics, Stacking Classifier achieved the highest specificity (0.911), accuracy (0.847), and F1-score (0.531) on the test set, while maintaining a high AUC value, demonstrating the best discriminative ability and stability. Therefore, Stacking Classifier was selected as the optimal model. We further evaluated the clinical utility and calibration performance of the Stacking Classifier model. DCA showed that the model delivered a positive net benefit across a wide range of threshold probabilities (0.2-0.8) in both training and test sets, outperforming the "treat-all" and "treat-none" strategies (Figure 6C-D). Calibration curves (Figure 6E-F) revealed good agreement between predicted probabilities and observed outcomes in both the training and test sets with Brier scores of 0.099 and 0.103 for the training and test sets, respectively. SHAP-based interpretation of the optimal model SHAP feature importance analysis (Figure 7A) identified the APSIII score as the most influential predictor of 30-day mortality, with the highest mean absolute SHAP value (0.045). This was followed by sepsis (0.041), age (0.037), WBC count (0.029), and statin use (0.024). The CHG index is considered the primary indicator of this study, with a mean absolute SHAP value of 0.011, placing it 8th among all features and indicating moderate predictive importance. The SHAP summary plot analysis was conducted, indicating that the use of statins was correlated with negative SHAP values. This finding suggests that statin use may function as a protective factor in terms of mortality. Conversely, patients exhibiting elevated APSIII scores, sepsis, advanced age, elevated WBC counts, and elevated CHG indices demonstrated positive SHAP values, signifying their role as independent risk factors for mortality. As illustrated in Figure 7B, the dependence plots provided additional information on the impact of each predictor on model output, and on potential interaction effects. 4. Discussion This study set out to ascertain the association between the CHG index and the 30-day and 90-day all-cause mortality rates of critically ill patients with ASCVD. The present findings indicate that an augmented CHG index is independently correlated with elevated mortality risk. Kaplan-Meier survival analysis showed statistically significant variations in survival probability between the CHG index quintiles. Multivariate Cox proportional hazards regression was employed to confirm that the CHG index continued to independently predict mortality following adjustment for potential confounders. Moreover, RCS analysis demonstrated a linear dose-response association of CHG index with mortality. It is noteworthy that the CHG index serves as a critical predictor in machine learning models, underscoring its potential application in risk stratification within clinical practice. Lipid and glucose metabolic disturbances are fundamental pathophysiological features of ASCVD. A systematic review covering 208 studies indicated that elevated total cholesterol is associated with an increased risk of premature coronary heart disease 24 . Regarding high-density lipoprotein cholesterol, epidemiological evidence has revealed a non-linear association with mortality risk, where reduced concentrations suggest a poor prognosis in patients with ASCVD 25 . Furthermore, a systematic review and meta-analysis confirmed that hyperglycaemia can independently predict mortality in patients with acute coronary syndromes 26 . These metabolic imbalances can occur simultaneously, collectively increasing the risk of CVD mortality. Notably, metabolic syndrome, characterized by the coexistence of dyslipidaemia and hyperglycaemia, is often significantly associated with elevated CVD mortality 27 . This epidemiological evidence highlights the importance of lipid and glucose metabolism in determining cardiovascular prognosis, providing a robust theoretical rationale for integrating these parameters into a unified prognostic index. It has been demonstrated that the CHG index, a composite indicator of cholesterol, high-density lipoprotein, and glucose metabolism parameters, exhibits a correlation with adverse outcomes across a range of conditions, including metabolic syndrome, stroke, and CVD. In the context of metabolic syndrome populations, a U-shaped nonlinear relationship has been documented between the CHG index and mortality risk. This relationship suggests that both low and elevated levels of the index correlate with higher cardiovascular and all-cause death risk 28 . The prognostic utility of the CHG index is further demonstrated in patients with metabolic dysfunction-associated steatotic liver disease, where it has been associated with mortality risk, particularly among individuals under 60 years of age and with lean body composition 29 . Within the context of the CHG index, it has been demonstrated that there exists an independent association with the occurrence of stroke in patients diagnosed with early cardiovascular-renal-metabolic syndrome. Furthermore, the index has been shown to be associated with an elevated risk of cardiovascular mortality and mortality from any cause in patients diagnosed with calcific aortic valve stenosis 30 , 31 . The index also functions as a novel predictor of in-stent restenosis following percutaneous coronary intervention, thereby emphasising its significance in the domain of interventional cardiology 32 . It is noteworthy that comparative analyses indicate that the CHG index performs comparably to the TyG index in predicting cardiovascular metabolic disease risk 33 . The present study is the first to specifically ascertain the relationship between CHG index and short-term mortality in critically ill patients with ASCVD, thereby highlighting the potential value of CHG index as a severity and prognosis marker. The pathophysiological mechanisms linking adverse outcomes in critically ill patients with ASCVD involve complex interactions among lipid metabolism, glucose homeostasis, and inflammatory responses. Elevated TC promotes the formation of foam cells by enhancing macrophage uptake of oxidized lipoproteins to a certain extent, thereby contributing to the development of atherosclerotic lesions 34 , 35 . Furthermore, elevated blood glucose upregulates inflammatory markers and increases the production of reactive oxygen species, leading to vascular dysfunction. Secondly, it accelerates the formation of advanced glycation end products, activating pro-inflammatory signaling cascades and consequently promoting atherosclerosis 36 , 37 . In contrast, HDL exerts protective effects by facilitating reverse cholesterol transport, thereby reducing lipid accumulation in the arterial wall 38 . Concurrently, HDL particles possess anti-inflammatory and antioxidant properties that counteract the effects of oxidized lipoproteins 39 , 40 . When these metabolic disturbances converge, a pro-inflammatory environment characterized by elevated cytokines and adhesion molecules is established, accelerating plaque progression and increasing the risk of rupture 41 , 42 . Consequently, in critically ill patients, these metabolic disturbances and inflammatory responses may further deteriorate an already compromised cardiovascular status. Machine learning has found increasingly widespread applications in medicine, demonstrating significant advantages in tasks ranging from disease diagnosis to prognostic prediction 43 . Traditional prognostic assessment tools used in the ICU, although widely applied in clinical practice, have limitations in accuracy 44 . In contrast, machine learning models often achieve superior predictive performance, as they are better at handling complex, nonlinear relationships among variables compared to traditional indicator assessments 45 , 46 . In the construction of machine learning models, ensemble methods enhance model stability and generalization capabilities by integrating outputs from multiple base models, thereby enabling more reliable predictions 47 , 48 . Furthermore, the "black-box" nature of machine learning models has remained a barrier to their clinical application 49 . The SHAP method, grounded in Shapley values from game theory, provides an interpretability framework for machine learning models by quantifying the contribution of each feature to model predictions 50 . In our study, SHAP analysis identified the CHG index as a significant predictor of mortality in critically ill patients with ASCVD. This finding corroborates the results of the Cox regression analysis, offering machine learning-based support for the prognostic value of the CHG index. The present study is subject to several limitations. Firstly, as a retrospective cohort study, it was not possible to completely exclude selection bias and residual confounding. Secondly, the study data were derived from a single center (MIMIC-IV database), and the external validity of the models needs to be further validated in independent cohorts with diverse ethnicities, geographic regions and healthcare systems. Thirdly, as a retrospective cohort study, it was not possible to make definitive conclusions regarding causal relationships. Based on the findings of this study, future research can be explored in the following directions. First, prospective multicenter cohort studies should be conducted to validate the prognostic value of the CHG index across diverse ethnicities, geographic regions, and healthcare systems, and to evaluate the utility of CHG index-based risk stratification strategies in guiding clinical decision-making. Second, time-series analysis methods should be integrated to construct dynamic CHG index monitoring models, systematically assessing the association between metabolic indicator trends and prognosis, and exploring the feasibility of dynamic risk early warning. Third, multi-omics technologies (such as metabolomics, proteomics, and genomics) should be employed to deeply elucidate the molecular mechanisms underlying the association between the CHG index and adverse outcomes, identifying potential therapeutic intervention targets. Fourth, randomized controlled trials or observational intervention studies should be conducted to evaluate whether metabolic management to reduce the CHG index can improve clinical outcomes in critically ill patients with ASCVD, providing evidence-based medical evidence for the CHG index as a potential therapeutic target. 5. Conclusion Elevated CHG index was found to have an independent linear relationship with greater all-cause mortality among critically ill ASCVD patients. Machine learning analysis further confirms CHG index serves as a critical mortality predictor in this cohort. This novel metabolic index holds promise as an effective indicator for patient risk assessment and treatment decisions. Abbreviations AdaBoost adaptive boosting APSIII Acute Physiology Score III APACHE Acute Physiology and Chronic Health Evaluation ASCVD atherosclerotic cardiovascular disease AUC area under the receiver operating characteristic curve Boruta Boruta algorithm BUN blood urea nitrogen CHG Cholesterol, High-Density Lipoprotein, and Glucose CI confidence interval CVD cardiovascular disease DCA decision curve analysis DBP diastolic blood pressure FBG fasting blood glucose GB gradient boosting GCS Glasgow Coma Score HCT hematocrit HDL high-density lipoprotein cholesterol HR hazard ratio ICD International Classification of Diseases ICU intensive care unit IQR interquartile range IR insulin resistance LDA linear discriminant analysis LDL low-density lipoprotein cholesterol LightGBM light gradient boosting machine LR logistic regression MASLD metabolic dysfunction-associated steatotic liver disease METS-IR metabolic score for insulin resistance MICE multiple imputation by chained equations MIMIC-IV Medical Information Mart for Intensive Care IV NB naive bayes NIH National Institutes of Health PREVENT American Heart Association's PREVENT Equations QDA quadratic discriminant analysis RBC red blood cell RCS restricted cubic splines RF random forest RDW red blood cell distribution width RR respiratory rate SBP systolic blood pressure SD standard deviation SHAP SHapley Additive exPlanations SOFA Sequential Organ Failure Assessment SpO2 peripheral oxygen saturation SQL structured query language TC total cholesterol TG triglyceride TyG triglyceride-glucose VIF variance inflation factor WBC white blood cell XGBoost extreme gradient boosting Declarations Acknowledgements Not applicable. Author Contributions JJL, MHL, and YHW came up with the article concept and design ideas, contributed to the methodology, and wrote the initial draft. JJL performed the data curation and led the formal analysis and visualization, with MHL and YHW contributing equally to the formal analysis and visualization; MHL also performed the validation. MHW acquired the funding, administered the project, and supervised the study. JGF carried out the investigation and provided the resources and software. MHW and JGF were pivotal in revising the manuscript. The final manuscript was reviewed and approved by all study contributors. Funding The work was supported by the Sichuan Province Science and Technology Support Program 2022YFS0632. Data availability No datasets were generated or analysed during the current study. Ethics approval and consent to participate Use of the MIMIC-IV database was approved by the review committees of Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center. As the data are publicly available, ethical approval and informed consent were waived for this study. Consent for publication Not applicable. Competing interests The authors report no conflicts of interest in this work. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Martin SS, Aday AW, Allen NB, et al. 2025 Heart Disease and Stroke Statistics: A Report of US and Global Data From the American Heart Association. Circulation . 2025;151(8):e41-e660. doi:10.1161/CIR.0000000000001303 GBD 2021 Diseases and Injuries Collaborators. Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. Lancet . 2024;403(10440):2133-2161. doi:10.1016/S0140-6736(24)00757-8 Wang H, Jiang G, Qu X, Zhou J. Association of triglyceride glucose index with mortality in critically ill patients with atherosclerotic cardiovascular disease: analysis of the MIMIC-IV database. Eur J Med Res . Published online February 5, 2026. doi:10.1186/s40001-026-03994-w Plummer MP, Bellomo R, Cousins CE, et al. Dysglycaemia in the critically ill and the interaction of chronic and acute glycaemia with mortality. Intensive Care Med . 2014;40(7):973-980. doi:10.1007/s00134-014-3287-7 Liu Y, Fu H, Wang Y, et al. Association between the low-density lipoprotein to high-density lipoprotein ratio and prognosis in critically ill intracerebral hemorrhage patients: a retrospective cohort study from the MIMIC-IV database. Lipids Health Dis . 2025;24(1):36. doi:10.1186/s12944-025-02459-z Wang F, Guo Y, Tang Y, et al. Combined assessment of stress hyperglycemia ratio and glycemic variability to predict all-cause mortality in critically ill patients with atherosclerotic cardiovascular diseases across different glucose metabolic states: an observational cohort study with machine learning. Cardiovasc Diabetol . 2025;24(1):199. doi:10.1186/s12933-025-02762-0 Lin Z, Yuan S, Li B, et al. Insulin-based or non-insulin-based insulin resistance indicators and risk of long-term cardiovascular and all-cause mortality in the general population: A 25-year cohort study. Diabetes Metab . 2024;50(5):101566. doi:10.1016/j.diabet.2024.101566 Singh B, Saxena A. Surrogate markers of insulin resistance: A review. World J Diabetes . 2010;1(2):36-47. doi:10.4239/wjd.v1.i2.36 Ouyang Q, Zhang C, Zhang H, Xiong T, Chen Y, Zhang P. Evaluation of four insulin resistance surrogate indicators for predicting short-term mortality in critically ill patients with ischemic stroke: a nationwide retrospective cohort study. Cardiovasc Diabetol . 2025;24(1):409. doi:10.1186/s12933-025-02956-6 He G, Zhang Z, Wang C, et al. Association of the triglyceride-glucose index with all-cause and cause-specific mortality: a population-based cohort study of 3.5 million adults in China. Lancet Reg Health West Pac . 2024;49:101135. doi:10.1016/j.lanwpc.2024.101135 Fahed G, Aoun L, Bou Zerdan M, et al. Metabolic Syndrome: Updates on Pathophysiology and Management in 2021. Int J Mol Sci . 2022;23(2):786. doi:10.3390/ijms23020786 Mo D, Zhang P, Zhang M, Dai H, Guan J. Cholesterol, high-density lipoprotein, and glucose index versus triglyceride-glucose index in predicting cardiovascular disease risk: a cohort study. Cardiovasc Diabetol . 2025;24(1):116. doi:10.1186/s12933-025-02675-y Żerdziński K, Janiec J, Jóźwik K, Łajczak P, Krzych ŁJ. Artificial Intelligence in Intensive Care: An Overview of Systematic Reviews with Clinical Maturity and Readiness Mapping. J Clin Med . 2025;15(1):185. doi:10.3390/jcm15010185 Sun H, Kang M, Zhang H, Jia J, Wang Q. Machine Learning for Predicting Mortality in Intensive Care Unit Patients: A Prognostic Performance Systematic Review and Meta-Analysis. Nurs Crit Care . 2025;30(6):e70206. doi:10.1111/nicc.70206 Mirzakhani F, Sadoughi F, Hatami M, Amirabadizadeh A. Which model is superior in predicting ICU survival: artificial intelligence versus conventional approaches. BMC Med Inform Decis Mak . 2022;22(1):167. doi:10.1186/s12911-022-01903-9 Ali S, Akhlaq F, Imran AS, Kastrati Z, Daudpota SM, Moosa M. The enlightening role of explainable artificial intelligence in medical & healthcare domains: A systematic literature review. Comput Biol Med . 2023;166:107555. doi:10.1016/j.compbiomed.2023.107555 Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data . 2023;10(1):1. doi:10.1038/s41597-022-01899-x Khan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association’s PREVENT Equations. Circulation . 2024;149(6):430-449. doi:10.1161/CIRCULATIONAHA.123.067626 Mansoori A, Nosrati M, Dorchin M, et al. A novel index for diagnosis of type 2 diabetes mellitus: Cholesterol, High density lipoprotein, and Glucose (CHG) index. J Diabetes Investig . 2025;16(2):309-314. doi:10.1111/jdi.14343 Austin PC, White IR, Lee DS, van Buuren S. Missing Data in Clinical Research: A Tutorial on Multiple Imputation. Can J Cardiol . 2021;37(9):1322-1331. doi:10.1016/j.cjca.2020.11.010 Xu J, Cai D, Jiao Y, et al. Insights into the complex relationship between triglyceride glucose-waist height ratio index, mean arterial pressure, and cardiovascular disease: a nationwide prospective cohort study. Cardiovasc Diabetol . 2025;24(1):93. doi:10.1186/s12933-025-02657-0 Degenhardt F, Seifert S, Szymczak S. Evaluation of variable selection methods for random forests and omics data sets. Brief Bioinform . 2019;20(2):492-503. doi:10.1093/bib/bbx124 Gisperg F, Klausser R, Elshazly M, Kopp J, Brichtová EP, Spadiut O. Bayesian Optimization in Bioprocess Engineering-Where Do We Stand Today? Biotechnol Bioeng . 2025;122(6):1313-1325. doi:10.1002/bit.28960 Khoja A, Andraweera PH, Lassi ZS, et al. Modifiable and Non-Modifiable Risk Factors for Premature Coronary Heart Disease (PCHD): Systematic Review and Meta-Analysis. Heart Lung Circ . 2024;33(3):265-280. doi:10.1016/j.hlc.2023.12.012 Ganjali S, Watts GF, Banach M, Reiner Ž, Nachtigal P, Sahebkar A. The Yin and Yang of High-density Lipoprotein and Atherosclerotic Cardiovascular Disease: Focusing on Functionality and Cholesterol Efflux to Reframe the HDL Hypothesis. Curr Med Chem . 2021;28(29):6066-6081. doi:10.2174/0929867328666210208182326 Alkatiri AH, Qalby N, Mappangara I, et al. Stress hyperglycemia and poor outcomes in patients with ST-elevation myocardial infarction: a systematic review and meta-analysis. Front Cardiovasc Med . 2024;11:1303685. doi:10.3389/fcvm.2024.1303685 Mottillo S, Filion KB, Genest J, et al. The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis. J Am Coll Cardiol . 2010;56(14):1113-1132. doi:10.1016/j.jacc.2010.05.034 Wei S, Yu S, Qian C, Xu Z, Jia Y, Chen B. Exploring the U-shaped nonlinear relationship of CHG index with metabolic syndrome and mortality risks in metabolic syndrome patients. Lipids Health Dis . 2025;24(1):294. doi:10.1186/s12944-025-02714-3 Zhu H, Liu L, Yang S, et al. Association of cholesterol, high-density lipoprotein and glucose (CHG) index with mortality risk in metabolic dysfunction-associated steatotic liver disease (MASLD) adults: results from two prospective cohorts. Cardiovasc Diabetol . 2026;25(1):23. doi:10.1186/s12933-026-03079-2 Ke L, Li Y, Jiang S, Xing W, Zhao L. Associations between the CHG index, its modified versions and incident stroke in patients with early-stage CKM syndrome: a nationwide cohort study. Eur J Med Res . 2026;31(1):206. doi:10.1186/s40001-025-03808-5 Guo Z, Xiong Z, Zhang W, et al. Association between cholesterol, high-density lipoprotein, and glucose index and risks of cardiovascular and all-cause mortality in patients with calcific aortic valve stenosis: the ARISTOTLE cohort study. Cardiovasc Diabetol . 2025;24(1):393. doi:10.1186/s12933-025-02906-2 Lin Y, Ke J, Chen S, et al. Cholesterol, high-density lipoprotein, glucose index as a novel marker for predicting in-stent restenosis after drug-eluting stent implantation in patients with acute coronary syndrome. Front Cardiovasc Med . 2026;13:1740216. doi:10.3389/fcvm.2026.1740216 Tang J, Wei S, Li Y, Zhou Y. Assessment of cholesterol-HDL-glucose index in anticipating risk of cardiometabolic diseases: a comparative study with triglyceride-glucose index. Sci Rep . 2025;15(1):45356. doi:10.1038/s41598-025-30024-2 Lu Y, Cui X, Zhang L, et al. The Functional Role of Lipoproteins in Atherosclerosis: Novel Directions for Diagnosis and Targeting Therapy. Aging Dis . 2022;13(2):491-520. doi:10.14336/AD.2021.0929 Chistiakov DA, Bobryshev YV, Orekhov AN. Macrophage-mediated cholesterol handling in atherosclerosis. J Cell Mol Med . 2016;20(1):17-28. doi:10.1111/jcmm.12689 Luc K, Schramm-Luc A, Guzik TJ, Mikolajczyk TP. Oxidative stress and inflammatory markers in prediabetes and diabetes. J Physiol Pharmacol . 2019;70(6). doi:10.26402/jpp.2019.6.01 Mengstie MA, Chekol Abebe E, Behaile Teklemariam A, et al. Endogenous advanced glycation end products in the pathogenesis of chronic diabetic complications. Front Mol Biosci . 2022;9:1002710. doi:10.3389/fmolb.2022.1002710 Adorni MP, Ronda N, Bernini F, Zimetti F. High Density Lipoprotein Cholesterol Efflux Capacity and Atherosclerosis in Cardiovascular Disease: Pathophysiological Aspects and Pharmacological Perspectives. Cells . 2021;10(3):574. doi:10.3390/cells10030574 Allard-Ratick MP, Kindya BR, Khambhati J, et al. HDL: Fact, fiction, or function? HDL cholesterol and cardiovascular risk. Eur J Prev Cardiol . 2021;28(2):166-173. doi:10.1177/2047487319848214 Denimal D. Antioxidant and Anti-Inflammatory Functions of High-Density Lipoprotein in Type 1 and Type 2 Diabetes. Antioxidants (Basel) . 2023;13(1):57. doi:10.3390/antiox13010057 Libby P. Inflammation during the life cycle of the atherosclerotic plaque. Cardiovasc Res . 2021;117(13):2525-2536. doi:10.1093/cvr/cvab303 Silveira Rossi JL, Barbalho SM, Reverete de Araujo R, Bechara MD, Sloan KP, Sloan LA. Metabolic syndrome and cardiovascular diseases: Going beyond traditional risk factors. Diabetes Metab Res Rev . 2022;38(3):e3502. doi:10.1002/dmrr.3502 Chafai N, Bonizzi L, Botti S, Badaoui B. Emerging applications of machine learning in genomic medicine and healthcare. Crit Rev Clin Lab Sci . 2024;61(2):140-163. doi:10.1080/10408363.2023.2259466 Salluh JIF, Moralez GM, Tracy A, Deliberato RO. ICU scoring systems: current perspectives and future directions. Curr Opin Crit Care . 2025;31(5):608-615. doi:10.1097/MCC.0000000000001305 Yu MY, Yoo HY, Han GI, Kim EJ, Son YJ. Comparing the Performance of Machine Learning Models and Conventional Risk Scores for Predicting Major Adverse Cardiovascular Cerebrovascular Events After Percutaneous Coronary Intervention in Patients With Acute Myocardial Infarction: Systematic Review and Meta-Analysis. J Med Internet Res . 2025;27:e76215. doi:10.2196/76215 Grant L, Joo P, Nemnom MJ, Thiruganasambandamoorthy V. Machine learning versus traditional methods for the development of risk stratification scores: a case study using original Canadian Syncope Risk Score data. Intern Emerg Med . 2022;17(4):1145-1153. doi:10.1007/s11739-021-02873-y Ganie SM, Pramanik PKD, Zhao Z. Ensemble learning with explainable AI for improved heart disease prediction based on multiple datasets. Sci Rep . 2025;15(1):13912. doi:10.1038/s41598-025-97547-6 Ensemble Learning for Disease Prediction: A Review - PubMed. Accessed February 25, 2026. https://pubmed.ncbi.nlm.nih.gov/37372925/ Ciobanu-Caraus O, Aicher A, Kernbach JM, Regli L, Serra C, Staartjes VE. A critical moment in machine learning in medicine: on reproducible and interpretable learning. Acta Neurochir (Wien) . 2024;166(1):14. doi:10.1007/s00701-024-05892-8 Bernal L, Rastelli G, Pinzi L. Improving Machine Learning Classification Predictions through SHAP and Features Analysis Interpretation. J Chem Inf Model . 2025;65(21):11716-11732. doi:10.1021/acs.jcim.5c02015 Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterial.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 27 Apr, 2026 Reviews received at journal 15 Apr, 2026 Reviewers agreed at journal 14 Apr, 2026 Reviews received at journal 14 Apr, 2026 Reviewers agreed at journal 13 Apr, 2026 Reviews received at journal 10 Apr, 2026 Reviewers agreed at journal 10 Apr, 2026 Reviewers agreed at journal 09 Apr, 2026 Reviews received at journal 02 Apr, 2026 Reviewers agreed at journal 02 Apr, 2026 Reviewers invited by journal 02 Apr, 2026 Editor invited by journal 09 Mar, 2026 Editor assigned by journal 07 Mar, 2026 Submission checks completed at journal 07 Mar, 2026 First submitted to journal 04 Mar, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9035595","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":618880513,"identity":"b56c4810-9dd8-45bd-8b88-9917a8a06afc","order_by":0,"name":"Jijun Lan","email":"","orcid":"","institution":"Affiliated Hospital of Southwest Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jijun","middleName":"","lastName":"Lan","suffix":""},{"id":618880514,"identity":"7a0b734a-7826-4ad0-bbb8-266059e49b2b","order_by":1,"name":"Meihan Liu","email":"","orcid":"","institution":"Affiliated Hospital of Southwest Medical University","correspondingAuthor":false,"prefix":"","firstName":"Meihan","middleName":"","lastName":"Liu","suffix":""},{"id":618880515,"identity":"5fdd8740-23d7-499d-a2eb-17359e7eaa22","order_by":2,"name":"Yihua Wang","email":"","orcid":"","institution":"Affiliated Hospital of Southwest Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yihua","middleName":"","lastName":"Wang","suffix":""},{"id":618880516,"identity":"b2c9f354-a1a5-48ff-9e39-5d9b50582d22","order_by":3,"name":"Jianguo Feng","email":"","orcid":"","institution":"Anesthesiology and Critical Care Medicine Key Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Jianguo","middleName":"","lastName":"Feng","suffix":""},{"id":618880517,"identity":"11707a9a-d60a-4d5b-aa14-fca18ad32a8f","order_by":4,"name":"Maohua Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA6klEQVRIiWNgGAWjYBCDBAYG5gMINpFa2BIQbCK18BgQp8Xg+NnDr27U2OQZHD/z+TNvzmEGfvYcA4afO/BoOZOXZp1zLK3Y4EzuNmnebYcZJHveGDD2nsGtxexAjplxDtvhxA0Hcrcx5wK1GNzIMWBmbMOj5fwboJZ/QC3n3zz+DNJiT1DLjRzjx7ltQC03chikwbZIENBif+ONGXNuX1rizBvPzKT/bkvnkTjzrOBgLx4tkv05xp9zvtkk9p1Pfvxx5jZrOf725I0PfuLRAgRsEsg8HhBxAK8GYEL5QEDBKBgFo2AUjHQAAEGLWOQncTaWAAAAAElFTkSuQmCC","orcid":"","institution":"Affiliated Hospital of Southwest Medical University","correspondingAuthor":true,"prefix":"","firstName":"Maohua","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2026-03-05 04:08:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9035595/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9035595/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":106533511,"identity":"ad8eb099-ad1a-45d3-a6b7-3da4c6e4594e","added_by":"auto","created_at":"2026-04-09 14:57:35","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1117053,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of the study design.\u003c/p\u003e","description":"","filename":"image1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/7ffe223e4d7baf04247ee547.jpeg"},{"id":106533507,"identity":"ae6503b3-ae4e-4576-b807-587d1b74d443","added_by":"auto","created_at":"2026-04-09 14:57:34","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1082541,"visible":true,"origin":"","legend":"\u003cp\u003eKaplan-Meier curves of CHG index quintiles for mortality. (A) 30-day mortality; (B) 90-day mortality.\u003c/p\u003e","description":"","filename":"image2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/1d02a54a585a355f354f0739.jpeg"},{"id":106533509,"identity":"45d3a691-8223-4fdc-9666-369726096565","added_by":"auto","created_at":"2026-04-09 14:57:35","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1081767,"visible":true,"origin":"","legend":"\u003cp\u003eMultivariable-adjusted RCS analyses of CHG index. (A) Based on the 30-day mortality; (B) Based on the 90-day mortality.\u003c/p\u003e","description":"","filename":"image3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/fbdb52d6c00ca022b10a8ff5.jpeg"},{"id":106533405,"identity":"1e2061c6-25b1-47e0-aada-29e8a1566093","added_by":"auto","created_at":"2026-04-09 14:57:20","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1854779,"visible":true,"origin":"","legend":"\u003cp\u003eForest plots illustrating stratified analyses of the association of CHG index and ASCVD mortality. (A) Based on the 30-day mortality; (B) Based on the 90-day mortality.\u003c/p\u003e","description":"","filename":"image4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/ec8f7e64219c22768c57855e.jpeg"},{"id":106533414,"identity":"c5cef989-81e1-4516-bfa7-cd4d16a95e3b","added_by":"auto","created_at":"2026-04-09 14:57:23","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":395281,"visible":true,"origin":"","legend":"\u003cp\u003eThe Boruta algorithm ranks potential risk factors for 30-day mortality by importance.\u003c/p\u003e","description":"","filename":"image5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/40228bc928bc1ce530fd1057.jpeg"},{"id":106533401,"identity":"54eeef43-5b0e-4a69-952f-20befee5e078","added_by":"auto","created_at":"2026-04-09 14:57:19","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1094712,"visible":true,"origin":"","legend":"\u003cp\u003eThe ROC curve of a 30-day mortality prediction model based on machine learning and the clinical practicality and calibration evaluation of the best performance model. (A) ROC curve on the training set; (B) ROC curve on the testing set; (C) Calibration curve of Stacking Classifier model on training set; (D) Calibration curve of Stacking Classifier model on testing set; (E) DCA of Stacking Classifier model on the training set; (F) DCA of Stacking Classifier model on the testing set.\u003c/p\u003e","description":"","filename":"image6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/df7466a92078face0842ad42.jpeg"},{"id":106533450,"identity":"cfc98806-812b-47a5-b11f-4cc1dfc8a68e","added_by":"auto","created_at":"2026-04-09 14:57:28","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1787481,"visible":true,"origin":"","legend":"\u003cp\u003eModel interpretation based on SHAP. (A) SHAP summary plot and honeycomb plot of feature importance; (B) SHAP dependency plot.\u003c/p\u003e","description":"","filename":"image7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/cb06ddcf629edddcab030175.jpeg"},{"id":106533657,"identity":"38609d4d-c96a-429c-b5a1-af6f082e8ae6","added_by":"auto","created_at":"2026-04-09 14:57:55","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":13841101,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/bd54ae16-b554-4613-af47-8bed0c01784c.pdf"},{"id":106533447,"identity":"b254e666-7c0e-4779-b190-011cbd5ef0f9","added_by":"auto","created_at":"2026-04-09 14:57:27","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":73224,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-9035595/v1/ba3ed9fe61b92c904c9c82dc.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Association between the Cholesterol, High-Density Lipoprotein, and Glucose Index and All-Cause Mortality in Critically Ill Patients with Atherosclerotic Cardiovascular Disease: A machine learning-based retrospective cohort study from the MIMIC-IV database","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eCardiovascular disease (CVD) is a major cause of mortality on a global scale\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Atherosclerotic cardiovascular disease (ASCVD), comprising coronary artery disease and ischemic stroke, represents the predominant subtype of CVD and imposes a substantial burden on global healthcare systems\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Critically ill patients with ASCVD admitted to the intensive care unit (ICU) constitute a high-risk population with poor prognosis, and metabolic dysregulation frequently complicates their clinical course\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e. Recent studies have highlighted a strong link between metabolic disturbances\u0026mdash;particularly in lipid and glucose regulation\u0026mdash;and mortality risk among critically ill populations\u003csup\u003e\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e,\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Dysfunction of glucose and lipid metabolism has been shown to be associated with disease severity and poor prognosis in ASCVD patients\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eInsulin resistance (IR) stands at the core of metabolic dysregulation and is increasingly recognized as a key contributor to poor cardiovascular outcomes\u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e. Although the hyperinsulinemic-euglycemic clamp remains the reference standard for assessing insulin sensitivity, its complexity and time demands restrict its use in everyday clinical practice\u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. In response, simpler surrogate indices have emerged\u0026mdash;such as the triglyceride-glucose (TyG) index and the metabolic score for insulin resistance (METS-IR)\u0026mdash;both of which have shown promise in predicting mortality among critically ill patients\u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e. Large epidemiological studies have further linked the TyG index to cardiovascular mortality\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. That said, these markers rely mainly on triglyceride (TG) and fasting blood glucose (FBG) levels, leaving cholesterol metabolism largely unaccounted for. Of note, low high-density lipoprotein (HDL) cholesterol often accompanies IR, and their combined presence amplifies cardiovascular risk\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e. The cholesterol, high-density lipoprotein, and glucose (CHG) index, which brings together total cholesterol (TC), HDL, and FBG, has been put forward as a more integrated measure of both lipid and glucose status\u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. Whether this index holds prognostic value for critically ill patients with established ASCVD, however, remains an open question.\u003c/p\u003e \u003cp\u003eMachine learning approaches, capable of capturing complex, nonlinear relationships among variables, are regarded as valuable tools for forecasting clinical outcomes\u003csup\u003e\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. Various machine learning models, including random forest (RF), extreme gradient boosting (XGBoost), and gradient boosting (GB), have been developed and validated with high accuracy for clinical prediction tasks\u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e. Their application to mortality prediction in ICU settings has grown considerably, and several studies suggest they may outperform established tools like the Acute Physiology and Chronic Health Evaluation (APACHE) and Sequential Organ Failure Assessment (SOFA) scores\u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Furthermore, the SHapley Additive exPlanations (SHAP) method addresses the \"black box\" concern often associated with machine learning in clinical settings by quantifying the contribution of each predictor to individual predictions\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. This interpretable machine learning approach offers a valuable framework for assessing the prognostic relevance of novel metabolic indices in critically ill populations.\u003c/p\u003e \u003cp\u003eUsing data drawn from the MIMIC-IV database, this study set out to assess whether the CHG index correlates with all-cause mortality in critically ill patients diagnosed with ASCVD. A secondary aim was to build interpretable machine learning models capable of identifying the key factors of prognosis in this vulnerable population.\u003c/p\u003e"},{"header":"2. Methods","content":"\u003cp\u003e \u003cb\u003eData source\u003c/b\u003e \u003c/p\u003e \u003cp\u003eWe retrieved data for this retrospective cohort study from the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.1). This publicly accessible critical care resource contains de-identified electronic health records from over 360,000 patients and 546,000 hospital admissions at Beth Israel Deaconess Medical Center (Boston, MA, USA) between 2008 and 2022, including 94,458 ICU stays\u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Patient privacy is safeguarded through the removal of all personal identifiers. The original data collection was approved by the institutional review boards of Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology, with a waiver of informed consent granted due to the anonymized nature of the data. Author Ji-Jun Lan completed the National Institutes of Health (NIH) web-based training course on protecting human research participants (certification number: 69132229) and obtained access to the database.\u003c/p\u003e \u003cp\u003e \u003cb\u003eStudy population\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe diagnosis of ASCVD was established through manual review of International Classification of Diseases (ICD), Ninth and Tenth Revisions (ICD-9 and ICD-10) (detailed in Supplementary Material Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e)\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e. Patients were excluded from participation according to the following criteria: (1) age\u0026thinsp;\u0026lt;\u0026thinsp;18 years; (2) multiple ICU admissions for ASCVD (only the first admission was included); (3) ICU length of stay\u0026thinsp;\u0026lt;\u0026thinsp;24 hours; (4) missing data on glucose, TC, or HDL; (5) presence of severe liver disease, metastatic solid tumor, or malignant cancer; (6) outliers of CHG index, defined as values below Q1\u0026thinsp;\u0026minus;\u0026thinsp;3IQR or above Q3\u0026thinsp;+\u0026thinsp;3IQR, where Q1 and Q3 respectively represent the first and third quartiles, and IQR stands for the interquartile range. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e illustrates the patient selection process.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eData acquisition\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe data extraction process was executed utilizing the structured query language (SQL) via Navicat Premium (version 17), with a particular emphasis placed on six distinct categories: (1) Demographics: age, sex, race; (2) Vital signs: heart rate, weight, temperature, respiratory rate (RR), systolic blood pressure (SBP), diastolic blood pressure (DBP), peripheral oxygen saturation (SpO2); (3) Clinical severity scores: Glasgow Coma Score (GCS), Acute Physiology Score III (APSIII); (4) Laboratory parameters: blood urea nitrogen (BUN), creatinine, potassium, sodium, anion gap, bicarbonate, white blood cell (WBC), red blood cell (RBC), platelets, hemoglobin, red blood cell distribution width (RDW), hematocrit (HCT), glucose, TG, TC, HDL, low-density lipoprotein cholesterol (LDL); (5) Comorbidities: sepsis, congestive heart failure, dementia, chronic pulmonary disease, rheumatic disease, diabetes, renal disease, mild liver disease, paraplegia, hypertension; (6) Medications: vasopressin, aspirin, clopidogrel, warfarin, and statins.\u003c/p\u003e \u003cp\u003eWe computed the CHG index according to the formula: CHG index\u0026thinsp;=\u0026thinsp;Ln [TC (mg/dL) \u0026times; FBG (mg/dL) / (2 \u0026times; HDL (mg/dL))]\u003csup\u003e19\u003c/sup\u003e. Vital signs and laboratory parameters were based on the first measurement within 24 hours of ICU admission to capture baseline physiological status. Clinical severity scores were calculated based on the mean values of measurements taken on the first day of ICU admission. The utilization of pharmaceuticals was defined as the record of whether the patient received the specified medications during their ICU stay. The main endpoint was designated as 30-day all-cause mortality, whereas the secondary endpoint was designated as 90-day all-cause mortality.\u003c/p\u003e \u003cp\u003e \u003cb\u003eStatistical analysis\u003c/b\u003e \u003c/p\u003e \u003cp\u003eThe CHG index was then grouped into quintiles: Q1 (3.44\u0026ndash;4.98), Q2 (4.98\u0026ndash;5.25), Q3 (5.25\u0026ndash;5.50), Q4 (5.50\u0026ndash;5.84), and Q5 (5.84\u0026ndash;7.53). The continuous variables were reported as the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation (SD) or the median (interquartile range (IQR)), and the group comparisons were performed using either the t-test or the Mann-Whitney U test. Categorical variables were expressed as frequencies (percentages) and compared using chi-square or Fisher's exact test. Multicollinearity was detected using the Variance Inflation Factor (VIF). Covariates with a VIF greater than 5 were omitted from multivariate analyses. Kaplan-Meier curves were employed to visualize survival differences, and Cox proportional hazards regression was applied with three adjustment models: The first model was a crude model; the second model was adjusted for age, sex and race; the third model was further adjusted for comorbidities, vital signs, medications, laboratory parameters and severity scores. The selection of covariates was based on clinical relevance and univariate significance, with multicollinearity controlled (VIF\u0026thinsp;\u0026lt;\u0026thinsp;5). Restricted cubic splines (RCS) based on Cox proportional hazards model (Model 3) were used, along with histograms, to evaluate the dose-response relationship between CHG index and all-cause mortality at different prognostic time points. Subgroup and sensitivity analyses were conducted to evaluate result robustness. The missing data were then imputed using multiple imputation by chained equations (MICE), which generated five imputed datasets with pooled estimates calculated via Rubin's rules\u003csup\u003e\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMachine learning model development and performance assessment\u003c/b\u003e \u003c/p\u003e \u003cp\u003eIn the context of machine learning-based prediction, the dataset was initially partitioned into a training and testing set at an 8:2 ratio, with the sets being randomly selected. The Boruta algorithm was subsequently employed on the training set to rank the importance of predictive features for 30-day mortality\u003csup\u003e\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. A total of nine base models were developed, including logistic regression (LR), RF, XGBoost, light gradient boosting machine (LightGBM), linear discriminant analysis (LDA), GB, adaptive boosting (AdaBoost), naive bayes (NB), and quadratic discriminant analysis (QDA). A stacking ensemble was also built, combining logistic regression (LR), linear discriminant analysis (LDA), XGBoost, RF, and adaptive boosting (AdaBoost) as base learners, with LR again serving as the meta-learner. Each model underwent hyperparameter tuning via Bayesian optimization\u0026mdash;100 iterations per model\u0026mdash;with the area under the receiver operating characteristic curve (AUC) set as the objective function to maximize\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e. Performance was evaluated using five repetitions of 10-fold stratified cross-validation. Several metrics were used to assess model performance, including AUC, accuracy, specificity, sensitivity, and F1 score. The calibration curves were utilized to assess the concordance between the observed and predicted outcomes. Decision curve analysis (DCA) was employed to evaluate the clinical net benefit. The best-performing model was further interpreted using SHAP for determining crucial predictive factors. All data analyses were performed using Python 3.12.4 and R 4.5.0 software. A two-tailed P value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003eBaseline characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study population comprised 1,959 critically ill patients with ASCVD were included, with a median age of 71.99 years (IQR: 61.33-82.63) and 52.8% (n=1,034) being male.\u0026nbsp;The statistics of the missing ratio before missing value interpolation are shown in Supplementary Material Table S2. The median CHG index was 5.38 (IQR: 5.05-5.73). The mortality rates within 30 and 90 days of ICU admission were 16.7% and 21.0%, respectively. Analysis revealed that non-survivors were older [80.74 (69.25-87.89) vs. 70.29 (59.83-81.17) years], had higher CHG index levels [5.44 (5.15-5.88) vs. 5.37 (5.04-5.71)], higher APSIII scores [48 (34-61) vs. 33 (25-43)], lower GCS scores [14.67 (12.46-15.00) vs. 14.84 (13.99-15.00)], higher heart rate, RR, WBC count, glucose, BUN, and RDW, lower RBC count, hemoglobin, TC, and LDL, higher prevalence of comorbidities including sepsis, congestive heart failure, and renal disease, higher vasopressin use, and lower use of aspirin, warfarin, and statins (all P\u0026lt;0.05). The baseline characteristics of the two groups are exhibited in Table 1 and Supplementary Material Table S3.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 1 Baseline characteristics according to 30-day mortality\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"615\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003eOverall(n=1959)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003eSurvivors(n=1632)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003eNon-survivors(n=327)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003eP value\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eCHG index\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 5.38 (5.05-5.73)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 5.37 (5.04-5.71)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 5.44 (5.15-5.88)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDemographic variables\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 164px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 165px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 185px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eAge (years)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;71.99 (61.33-82.63)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;70.29 (59.83-81.17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;80.74 (69.25-87.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eMale (n, %)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1034 (52.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;882 (54.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;152 (46.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRace (White)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1031 (52.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;881 (54.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;150 (45.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eVital signs\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eHeart rate (b/min)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;80.00 (69.00-93.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;79.00 (69.00-92.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;85.00 (71.00-97.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRR (b/min)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;18.00 (15.00-22.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;18.00 (15.00-22.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;19.00 (16.00-23.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSBP (mmHg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e140.00 (122.00-156.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e140.00 (123.00-156.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e137.00 (117.00-155.25)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDBP (mmHg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;77.00 (66.00-90.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;78.00 (67.00-91.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;74.00 (61.75-86.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSpo2 (b/min)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;98.00 (96.00-99.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;97.00 (96.00-99.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;98.00 (96.00-100.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eWeight (kg)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;78.00 (65.50-92.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;78.70 (66.68-93.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;73.30 (59.95-86.75)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eTemperature (℃)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;36.78 (36.50-37.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;36.78 (36.50-37.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;36.72 (36.44-37.06)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.129\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eClinical severity scores\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eGCS\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.83 (13.79-15.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.84 (13.99-15.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.67 (12.46-15.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eAPSIII\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;35.00 (26.00-46.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;33.00 (25.00-43.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;48.00 (34.00-61.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eLaboratory parameters\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 164px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 165px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd colspan=\"2\" style=\"width: 185px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eBUN (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;17.00 (13.00-25.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;16.00 (12.00-23.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;22.00 (16.00-34.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eCreatinine (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 0.90 (0.80-1.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 0.90 (0.70-1.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1.10 (0.80-1.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003ePotassium (mEq/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 4.10 (3.70-4.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 4.00 (3.70-4.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 4.20 (3.80-4.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSodium (mEq/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e139.00 (137.00-141.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e139.00 (137.00-141.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e139.00 (137.00-142.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.282\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eAnion gap (mmol/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.00 (12.00-16.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.00 (12.00-16.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;15.00 (13.00-18.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eBicarbonate (mmol/L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;23.00 (21.00-25.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;23.00 (21.00-25.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;22.00 (19.00-24.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eWBC (\u0026times;10\u003csup\u003e3\u003c/sup\u003e/\u0026micro;L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;10.00 (7.70-13.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 9.60 (7.50-12.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;11.90 (9.20-15.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRBC (\u0026times;10\u003csup\u003e6\u003c/sup\u003e/\u0026micro;L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 4.17 (3.70-4.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 4.21 (3.75-4.63)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 3.99 (3.49-4.46)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003ePlatelets (\u0026times;10\u003csup\u003e3\u003c/sup\u003e/\u0026micro;L)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e215.00 (171.00-266.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e216.00 (173.00-265.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e213.00 (161.00-270.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eHemoglobin (g/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;12.50 (11.00-13.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;12.60 (11.10-13.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;11.80 (10.50-13.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRDW (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;13.70 (13.00-14.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;13.60 (12.90-14.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;14.20 (13.40-15.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eHCT (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;38.10 (33.80-41.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;38.40 (34.20-41.60)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;36.10 (32.40-40.55)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eGlucose (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e125.00 (103.00-162.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e122.00 (100.75-154.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e145.00 (119.00-196.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eTG (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e102.00 (74.00-143.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e103.00 (75.75-145.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;94.00 (68.50-135.50)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eTC (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e156.00 (127.00-189.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e159.50 (129.00-191.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e140.00 (112.00-172.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eHDL (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;45.00 (36.00-57.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;46.00 (37.00-57.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;45.00 (35.00-54.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.022\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eLDL (mg/dL)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;83.00 (60.00-113.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;86.00 (62.00-114.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp;69.00 (51.00-101.00)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eComorbidities\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eSepsis (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;607 (31.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;421 (25.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;186 (56.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eCongestive heart failure (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;625 (31.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;483 (29.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;142 (43.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDementia (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;110 (5.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 74 (4.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 36 (11.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eChronic pulmonary disease (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;327 (16.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;264 (16.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 63 (19.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.198\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRheumatic disease (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 48 (2.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 39 (2.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;9 (2.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.848\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eDiabetes (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;659 (33.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;533 (32.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;126 (38.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.047\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eRenal disease (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;380 (19.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;289 (17.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 91 (27.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eMild liver disease (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 48 (2.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 40 (2.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; \u0026nbsp;8 (2.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eParaplegia (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;961 (49.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;770 (47.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;191 (58.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eHypertension (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1525 (77.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1254 (76.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;271 (82.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eMedications\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eVasopressin (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 55 (2.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 16 (1.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 39 (11.9)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eAspirin (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1314 (67.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1119 (68.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;195 (59.6)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eClopidogrel (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;357 (18.2)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;311 (19.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 46 (14.1)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eWarfarin (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;190 (9.7)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;177 (10.8)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp; 13 (4.0)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 102px;\"\u003e\n \u003cp\u003eStatins (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1339 (68.4)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; 1164 (71.3)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 152px;\"\u003e\n \u003cp\u003e\u0026nbsp; \u0026nbsp;175 (53.5)\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 57px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd style=\"width: 102px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 152px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 12px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 140px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 24px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 128px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd style=\"width: 57px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eData are expressed as median (IQR), or n (%). CHG index: Cholesterol, high-density lipoprotein, and glucose index; RR: respiratory rate; SBP: systolic blood pressure; DBP: diastolic blood pressure; Spo2: peripheral oxygen saturation; GCS: Glasgow Coma Score; APSIII: Acute Physiology Score III; BUN: blood urea nitrogen; WBC: white blood cell; RBC: red blood cell; RDW: red blood cell distribution width; HCT: hematocrit; TG: triglyceride; TC: total cholesterol; HDL: high-density lipoprotein cholesterol; LDL: low-density lipoprotein cholesterol\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAssociation between CHG index and all-cause mortality\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2 The association\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ebetween the CHG index and all-cause mortality\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"687\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" rowspan=\"2\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eVariables\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 183px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 1\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 178px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 2\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"3\" valign=\"top\" style=\"width: 180px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModel 3\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 2px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHR (95%CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHR (95%CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eHR (95%CI)\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eP\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003e30-day mortality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eCHG (per 1 unit)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.62 (1.33-1.97)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e2.24 (1.82-2.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.58 (1.22-2.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eQuintile groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.55 (1.06-2.26)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.023\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.61 (1.10-2.35)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.014\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.75 (1.19-2.59)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.005\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.37 (0.93-2.02)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.114\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.54 (1.04-2.27)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.54 (1.03-2.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.037\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.52 (1.04-2.22)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.88 (1.28-2.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.76 (1.16-2.67)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e2.13 (1.49-3.05)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e3.12 (2.16-4.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e2.10 (1.34-3.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eP for trend\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.62 (1.27-2.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e2.22 (1.73-2.87)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.60 (1.16-2.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.004\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003e90-day mortality\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eCHG (per 1 unit)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.53 (1.28-1.83)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e2.13 (1.77-2.58)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.61 (1.28-2.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eQuintile groups\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" valign=\"top\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.00 (Reference)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.49 (1.07-2.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.019\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.56 (1.12-2.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.70 (1.21-2.40)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.38 (0.98-1.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.061\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.57 (1.12-2.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.61 (1.13-2.29)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.008\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.48 (1.06-2.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 58px;\"\u003e\n \u003cp\u003e0.021\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 123px;\"\u003e\n \u003cp\u003e1.86 (1.33-2.61)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.78 (1.23-2.58)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eGroup 5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.96 (1.42-2.70)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 123px;\"\u003e\n \u003cp\u003e2.91 (2.10-4.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e2.07 (1.39-3.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd nowrap=\"\" valign=\"top\" style=\"width: 144px;\"\u003e\n \u003cp\u003eP for trend\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.53 (1.24-1.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 123px;\"\u003e\n \u003cp\u003e2.12 (1.69-2.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 58px;\"\u003e\n \u003cp\u003e\u0026lt;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" style=\"width: 124px;\"\u003e\n \u003cp\u003e1.60 (1.20-2.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd nowrap=\"\" colspan=\"2\" style=\"width: 56px;\"\u003e\n \u003cp\u003e0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eHR: Hazard Ratio, CI: Confidence Interval. Model 1: unadjusted; Model 2: adjusted for age, gender, race; Model 3: adjusted for Model 2 plus Sepsis, Congestive heart failure, dementia, diabetes, renal disease, paraplegia, hypertension, heart rate, DBP, RR, weight, BUN, creatinine, potassium, anion gap, bicarbonate, WBC, RBC, RDW, LDL, vasopressin, aspirin, clopidogrel, warfarin, statins, GCS, APSIII.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eKaplan-Meier survival analysis showed significant differences in survival rates among CHG index quintiles, with patients in the highest CHG index groups demonstrating higher 30-day and 90-day all-cause mortality than those in the lowest groups (log-rank test, P = 0.00079 and P = 0.0012, respectively); the detailed findings are illustrated in Figure 2. The results of multivariate Cox proportional hazards regression analysis are shown in Table 2, and the selection criteria for covariates in Model 3 are provided in Supplementary Materials Table S4-6. In the fully adjusted model (Model 3), a 1-unit increase in the CHG index showed a 58% increased risk of 30-day all-cause mortality (HR=1.58, 95% CI: 1.22-2.05) and a 61% higher risk of 90-day all-cause mortality (HR=1.61, 95% CI: 1.28-2.03) (both P\u0026lt;0.001). For 30-day mortality, utilizing Q1 group as the reference, the hazard ratios (HR) and 95% confidence intervals (CI) for Q5 group in Model 1, Model 2, and Model 3 were 2.13 (1.49-3.05), 3.12 (2.16-4.51), and 2.10 (1.34-3.29), respectively. For 90-day mortality, the HR and 95% CI for Q5 group in Model 1, Model 2, and Model 3 were 1.96 (1.42-2.70), 2.91 (2.10-4.04), and 2.07 (1.39-3.07), respectively. The trend test indicated a dose-response relationship between CHG index and all-cause mortality risk (P = 0.004 and P = 0.001, respectively). RCS regression analysis confirmed a linear dose-response relationship between CHG index and all-cause mortality risk (Figure 3). In the fully adjusted model (Model 3), the CHG index exhibited a linear correlation with both 30-day mortality (P-nonlinear=0.250) and 90-day mortality (P-nonlinear=0.076).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSubgroup and sensitivity analyses\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe results of the subgroup analysis are shown in Figure 4. Subgroup analysis revealed a significant interaction between the CHG index and 30-day mortality risk depending on clopidogrel use (P for interaction = 0.049) and statin use (P for interaction = 0.008). Among patients not receiving clopidogrel, a higher CHG index was associated with increased mortality risk (HR = 1.38, 95% CI: 1.09\u0026ndash;1.74); this association was not observed in those on clopidogrel (HR = 0.79, 95% CI: 0.40\u0026ndash;1.53). Similarly, the association was stronger in patients not using statins (HR=1.69, 95% CI: 1.19-2.39) compared with those using statins (HR=1.14, 95% CI: 0.84-1.56). For 90-day mortality, significant interactions were found across age (P for interaction=0.033), aspirin use (P for interaction=0.030), and statin use (P for interaction=0.013) subgroups. The CHG index demonstrated a significant association with mortality risk in patients aged \u0026ge;65 years (HR=1.35, 95% CI: 1.08-1.70), but not in those aged \u0026lt;65 years (HR=1.47, 95% CI: 0.95-2.27).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWe conducted multiple sensitivity analyses to evaluate the stability of our results. Firstly, following the elimination of missing values, the outcomes of the Cox proportional hazards regression analysis were found to be in alignment with the primary analysis (Supplementary Material Table S7). Secondly, utilizing the technique of logistic regression, the correlation between the CHG index and all-cause mortality persisted in alignment with the primary findings (Supplementary Material Table S8).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eModel performance evaluation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFeature importance ranking was performed on the training set using the Boruta algorithm (Figure 5). We identified the top 16 important features, including APSIII, vasopressin use, sepsis, GCS score, WBC count, age, glucose, BUN, weight, temperature, CHG index, bicarbonate, SpO2, HCT, TC, and statin use, and subsequently constructed 10 machine learning models for predicting 30-day mortality. The performance metrics of each model on the training and test sets are presented in Supplementary Material Table S9, and the optimal hyperparameter settings for each model are shown in Supplementary Material Table S10. On the training set, RF achieved the highest AUC (0.912, 95% CI: 0.898-0.926), followed by GB (AUC=0.907) and Stacking Classifier (AUC=0.882). However, on the test set, Stacking Classifier demonstrated the most stable performance with an AUC of 0.833 (95% CI: 0.796-0.870), which was comparable to its training set result (AUC difference=0.049), indicating good generalization ability (Figure 6A-B). In contrast, RF and GB showed decreased AUC values of 0.821 and 0.795 on the test set (AUC differences of 0.091 and 0.112, respectively), suggesting some degree of overfitting. Considering multiple performance metrics, Stacking Classifier achieved the highest specificity (0.911), accuracy (0.847), and F1-score (0.531) on the test set, while maintaining a high AUC value, demonstrating the best discriminative ability and stability. Therefore, Stacking Classifier was selected as the optimal model. We further evaluated the clinical utility and calibration performance of the Stacking Classifier model. DCA showed that the model delivered a positive net benefit across a wide range of threshold probabilities (0.2-0.8) in both training and test sets, outperforming the \u0026quot;treat-all\u0026quot; and \u0026quot;treat-none\u0026quot; strategies (Figure 6C-D). Calibration curves (Figure 6E-F) revealed good agreement between predicted probabilities and observed outcomes in both the training and test sets with Brier scores of 0.099 and 0.103 for the training and test sets, respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSHAP-based interpretation of the optimal model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSHAP feature importance analysis (Figure 7A) identified the APSIII score as the most influential predictor of 30-day mortality, with the highest mean absolute SHAP value (0.045). This was followed by sepsis (0.041), age (0.037), WBC count (0.029), and statin use (0.024). The CHG index is considered the primary indicator of this study, with a mean absolute SHAP value of 0.011, placing it 8th among all features and indicating moderate predictive importance. The SHAP summary plot analysis was conducted, indicating that the use of statins was correlated with negative SHAP values. This finding suggests that statin use may function as a protective factor in terms of mortality. Conversely, patients exhibiting elevated APSIII scores, sepsis, advanced age, elevated WBC counts, and elevated CHG indices demonstrated positive SHAP values, signifying their role as independent risk factors for mortality. As illustrated in Figure 7B, the dependence plots provided additional information on the impact of each predictor on model output, and on potential interaction effects.\u003c/p\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study set out to ascertain the association between the CHG index and the 30-day and 90-day all-cause mortality rates of critically ill patients with ASCVD. The present findings indicate that an augmented CHG index is independently correlated with elevated mortality risk. Kaplan-Meier survival analysis showed statistically significant variations in survival probability between the CHG index quintiles. Multivariate Cox proportional hazards regression was employed to confirm that the CHG index continued to independently predict mortality following adjustment for potential confounders. Moreover, RCS analysis demonstrated a linear dose-response association of CHG index with mortality. It is noteworthy that the CHG index serves as a critical predictor in machine learning models, underscoring its potential application in risk stratification within clinical practice.\u003c/p\u003e \u003cp\u003eLipid and glucose metabolic disturbances are fundamental pathophysiological features of ASCVD. A systematic review covering 208 studies indicated that elevated total cholesterol is associated with an increased risk of premature coronary heart disease\u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Regarding high-density lipoprotein cholesterol, epidemiological evidence has revealed a non-linear association with mortality risk, where reduced concentrations suggest a poor prognosis in patients with ASCVD\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Furthermore, a systematic review and meta-analysis confirmed that hyperglycaemia can independently predict mortality in patients with acute coronary syndromes\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u003c/sup\u003e. These metabolic imbalances can occur simultaneously, collectively increasing the risk of CVD mortality. Notably, metabolic syndrome, characterized by the coexistence of dyslipidaemia and hyperglycaemia, is often significantly associated with elevated CVD mortality\u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. This epidemiological evidence highlights the importance of lipid and glucose metabolism in determining cardiovascular prognosis, providing a robust theoretical rationale for integrating these parameters into a unified prognostic index.\u003c/p\u003e \u003cp\u003eIt has been demonstrated that the CHG index, a composite indicator of cholesterol, high-density lipoprotein, and glucose metabolism parameters, exhibits a correlation with adverse outcomes across a range of conditions, including metabolic syndrome, stroke, and CVD. In the context of metabolic syndrome populations, a U-shaped nonlinear relationship has been documented between the CHG index and mortality risk. This relationship suggests that both low and elevated levels of the index correlate with higher cardiovascular and all-cause death risk\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u003c/sup\u003e. The prognostic utility of the CHG index is further demonstrated in patients with metabolic dysfunction-associated steatotic liver disease, where it has been associated with mortality risk, particularly among individuals under 60 years of age and with lean body composition\u003csup\u003e\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e. Within the context of the CHG index, it has been demonstrated that there exists an independent association with the occurrence of stroke in patients diagnosed with early cardiovascular-renal-metabolic syndrome. Furthermore, the index has been shown to be associated with an elevated risk of cardiovascular mortality and mortality from any cause in patients diagnosed with calcific aortic valve stenosis\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. The index also functions as a novel predictor of in-stent restenosis following percutaneous coronary intervention, thereby emphasising its significance in the domain of interventional cardiology\u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. It is noteworthy that comparative analyses indicate that the CHG index performs comparably to the TyG index in predicting cardiovascular metabolic disease risk\u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. The present study is the first to specifically ascertain the relationship between CHG index and short-term mortality in critically ill patients with ASCVD, thereby highlighting the potential value of CHG index as a severity and prognosis marker.\u003c/p\u003e \u003cp\u003eThe pathophysiological mechanisms linking adverse outcomes in critically ill patients with ASCVD involve complex interactions among lipid metabolism, glucose homeostasis, and inflammatory responses. Elevated TC promotes the formation of foam cells by enhancing macrophage uptake of oxidized lipoproteins to a certain extent, thereby contributing to the development of atherosclerotic lesions\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Furthermore, elevated blood glucose upregulates inflammatory markers and increases the production of reactive oxygen species, leading to vascular dysfunction. Secondly, it accelerates the formation of advanced glycation end products, activating pro-inflammatory signaling cascades and consequently promoting atherosclerosis\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. In contrast, HDL exerts protective effects by facilitating reverse cholesterol transport, thereby reducing lipid accumulation in the arterial wall\u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. Concurrently, HDL particles possess anti-inflammatory and antioxidant properties that counteract the effects of oxidized lipoproteins\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e,\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. When these metabolic disturbances converge, a pro-inflammatory environment characterized by elevated cytokines and adhesion molecules is established, accelerating plaque progression and increasing the risk of rupture\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e,\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u003c/sup\u003e. Consequently, in critically ill patients, these metabolic disturbances and inflammatory responses may further deteriorate an already compromised cardiovascular status.\u003c/p\u003e \u003cp\u003eMachine learning has found increasingly widespread applications in medicine, demonstrating significant advantages in tasks ranging from disease diagnosis to prognostic prediction\u003csup\u003e\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Traditional prognostic assessment tools used in the ICU, although widely applied in clinical practice, have limitations in accuracy\u003csup\u003e\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u003c/sup\u003e. In contrast, machine learning models often achieve superior predictive performance, as they are better at handling complex, nonlinear relationships among variables compared to traditional indicator assessments\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e,\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. In the construction of machine learning models, ensemble methods enhance model stability and generalization capabilities by integrating outputs from multiple base models, thereby enabling more reliable predictions\u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e,\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e. Furthermore, the \"black-box\" nature of machine learning models has remained a barrier to their clinical application\u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. The SHAP method, grounded in Shapley values from game theory, provides an interpretability framework for machine learning models by quantifying the contribution of each feature to model predictions\u003csup\u003e\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. In our study, SHAP analysis identified the CHG index as a significant predictor of mortality in critically ill patients with ASCVD. This finding corroborates the results of the Cox regression analysis, offering machine learning-based support for the prognostic value of the CHG index.\u003c/p\u003e \u003cp\u003eThe present study is subject to several limitations. Firstly, as a retrospective cohort study, it was not possible to completely exclude selection bias and residual confounding. Secondly, the study data were derived from a single center (MIMIC-IV database), and the external validity of the models needs to be further validated in independent cohorts with diverse ethnicities, geographic regions and healthcare systems. Thirdly, as a retrospective cohort study, it was not possible to make definitive conclusions regarding causal relationships. Based on the findings of this study, future research can be explored in the following directions. First, prospective multicenter cohort studies should be conducted to validate the prognostic value of the CHG index across diverse ethnicities, geographic regions, and healthcare systems, and to evaluate the utility of CHG index-based risk stratification strategies in guiding clinical decision-making. Second, time-series analysis methods should be integrated to construct dynamic CHG index monitoring models, systematically assessing the association between metabolic indicator trends and prognosis, and exploring the feasibility of dynamic risk early warning. Third, multi-omics technologies (such as metabolomics, proteomics, and genomics) should be employed to deeply elucidate the molecular mechanisms underlying the association between the CHG index and adverse outcomes, identifying potential therapeutic intervention targets. Fourth, randomized controlled trials or observational intervention studies should be conducted to evaluate whether metabolic management to reduce the CHG index can improve clinical outcomes in critically ill patients with ASCVD, providing evidence-based medical evidence for the CHG index as a potential therapeutic target.\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eElevated CHG index was found to have an independent linear relationship with greater all-cause mortality among critically ill ASCVD patients. Machine learning analysis further confirms CHG index serves as a critical mortality predictor in this cohort. This novel metabolic index holds promise as an effective indicator for patient risk assessment and treatment decisions.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAdaBoost\u003c/p\u003e\n\u003cp\u003eadaptive boosting\u003c/p\u003e\n\u003cp\u003eAPSIII\u003c/p\u003e\n\u003cp\u003eAcute Physiology Score III\u003c/p\u003e\n\u003cp\u003eAPACHE\u003c/p\u003e\n\u003cp\u003eAcute Physiology and Chronic Health Evaluation\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eASCVD\u003c/p\u003e\n\u003cp\u003eatherosclerotic cardiovascular disease\u003c/p\u003e\n\u003cp\u003eAUC\u003c/p\u003e\n\u003cp\u003earea under the receiver operating characteristic curve\u003c/p\u003e\n\u003cp\u003eBoruta\u003c/p\u003e\n\u003cp\u003eBoruta algorithm\u003c/p\u003e\n\u003cp\u003eBUN\u003c/p\u003e\n\u003cp\u003eblood urea nitrogen\u003c/p\u003e\n\u003cp\u003eCHG\u003c/p\u003e\n\u003cp\u003eCholesterol, High-Density Lipoprotein, and Glucose\u003c/p\u003e\n\u003cp\u003eCI\u003c/p\u003e\n\u003cp\u003econfidence interval\u003c/p\u003e\n\u003cp\u003eCVD\u003c/p\u003e\n\u003cp\u003ecardiovascular disease\u003c/p\u003e\n\u003cp\u003eDCA\u003c/p\u003e\n\u003cp\u003edecision curve analysis\u003c/p\u003e\n\u003cp\u003eDBP\u003c/p\u003e\n\u003cp\u003ediastolic blood pressure\u003c/p\u003e\n\u003cp\u003eFBG\u003c/p\u003e\n\u003cp\u003efasting blood glucose\u003c/p\u003e\n\u003cp\u003eGB\u003c/p\u003e\n\u003cp\u003egradient boosting\u003c/p\u003e\n\u003cp\u003eGCS\u003c/p\u003e\n\u003cp\u003eGlasgow Coma Score\u003c/p\u003e\n\u003cp\u003eHCT\u003c/p\u003e\n\u003cp\u003ehematocrit\u003c/p\u003e\n\u003cp\u003eHDL\u003c/p\u003e\n\u003cp\u003ehigh-density lipoprotein cholesterol\u003c/p\u003e\n\u003cp\u003eHR\u003c/p\u003e\n\u003cp\u003ehazard ratio\u003c/p\u003e\n\u003cp\u003eICD\u003c/p\u003e\n\u003cp\u003eInternational Classification of Diseases\u003c/p\u003e\n\u003cp\u003eICU\u003c/p\u003e\n\u003cp\u003eintensive care unit\u003c/p\u003e\n\u003cp\u003eIQR\u003c/p\u003e\n\u003cp\u003einterquartile range\u003c/p\u003e\n\u003cp\u003eIR\u003c/p\u003e\n\u003cp\u003einsulin resistance\u003c/p\u003e\n\u003cp\u003eLDA\u003c/p\u003e\n\u003cp\u003elinear discriminant analysis\u003c/p\u003e\n\u003cp\u003eLDL\u003c/p\u003e\n\u003cp\u003elow-density lipoprotein cholesterol\u003c/p\u003e\n\u003cp\u003eLightGBM\u003c/p\u003e\n\u003cp\u003elight gradient boosting machine\u003c/p\u003e\n\u003cp\u003eLR\u003c/p\u003e\n\u003cp\u003elogistic regression\u003c/p\u003e\n\u003cp\u003eMASLD\u003c/p\u003e\n\u003cp\u003emetabolic dysfunction-associated steatotic liver disease\u003c/p\u003e\n\u003cp\u003eMETS-IR\u003c/p\u003e\n\u003cp\u003emetabolic score for insulin resistance\u003c/p\u003e\n\u003cp\u003eMICE\u003c/p\u003e\n\u003cp\u003emultiple imputation by chained equations\u003c/p\u003e\n\u003cp\u003eMIMIC-IV\u003c/p\u003e\n\u003cp\u003eMedical Information Mart for Intensive Care IV\u003c/p\u003e\n\u003cp\u003eNB\u003c/p\u003e\n\u003cp\u003enaive bayes\u003c/p\u003e\n\u003cp\u003eNIH\u003c/p\u003e\n\u003cp\u003eNational Institutes of Health\u003c/p\u003e\n\u003cp\u003ePREVENT\u003c/p\u003e\n\u003cp\u003eAmerican Heart Association\u0026apos;s PREVENT Equations\u003c/p\u003e\n\u003cp\u003eQDA\u003c/p\u003e\n\u003cp\u003equadratic discriminant analysis\u003c/p\u003e\n\u003cp\u003eRBC\u003c/p\u003e\n\u003cp\u003ered blood cell\u003c/p\u003e\n\u003cp\u003eRCS\u003c/p\u003e\n\u003cp\u003erestricted cubic splines\u003c/p\u003e\n\u003cp\u003eRF\u003c/p\u003e\n\u003cp\u003erandom forest\u003c/p\u003e\n\u003cp\u003eRDW\u003c/p\u003e\n\u003cp\u003ered blood cell distribution width\u003c/p\u003e\n\u003cp\u003eRR\u003c/p\u003e\n\u003cp\u003erespiratory rate\u003c/p\u003e\n\u003cp\u003eSBP\u003c/p\u003e\n\u003cp\u003esystolic blood pressure\u003c/p\u003e\n\u003cp\u003eSD\u003c/p\u003e\n\u003cp\u003estandard deviation\u003c/p\u003e\n\u003cp\u003eSHAP\u003c/p\u003e\n\u003cp\u003eSHapley Additive exPlanations\u003c/p\u003e\n\u003cp\u003eSOFA\u003c/p\u003e\n\u003cp\u003eSequential Organ Failure Assessment\u003c/p\u003e\n\u003cp\u003eSpO2\u003c/p\u003e\n\u003cp\u003eperipheral oxygen saturation\u003c/p\u003e\n\u003cp\u003eSQL\u003c/p\u003e\n\u003cp\u003estructured query language\u003c/p\u003e\n\u003cp\u003eTC\u003c/p\u003e\n\u003cp\u003etotal cholesterol\u003c/p\u003e\n\u003cp\u003eTG\u003c/p\u003e\n\u003cp\u003etriglyceride\u003c/p\u003e\n\u003cp\u003eTyG\u003c/p\u003e\n\u003cp\u003etriglyceride-glucose\u003c/p\u003e\n\u003cp\u003eVIF\u003c/p\u003e\n\u003cp\u003evariance inflation factor\u003c/p\u003e\n\u003cp\u003eWBC\u003c/p\u003e\n\u003cp\u003ewhite blood cell\u003c/p\u003e\n\u003cp\u003eXGBoost\u003c/p\u003e\n\u003cp\u003eextreme gradient boosting\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor Contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eJJL, MHL, and YHW came up with the article concept and design ideas, contributed to the methodology, and wrote the initial draft. JJL performed the data curation and led the formal analysis and visualization, with MHL and YHW contributing equally to the formal analysis and visualization; MHL also performed the validation. MHW acquired the funding, administered the project, and supervised the study. JGF carried out the investigation and provided the resources and software. MHW and JGF were pivotal in revising the manuscript. The final manuscript was reviewed and approved by all study contributors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe work was supported by the Sichuan Province Science and Technology Support Program 2022YFS0632.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNo datasets were generated or analysed during the current study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eUse of the MIMIC-IV database was approved by the review committees of Massachusetts Institute of Technology and Beth Israel Deaconess Medical Center. As the data are publicly available, ethical approval and informed consent were waived for this study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors report no conflicts of interest in this work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFootnotes\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePublisher\u0026rsquo;s note\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSpringer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eMartin SS, Aday AW, Allen NB, et al. 2025 Heart Disease and Stroke Statistics: A Report of US and Global Data From the American Heart Association. \u003cem\u003eCirculation\u003c/em\u003e. 2025;151(8):e41-e660. doi:10.1161/CIR.0000000000001303\u003c/li\u003e\n\u003cli\u003eGBD 2021 Diseases and Injuries Collaborators. Global incidence, prevalence, years lived with disability (YLDs), disability-adjusted life-years (DALYs), and healthy life expectancy (HALE) for 371 diseases and injuries in 204 countries and territories and 811 subnational locations, 1990-2021: a systematic analysis for the Global Burden of Disease Study 2021. \u003cem\u003eLancet\u003c/em\u003e. 2024;403(10440):2133-2161. doi:10.1016/S0140-6736(24)00757-8\u003c/li\u003e\n\u003cli\u003eWang H, Jiang G, Qu X, Zhou J. Association of triglyceride glucose index with mortality in critically ill patients with atherosclerotic cardiovascular disease: analysis of the MIMIC-IV database. \u003cem\u003eEur J Med Res\u003c/em\u003e. Published online February 5, 2026. doi:10.1186/s40001-026-03994-w\u003c/li\u003e\n\u003cli\u003ePlummer MP, Bellomo R, Cousins CE, et al. Dysglycaemia in the critically ill and the interaction of chronic and acute glycaemia with mortality. \u003cem\u003eIntensive Care Med\u003c/em\u003e. 2014;40(7):973-980. doi:10.1007/s00134-014-3287-7\u003c/li\u003e\n\u003cli\u003eLiu Y, Fu H, Wang Y, et al. Association between the low-density lipoprotein to high-density lipoprotein ratio and prognosis in critically ill intracerebral hemorrhage patients: a retrospective cohort study from the MIMIC-IV database. \u003cem\u003eLipids Health Dis\u003c/em\u003e. 2025;24(1):36. doi:10.1186/s12944-025-02459-z\u003c/li\u003e\n\u003cli\u003eWang F, Guo Y, Tang Y, et al. Combined assessment of stress hyperglycemia ratio and glycemic variability to predict all-cause mortality in critically ill patients with atherosclerotic cardiovascular diseases across different glucose metabolic states: an observational cohort study with machine learning. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2025;24(1):199. doi:10.1186/s12933-025-02762-0\u003c/li\u003e\n\u003cli\u003eLin Z, Yuan S, Li B, et al. Insulin-based or non-insulin-based insulin resistance indicators and risk of long-term cardiovascular and all-cause mortality in the general population: A 25-year cohort study. \u003cem\u003eDiabetes Metab\u003c/em\u003e. 2024;50(5):101566. doi:10.1016/j.diabet.2024.101566\u003c/li\u003e\n\u003cli\u003eSingh B, Saxena A. Surrogate markers of insulin resistance: A review. \u003cem\u003eWorld J Diabetes\u003c/em\u003e. 2010;1(2):36-47. doi:10.4239/wjd.v1.i2.36\u003c/li\u003e\n\u003cli\u003eOuyang Q, Zhang C, Zhang H, Xiong T, Chen Y, Zhang P. Evaluation of four insulin resistance surrogate indicators for predicting short-term mortality in critically ill patients with ischemic stroke: a nationwide retrospective cohort study. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2025;24(1):409. doi:10.1186/s12933-025-02956-6\u003c/li\u003e\n\u003cli\u003eHe G, Zhang Z, Wang C, et al. Association of the triglyceride-glucose index with all-cause and cause-specific mortality: a population-based cohort study of 3.5 million adults in China. \u003cem\u003eLancet Reg Health West Pac\u003c/em\u003e. 2024;49:101135. doi:10.1016/j.lanwpc.2024.101135\u003c/li\u003e\n\u003cli\u003eFahed G, Aoun L, Bou Zerdan M, et al. Metabolic Syndrome: Updates on Pathophysiology and Management in 2021. \u003cem\u003eInt J Mol Sci\u003c/em\u003e. 2022;23(2):786. doi:10.3390/ijms23020786\u003c/li\u003e\n\u003cli\u003eMo D, Zhang P, Zhang M, Dai H, Guan J. Cholesterol, high-density lipoprotein, and glucose index versus triglyceride-glucose index in predicting cardiovascular disease risk: a cohort study. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2025;24(1):116. doi:10.1186/s12933-025-02675-y\u003c/li\u003e\n\u003cli\u003eŻerdziński K, Janiec J, J\u0026oacute;źwik K, Łajczak P, Krzych ŁJ. Artificial Intelligence in Intensive Care: An Overview of Systematic Reviews with Clinical Maturity and Readiness Mapping. \u003cem\u003eJ Clin Med\u003c/em\u003e. 2025;15(1):185. doi:10.3390/jcm15010185\u003c/li\u003e\n\u003cli\u003eSun H, Kang M, Zhang H, Jia J, Wang Q. Machine Learning for Predicting Mortality in Intensive Care Unit Patients: A Prognostic Performance Systematic Review and Meta-Analysis. \u003cem\u003eNurs Crit Care\u003c/em\u003e. 2025;30(6):e70206. doi:10.1111/nicc.70206\u003c/li\u003e\n\u003cli\u003eMirzakhani F, Sadoughi F, Hatami M, Amirabadizadeh A. Which model is superior in predicting ICU survival: artificial intelligence versus conventional approaches. \u003cem\u003eBMC Med Inform Decis Mak\u003c/em\u003e. 2022;22(1):167. doi:10.1186/s12911-022-01903-9\u003c/li\u003e\n\u003cli\u003eAli S, Akhlaq F, Imran AS, Kastrati Z, Daudpota SM, Moosa M. The enlightening role of explainable artificial intelligence in medical \u0026amp; healthcare domains: A systematic literature review. \u003cem\u003eComput Biol Med\u003c/em\u003e. 2023;166:107555. doi:10.1016/j.compbiomed.2023.107555\u003c/li\u003e\n\u003cli\u003eJohnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. \u003cem\u003eSci Data\u003c/em\u003e. 2023;10(1):1. doi:10.1038/s41597-022-01899-x\u003c/li\u003e\n\u003cli\u003eKhan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association\u0026rsquo;s PREVENT Equations. \u003cem\u003eCirculation\u003c/em\u003e. 2024;149(6):430-449. doi:10.1161/CIRCULATIONAHA.123.067626\u003c/li\u003e\n\u003cli\u003eMansoori A, Nosrati M, Dorchin M, et al. A novel index for diagnosis of type 2 diabetes mellitus: Cholesterol, High density lipoprotein, and Glucose (CHG) index. \u003cem\u003eJ Diabetes Investig\u003c/em\u003e. 2025;16(2):309-314. doi:10.1111/jdi.14343\u003c/li\u003e\n\u003cli\u003eAustin PC, White IR, Lee DS, van Buuren S. Missing Data in Clinical Research: A Tutorial on Multiple Imputation. \u003cem\u003eCan J Cardiol\u003c/em\u003e. 2021;37(9):1322-1331. doi:10.1016/j.cjca.2020.11.010\u003c/li\u003e\n\u003cli\u003eXu J, Cai D, Jiao Y, et al. Insights into the complex relationship between triglyceride glucose-waist height ratio index, mean arterial pressure, and cardiovascular disease: a nationwide prospective cohort study. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2025;24(1):93. doi:10.1186/s12933-025-02657-0\u003c/li\u003e\n\u003cli\u003eDegenhardt F, Seifert S, Szymczak S. Evaluation of variable selection methods for random forests and omics data sets. \u003cem\u003eBrief Bioinform\u003c/em\u003e. 2019;20(2):492-503. doi:10.1093/bib/bbx124\u003c/li\u003e\n\u003cli\u003eGisperg F, Klausser R, Elshazly M, Kopp J, Brichtov\u0026aacute; EP, Spadiut O. Bayesian Optimization in Bioprocess Engineering-Where Do We Stand Today? \u003cem\u003eBiotechnol Bioeng\u003c/em\u003e. 2025;122(6):1313-1325. doi:10.1002/bit.28960\u003c/li\u003e\n\u003cli\u003eKhoja A, Andraweera PH, Lassi ZS, et al. Modifiable and Non-Modifiable Risk Factors for Premature Coronary Heart Disease (PCHD): Systematic Review and Meta-Analysis. \u003cem\u003eHeart Lung Circ\u003c/em\u003e. 2024;33(3):265-280. doi:10.1016/j.hlc.2023.12.012\u003c/li\u003e\n\u003cli\u003eGanjali S, Watts GF, Banach M, Reiner Ž, Nachtigal P, Sahebkar A. The Yin and Yang of High-density Lipoprotein and Atherosclerotic Cardiovascular Disease: Focusing on Functionality and Cholesterol Efflux to Reframe the HDL Hypothesis. \u003cem\u003eCurr Med Chem\u003c/em\u003e. 2021;28(29):6066-6081. doi:10.2174/0929867328666210208182326\u003c/li\u003e\n\u003cli\u003eAlkatiri AH, Qalby N, Mappangara I, et al. Stress hyperglycemia and poor outcomes in patients with ST-elevation myocardial infarction: a systematic review and meta-analysis. \u003cem\u003eFront Cardiovasc Med\u003c/em\u003e. 2024;11:1303685. doi:10.3389/fcvm.2024.1303685\u003c/li\u003e\n\u003cli\u003eMottillo S, Filion KB, Genest J, et al. The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e. 2010;56(14):1113-1132. doi:10.1016/j.jacc.2010.05.034\u003c/li\u003e\n\u003cli\u003eWei S, Yu S, Qian C, Xu Z, Jia Y, Chen B. Exploring the U-shaped nonlinear relationship of CHG index with metabolic syndrome and mortality risks in metabolic syndrome patients. \u003cem\u003eLipids Health Dis\u003c/em\u003e. 2025;24(1):294. doi:10.1186/s12944-025-02714-3\u003c/li\u003e\n\u003cli\u003eZhu H, Liu L, Yang S, et al. Association of cholesterol, high-density lipoprotein and glucose (CHG) index with mortality risk in metabolic dysfunction-associated steatotic liver disease (MASLD) adults: results from two prospective cohorts. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2026;25(1):23. doi:10.1186/s12933-026-03079-2\u003c/li\u003e\n\u003cli\u003eKe L, Li Y, Jiang S, Xing W, Zhao L. Associations between the CHG index, its modified versions and incident stroke in patients with early-stage CKM syndrome: a nationwide cohort study. \u003cem\u003eEur J Med Res\u003c/em\u003e. 2026;31(1):206. doi:10.1186/s40001-025-03808-5\u003c/li\u003e\n\u003cli\u003eGuo Z, Xiong Z, Zhang W, et al. Association between cholesterol, high-density lipoprotein, and glucose index and risks of cardiovascular and all-cause mortality in patients with calcific aortic valve stenosis: the ARISTOTLE cohort study. \u003cem\u003eCardiovasc Diabetol\u003c/em\u003e. 2025;24(1):393. doi:10.1186/s12933-025-02906-2\u003c/li\u003e\n\u003cli\u003eLin Y, Ke J, Chen S, et al. Cholesterol, high-density lipoprotein, glucose index as a novel marker for predicting in-stent restenosis after drug-eluting stent implantation in patients with acute coronary syndrome. \u003cem\u003eFront Cardiovasc Med\u003c/em\u003e. 2026;13:1740216. doi:10.3389/fcvm.2026.1740216\u003c/li\u003e\n\u003cli\u003eTang J, Wei S, Li Y, Zhou Y. Assessment of cholesterol-HDL-glucose index in anticipating risk of cardiometabolic diseases: a comparative study with triglyceride-glucose index. \u003cem\u003eSci Rep\u003c/em\u003e. 2025;15(1):45356. doi:10.1038/s41598-025-30024-2\u003c/li\u003e\n\u003cli\u003eLu Y, Cui X, Zhang L, et al. The Functional Role of Lipoproteins in Atherosclerosis: Novel Directions for Diagnosis and Targeting Therapy. \u003cem\u003eAging Dis\u003c/em\u003e. 2022;13(2):491-520. doi:10.14336/AD.2021.0929\u003c/li\u003e\n\u003cli\u003eChistiakov DA, Bobryshev YV, Orekhov AN. Macrophage-mediated cholesterol handling in atherosclerosis. \u003cem\u003eJ Cell Mol Med\u003c/em\u003e. 2016;20(1):17-28. doi:10.1111/jcmm.12689\u003c/li\u003e\n\u003cli\u003eLuc K, Schramm-Luc A, Guzik TJ, Mikolajczyk TP. Oxidative stress and inflammatory markers in prediabetes and diabetes. \u003cem\u003eJ Physiol Pharmacol\u003c/em\u003e. 2019;70(6). doi:10.26402/jpp.2019.6.01\u003c/li\u003e\n\u003cli\u003eMengstie MA, Chekol Abebe E, Behaile Teklemariam A, et al. Endogenous advanced glycation end products in the pathogenesis of chronic diabetic complications. \u003cem\u003eFront Mol Biosci\u003c/em\u003e. 2022;9:1002710. doi:10.3389/fmolb.2022.1002710\u003c/li\u003e\n\u003cli\u003eAdorni MP, Ronda N, Bernini F, Zimetti F. High Density Lipoprotein Cholesterol Efflux Capacity and Atherosclerosis in Cardiovascular Disease: Pathophysiological Aspects and Pharmacological Perspectives. \u003cem\u003eCells\u003c/em\u003e. 2021;10(3):574. doi:10.3390/cells10030574\u003c/li\u003e\n\u003cli\u003eAllard-Ratick MP, Kindya BR, Khambhati J, et al. HDL: Fact, fiction, or function? HDL cholesterol and cardiovascular risk. \u003cem\u003eEur J Prev Cardiol\u003c/em\u003e. 2021;28(2):166-173. doi:10.1177/2047487319848214\u003c/li\u003e\n\u003cli\u003eDenimal D. Antioxidant and Anti-Inflammatory Functions of High-Density Lipoprotein in Type 1 and Type 2 Diabetes. \u003cem\u003eAntioxidants (Basel)\u003c/em\u003e. 2023;13(1):57. doi:10.3390/antiox13010057\u003c/li\u003e\n\u003cli\u003eLibby P. Inflammation during the life cycle of the atherosclerotic plaque. \u003cem\u003eCardiovasc Res\u003c/em\u003e. 2021;117(13):2525-2536. doi:10.1093/cvr/cvab303\u003c/li\u003e\n\u003cli\u003eSilveira Rossi JL, Barbalho SM, Reverete de Araujo R, Bechara MD, Sloan KP, Sloan LA. Metabolic syndrome and cardiovascular diseases: Going beyond traditional risk factors. \u003cem\u003eDiabetes Metab Res Rev\u003c/em\u003e. 2022;38(3):e3502. doi:10.1002/dmrr.3502\u003c/li\u003e\n\u003cli\u003eChafai N, Bonizzi L, Botti S, Badaoui B. Emerging applications of machine learning in genomic medicine and healthcare. \u003cem\u003eCrit Rev Clin Lab Sci\u003c/em\u003e. 2024;61(2):140-163. doi:10.1080/10408363.2023.2259466\u003c/li\u003e\n\u003cli\u003eSalluh JIF, Moralez GM, Tracy A, Deliberato RO. ICU scoring systems: current perspectives and future directions. \u003cem\u003eCurr Opin Crit Care\u003c/em\u003e. 2025;31(5):608-615. doi:10.1097/MCC.0000000000001305\u003c/li\u003e\n\u003cli\u003eYu MY, Yoo HY, Han GI, Kim EJ, Son YJ. Comparing the Performance of Machine Learning Models and Conventional Risk Scores for Predicting Major Adverse Cardiovascular Cerebrovascular Events After Percutaneous Coronary Intervention in Patients With Acute Myocardial Infarction: Systematic Review and Meta-Analysis. \u003cem\u003eJ Med Internet Res\u003c/em\u003e. 2025;27:e76215. doi:10.2196/76215\u003c/li\u003e\n\u003cli\u003eGrant L, Joo P, Nemnom MJ, Thiruganasambandamoorthy V. Machine learning versus traditional methods for the development of risk stratification scores: a case study using original Canadian Syncope Risk Score data. \u003cem\u003eIntern Emerg Med\u003c/em\u003e. 2022;17(4):1145-1153. doi:10.1007/s11739-021-02873-y\u003c/li\u003e\n\u003cli\u003eGanie SM, Pramanik PKD, Zhao Z. Ensemble learning with explainable AI for improved heart disease prediction based on multiple datasets. \u003cem\u003eSci Rep\u003c/em\u003e. 2025;15(1):13912. doi:10.1038/s41598-025-97547-6\u003c/li\u003e\n\u003cli\u003eEnsemble Learning for Disease Prediction: A Review - PubMed. Accessed February 25, 2026. https://pubmed.ncbi.nlm.nih.gov/37372925/\u003c/li\u003e\n\u003cli\u003eCiobanu-Caraus O, Aicher A, Kernbach JM, Regli L, Serra C, Staartjes VE. A critical moment in machine learning in medicine: on reproducible and interpretable learning. \u003cem\u003eActa Neurochir (Wien)\u003c/em\u003e. 2024;166(1):14. doi:10.1007/s00701-024-05892-8\u003c/li\u003e\n\u003cli\u003eBernal L, Rastelli G, Pinzi L. Improving Machine Learning Classification Predictions through SHAP and Features Analysis Interpretation. \u003cem\u003eJ Chem Inf Model\u003c/em\u003e. 2025;65(21):11716-11732. doi:10.1021/acs.jcim.5c02015\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-cardiovascular-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcar","sideBox":"Learn more about [BMC Cardiovascular Disorders](http://bmccardiovascdisord.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcar/default.aspx","title":"BMC Cardiovascular Disorders","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Atherosclerotic cardiovascular disease, All-Cause mortality, CHG, Machine learning, MIMIC-IV","lastPublishedDoi":"10.21203/rs.3.rs-9035595/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9035595/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground:\u003c/strong\u003e The cholesterol, high-density lipoprotein, and glucose (CHG) index, as a novel comprehensive marker of lipid and glucose metabolism, has yet to be fully elucidated regarding its prognostic value for all-cause mortality in critically ill ASCVD patients. The present study aims to ascertain the correlation between CHG index and mortality in this population and to identify key predictors using machine learning techniques.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods:\u003c/strong\u003e Patients diagnosed with ASCVD were enrolled from the Medical Information Mart for Intensive Care (MIMIC)-IV database. Patients were divided into five groups based on CHG index values. The association between the CHG index and mortality was evaluated using Kaplan-Meier curves, Cox proportional hazards models, restricted cubic splines (RCS), and subgroup analyses. Ten machine learning models were applied to predict mortality risk, with the SHapley Additive exPlanations (SHAP) method used to identify key predictors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults:\u003c/strong\u003e 1,959 patients were involved (median age 71.99 years; 52.8% male). Following multivariate adjustment, a one-unit increase in the CHG index was found to be significantly associated with an elevated mortality risk (30-day HR: 1.58, 95% CI: 1.22–2.05; 90-day HR: 1.61, 95% CI: 1.28–2.03). In comparison with patients in the first quintile, those in the fifth quintile demonstrated the highest mortality risk (30-day HR: 2.10, 95% CI: 1.34–3.29; 90-day HR: 2.07, 95% CI: 1.39–3.07). RCS analysis demonstrated a linear positive association between CHG index and mortality. Among machine learning models, the Stacking Classifier demonstrated the most optimal predictive performance for 30-day mortality, with an Area Under the Curve (AUC) of 0.882 in the training set and 0.833 in the test set. The SHAP analysis identified the CHG index as a key predictor.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions:\u003c/strong\u003e The elevated CHG index demonstrated a direct correlation with an augmented mortality risk in critically ill ASCVD patients. The CHG index has the potential to be a valuable predictor for mortality risk assessment in this population.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eClinical trial number:\u003c/strong\u003e not applicable.\u003c/p\u003e","manuscriptTitle":"Association between the Cholesterol, High-Density Lipoprotein, and Glucose Index and All-Cause Mortality in Critically Ill Patients with Atherosclerotic Cardiovascular Disease: A machine learning-based retrospective cohort study from the MIMIC-IV database","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-09 14:56:18","doi":"10.21203/rs.3.rs-9035595/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-04-27T07:49:12+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-15T07:01:32+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"27990774577391018903061123868575976388","date":"2026-04-15T02:00:51+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-14T06:57:43+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"227257953678695009171681848505490494431","date":"2026-04-14T03:57:11+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-10T17:10:10+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"169353025226286459753685233159890765670","date":"2026-04-10T16:49:22+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"106249807104638331032059458078459240861","date":"2026-04-09T08:09:12+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-04-02T11:14:34+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"100602981805963110016278366726554496142","date":"2026-04-02T11:00:11+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-02T10:24:03+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-09T14:00:00+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-07T14:25:48+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-07T14:24:47+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Cardiovascular Disorders","date":"2026-03-05T04:04:33+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-cardiovascular-disorders","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bcar","sideBox":"Learn more about [BMC Cardiovascular Disorders](http://bmccardiovascdisord.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bcar/default.aspx","title":"BMC Cardiovascular Disorders","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"0f3232a2-558e-4631-b37c-d1ec8b439af2","owner":[],"postedDate":"April 9th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-04-09T14:56:18+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-09 14:56:18","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9035595","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9035595","identity":"rs-9035595","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}